Lab1-XGBoost-For-Timeseries (1).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<img src=\"images/header.png\" alt=\"Header\" style=\"width: 400px;\"/>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h1 align=\"center\">Predictive Maintenance using NVIDIA RAPIDS and Deep Learning Models</h1>\n",
    "<h4 align=\"center\">Part 1: Training GPU XGBoost models with RAPIDS for Time Series</h4>\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Working with time series algorithms can be difficult to configure and apply to arbitrary sequence prediction problems, even with well-defined and “easy to use” interfaces like those provided in the tf.keras deep learning library in Python.  One reason for this difficulty is the need for Deep Learning layers to input and return sequences rather than single values.\n",
    "\n",
    "In this course, we will take a publicly available time series dataset of hard drive operational metrics provided by <a href=\"https://www.backblaze.com/b2/hard-drive-test-data.html\">Backblaze</a>, to build out and compare machine learning vs deep learning approaches to predictive maintenance problems. \n",
    "\n",
    "The course materials are categorized in three sections. In this lab, we take a machine learning approach called “XGBoost” for classification of faulty/normal hard drives and discuss performance optimizations by utilizing accelerated computing. In lab 2, we move on to deep learning models and start with recurrent networks to predict the time series. We further explore variations of recurrent networks such as CNN-LSTM models. Lab 3, addresses applications of autoencoders to classify hard drives into faulty and normal bins. Throughout these labs, you will have the opportunity to get familiarized with details of implementations for each model, and learn how to curate the data and evaluate results. To begin, we start with \"Training GPU XGBoost models with RAPIDS for Time Series\". This lab covers the following topics:\n",
    "\n",
    "* [Background](#1)\n",
    "* [Predictive Maintenance: Predicting Anomalies Accurately and Early](#2)\n",
    "* [Lab Overview](#3)\n",
    "* [Dataset Loading and Pre-Processing](#4)\n",
    "* [Training GPU XGBoost models with RAPIDS for Time Series](#5)\n",
    "* [XGBoost](#6)\n",
    "* [More on Data Preparation](#7)\n",
    "    * [Exercise 1: Remove \"zero-valued\" columns](#e1)\n",
    "* [Define XGBoost Model Parameters](#8)\n",
    "* [Train XGBoost on CPUs](#9)\n",
    "* [Training with GPU without cuDF DataFrame](#10)\n",
    "* [Training with GPU and cuDF](#11)\n",
    "* [XGBoost Model Accuracy](#12)\n",
    "* [Measuring Model Accuracy](#13)\n",
    "* [Interpreting the results](#14)\n",
    "* [Dealing with Imbalanced Data](#15)\n",
    "    * [Exercise 2: Implications of balancing data samples](#e2)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"1\"></a>\n",
    "## Background\n",
    "\n",
    "One of the important recent topics of interest for industrial organizations  is to move away from reactive and schedule-based maintenance to <b>predictive maintenance</b> that helps optimize the operations and improve productivity.\n",
    "\n",
    "<a name=\"2\"></a>\n",
    "### Predictive Maintenance: Predicting Anomalies Accurately and Early\n",
    "\n",
    "Whereas schedule-based maintenance provides a regularly scheduled time for part replacements and repairs, predictive maintenance represents a condition based approach driven by data and analytics. Predictive maintenance strives to identify anomalies that can lead to costly breakdowns, such as improper machinery lubrication, contamination, misalignment, or suboptimal humidity and temperature conditions.\n",
    "\n",
    "By identifying such anomalies of many part failures, estimating remaining useful life of these parts and then mapping anomalies to failure conditions by interpreting service logs, predictive maintenance helps to manage failures and avoid costly unplanned downtime. Other benefits include fewer unnecessary repairs, optimized spare parts inventory, and longer lifespans of equipment and parts—ultimately increasing equipment availability while improving operational efficiency. \n",
    "\n",
    "In this regard, our goal in this lab is to show you how you can take your company's time series data and leverage it to predict outcomes.\n",
    "\n",
    "<a name=\"3\"></a>\n",
    "## Lab Overview\n",
    "\n",
    "One of the common problems in a datacenter is failing data storage disks. Current business applications transfer large amounts of data and put very big pressure on the disks.  Using hard drive S.M.A.R.T. (Self-Monitoring Analysis and Reporting Technology) data provided from Backblaze, we will build conventional Machine Learning models in addition to Deep Learning methods to predict RUL (Remaining Useful Life) of a hard drive.\n",
    "\n",
    "- Dataset - https://www.backblaze.com/b2/hard-drive-test-data.html\n",
    "- Research paper from IBM - https://researcher.watson.ibm.com/researcher/files/us-mqiao/BigDataCongress_2018.pdf\n",
    "\n",
    "After completing this Lab, you will know how to take time series data and:\n",
    "\n",
    "- Predict Failure using a machine learning classification model with XGBoost\n",
    "- Predict Failure using a deep learning classification model using LSTM\n",
    "- Detect Anomalies using an Autoencoder (AE) or Seq2Seq model\n",
    "\n",
    "\n",
    "Let's start by importing some libraries used throughout this lab."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import libraries that will be needed for the Lab\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "import sys\n",
    "import glob\n",
    "import os\n",
    "\n",
    "import datetime\n",
    "import time\n",
    "import gc\n",
    "import math\n",
    "import random\n",
    "from random import shuffle, randrange\n",
    "\n",
    "from sklearn.utils import shuffle\n",
    "from sklearn import preprocessing\n",
    "from sklearn.utils import shuffle\n",
    "from sklearn.metrics import *\n",
    "\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"4\"></a>\n",
    "## Dataset Loading and Pre-Processing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The data set consists of a daily snapshot of each operational hard drive in the Backblaze datacenter at the time of measurement.  Each snapshot includes basic drive information along with S.M.A.R.T. statistics reported by that drive.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define location of our input and prepared data files\n",
    "data_dir = './data/'\n",
    "\n",
    "csv_train_file = 'Lab1-2017-Full_data.csv.gz'\n",
    "csv_test_file = 'Lab1-2016-Q4_data.csv.gz'\n",
    "\n",
    "pkl_train_file = 'Lab1-2017-Full-ST4000DM000.pkl'\n",
    "pkl_test_file = 'Lab1-2016-Q4-ST4000DM000.pkl'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reading training data set...\n",
      "Finished reading training data set\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>serial_number</th>\n",
       "      <th>model</th>\n",
       "      <th>capacity_bytes</th>\n",
       "      <th>failure</th>\n",
       "      <th>smart_1_normalized</th>\n",
       "      <th>smart_1_raw</th>\n",
       "      <th>smart_2_normalized</th>\n",
       "      <th>smart_2_raw</th>\n",
       "      <th>smart_3_normalized</th>\n",
       "      <th>...</th>\n",
       "      <th>smart_250_normalized</th>\n",
       "      <th>smart_250_raw</th>\n",
       "      <th>smart_251_normalized</th>\n",
       "      <th>smart_251_raw</th>\n",
       "      <th>smart_252_normalized</th>\n",
       "      <th>smart_252_raw</th>\n",
       "      <th>smart_254_normalized</th>\n",
       "      <th>smart_254_raw</th>\n",
       "      <th>smart_255_normalized</th>\n",
       "      <th>smart_255_raw</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2017-02-01</td>\n",
       "      <td>MJ0351YNG9Z0XA</td>\n",
       "      <td>Hitachi HDS5C3030ALA630</td>\n",
       "      <td>3000592982016</td>\n",
       "      <td>0</td>\n",
       "      <td>100.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>135.0</td>\n",
       "      <td>108.0</td>\n",
       "      <td>127.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2017-02-01</td>\n",
       "      <td>MJ0351YNG9WJSA</td>\n",
       "      <td>Hitachi HDS5C3030ALA630</td>\n",
       "      <td>3000592982016</td>\n",
       "      <td>0</td>\n",
       "      <td>100.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>136.0</td>\n",
       "      <td>104.0</td>\n",
       "      <td>126.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2017-02-01</td>\n",
       "      <td>PL1321LAG34XWH</td>\n",
       "      <td>Hitachi HDS5C4040ALE630</td>\n",
       "      <td>4000787030016</td>\n",
       "      <td>0</td>\n",
       "      <td>100.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>134.0</td>\n",
       "      <td>101.0</td>\n",
       "      <td>130.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2017-02-01</td>\n",
       "      <td>MJ0351YNGABYAA</td>\n",
       "      <td>Hitachi HDS5C3030ALA630</td>\n",
       "      <td>3000592982016</td>\n",
       "      <td>0</td>\n",
       "      <td>100.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>136.0</td>\n",
       "      <td>104.0</td>\n",
       "      <td>137.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2017-02-01</td>\n",
       "      <td>Z305B2QN</td>\n",
       "      <td>ST4000DM000</td>\n",
       "      <td>4000787030016</td>\n",
       "      <td>0</td>\n",
       "      <td>117.0</td>\n",
       "      <td>124084032.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>91.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 95 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         date   serial_number                    model  capacity_bytes  \\\n",
       "0  2017-02-01  MJ0351YNG9Z0XA  Hitachi HDS5C3030ALA630   3000592982016   \n",
       "1  2017-02-01  MJ0351YNG9WJSA  Hitachi HDS5C3030ALA630   3000592982016   \n",
       "2  2017-02-01  PL1321LAG34XWH  Hitachi HDS5C4040ALE630   4000787030016   \n",
       "3  2017-02-01  MJ0351YNGABYAA  Hitachi HDS5C3030ALA630   3000592982016   \n",
       "4  2017-02-01        Z305B2QN              ST4000DM000   4000787030016   \n",
       "\n",
       "   failure  smart_1_normalized  smart_1_raw  smart_2_normalized  smart_2_raw  \\\n",
       "0        0               100.0          0.0               135.0        108.0   \n",
       "1        0               100.0          0.0               136.0        104.0   \n",
       "2        0               100.0          0.0               134.0        101.0   \n",
       "3        0               100.0          0.0               136.0        104.0   \n",
       "4        0               117.0  124084032.0                 NaN          NaN   \n",
       "\n",
       "   smart_3_normalized  ...  smart_250_normalized  smart_250_raw  \\\n",
       "0               127.0  ...                   NaN            NaN   \n",
       "1               126.0  ...                   NaN            NaN   \n",
       "2               130.0  ...                   NaN            NaN   \n",
       "3               137.0  ...                   NaN            NaN   \n",
       "4                91.0  ...                   NaN            NaN   \n",
       "\n",
       "   smart_251_normalized  smart_251_raw  smart_252_normalized  smart_252_raw  \\\n",
       "0                   NaN            NaN                   NaN            NaN   \n",
       "1                   NaN            NaN                   NaN            NaN   \n",
       "2                   NaN            NaN                   NaN            NaN   \n",
       "3                   NaN            NaN                   NaN            NaN   \n",
       "4                   NaN            NaN                   NaN            NaN   \n",
       "\n",
       "   smart_254_normalized  smart_254_raw  smart_255_normalized  smart_255_raw  \n",
       "0                   NaN            NaN                   NaN            NaN  \n",
       "1                   NaN            NaN                   NaN            NaN  \n",
       "2                   NaN            NaN                   NaN            NaN  \n",
       "3                   NaN            NaN                   NaN            NaN  \n",
       "4                   NaN            NaN                   NaN            NaN  \n",
       "\n",
       "[5 rows x 95 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Read entire SMART training data set\n",
    "# - Full 2017 dataset takes about 4 minutes to read in\n",
    "\n",
    "print('Reading training data set...')\n",
    "df = pd.read_csv(data_dir + csv_train_file)\n",
    "print('Finished reading training data set')\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(30471787, 95)\n"
     ]
    }
   ],
   "source": [
    "print(df.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Set Overview<br>\n",
    "\n",
    "The dataset in this DLI course is provided by Backblaze. More documentation can be found [here](https://www.backblaze.com/b2/hard-drive-test-data.html) about the type of the data and the different sensors which indicate the health of the hard drives. Following is a list of key parameters reported in the dataset. \n",
    "\n",
    "- <b>Date:</b> The date of the entrie recording in yyyy-mm-dd format.\n",
    "\n",
    "- <b>Serial Number:</b> The manufacturer-assigned serial number of the drive.\n",
    "\n",
    "- <b>Model:</b> The manufacturer-assigned model number of the drive.\n",
    "\n",
    "- <b>Capacity:</b> The drive capacity in bytes.\n",
    "\n",
    "- <b>Failure:</b> Contains a “0” if the drive is OK. Contains a “1” if this is the last day the drive was operational before failing.\n",
    "\n",
    "- <b>SMART Stats:</b> – 90 columns of data, which are the Raw and Normalized values for 45 different SMART stats as reported by the given drive. Each value is the number reported by the drive.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the snapshot above, one important thing to notice is that we have several different models for which data is recorded. First, let's take a look at number of samples for each model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ST4000DM000             12237899\n",
      "HGST HMS5C4040BLE640     5154696\n",
      "ST8000DM002              3523493\n",
      "HGST HMS5C4040ALE640     2728104\n",
      "ST8000NM0055             2664020\n",
      "                          ...   \n",
      "TOSHIBA HDWE160               64\n",
      "WDC WD10EADX                  59\n",
      "WDC WD10EARS                  32\n",
      " 00MD00                       14\n",
      "WDC WD1000FYPS                11\n",
      "Name: model, Length: 68, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "# let's review the available hard drive models and select one with lot of samples for this exercise\n",
    "print(df.model.value_counts())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Because SMART stats usage can vary in meaning from manufacturer to manufacturer, we will limit our training data to a specific hard drive model, to reduce  potential of other manufactures' usage from influencing our model. We will pick the `ST4000DM000` model, which has the largest number of samples."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reading test data set...\n",
      "Finished reading test data set\n"
     ]
    }
   ],
   "source": [
    "# ST4000DM000 is the most popular drive so let's use this to train\n",
    "harddrive_model = 'ST4000DM000'\n",
    "\n",
    "# keep only the selected hard drive model we want in our training data set\n",
    "df = df[df.model == harddrive_model]\n",
    "\n",
    "# Load test set from Q2 data and pre-selected model\n",
    "print('Reading test data set...')\n",
    "df_t = pd.read_csv(data_dir + csv_test_file)\n",
    "print('Finished reading test data set')\n",
    "\n",
    "# keep only the selected hard drive model we want in our test data set\n",
    "df_t = df_t[df_t.model == harddrive_model]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "#converting string to date time format for sorting and plotting time series data \n",
    "\n",
    "df['date'] = pd.to_datetime(df['date'])\n",
    "df_t['date'] = pd.to_datetime(df_t['date'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_cmap(n, name='hsv'):\n",
    "    '''Returns a function that maps each index in 0, 1, ..., n-1 to a distinct \n",
    "    RGB color; the keyword argument name must be a standard mpl colormap name.'''\n",
    "    return plt.cm.get_cmap(name, n)\n",
    "\n",
    "def plot_df(df,loc_col_list,axs):   \n",
    "    df_normal = df[df['failure']==0]\n",
    "    df_failed = df[df['failure']==1]\n",
    "    \n",
    "    n_rows = len(loc_col_list)\n",
    "    cmap = get_cmap(n_rows) \n",
    "    color_list = random.sample(range(n_rows),n_rows)\n",
    "    \n",
    "    for i, cur_col in enumerate(loc_col_list):\n",
    "        cur_min = df[cur_col].min() \n",
    "        cur_max = df[cur_col].max()\n",
    "        \n",
    "        axs[i].plot(df_normal['date'], df_normal[cur_col],'.', color=cmap(color_list[i]), markersize=5,alpha=0.2)\n",
    "        axs[i].plot(df_failed['date'], df_failed[cur_col],'.', color=cmap(color_list[i]), markersize=20,alpha=0.6)\n",
    "        axs[i].set_ylabel(cur_col)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "def remove_constant_values(df,check_str): \n",
    "    col_list = df.columns.tolist()\n",
    "    loc_col_list = [] \n",
    "    \n",
    "    print(\"Range for    : {:>25s} {:>15s}\".format(\"min\",\"max\"))\n",
    "    \n",
    "    for cur_col in col_list:\n",
    "        if check_str in cur_col:\n",
    "            cur_min = df[cur_col].min() \n",
    "            cur_max = df[cur_col].max()\n",
    "\n",
    "            if cur_min != cur_max : \n",
    "                print(\"  {:20s}   {:15d} {:15d} \".format(cur_col,int(cur_min),int(cur_max)))\n",
    "                loc_col_list.append(cur_col)\n",
    "                \n",
    "    print(\"\")\n",
    "    return loc_col_list "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Cleaning and understanding input data \n",
    "While building any machine learning/deep learning model, it is important to understand the quality and structure of the input data. For example, there can be missing data for several days for some of the sensors. There can also be sensors which do not have any data. The input data needs to be curated to remove such sensors and adjust the data for the sensors with missing values. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Range for    :                       min             max\n",
      "  smart_1_raw                      20608       243737992 \n",
      "  smart_4_raw                          1             297 \n",
      "  smart_5_raw                          0           43248 \n",
      "  smart_7_raw                          7 218859378419151 \n",
      "  smart_9_raw                          0           39525 \n",
      "  smart_12_raw                         0              76 \n",
      "  smart_183_raw                        0             684 \n",
      "  smart_184_raw                        0              62 \n",
      "  smart_187_raw                        0             690 \n",
      "  smart_188_raw                        0               7 \n",
      "  smart_189_raw                        0             264 \n",
      "  smart_190_raw                       14              42 \n",
      "  smart_192_raw                        0             285 \n",
      "  smart_193_raw                        5          574352 \n",
      "  smart_194_raw                       14              42 \n",
      "  smart_197_raw                        0           21608 \n",
      "  smart_198_raw                        0           21608 \n",
      "  smart_199_raw                        0            2683 \n",
      "  smart_240_raw                        0           39339 \n",
      "  smart_241_raw                        0     51579811864 \n",
      "  smart_242_raw                     3912    417629711029 \n",
      "\n",
      "Range for    :                       min             max\n",
      "  smart_1_normalized                  79             120 \n",
      "  smart_3_normalized                  91             100 \n",
      "  smart_5_normalized                  67             100 \n",
      "  smart_7_normalized                  28             100 \n",
      "  smart_9_normalized                  55             100 \n",
      "  smart_183_normalized                 1             100 \n",
      "  smart_184_normalized                38             100 \n",
      "  smart_187_normalized                 1             100 \n",
      "  smart_189_normalized                 1             100 \n",
      "  smart_190_normalized                58              86 \n",
      "  smart_193_normalized                 1             100 \n",
      "  smart_194_normalized                14              42 \n",
      "  smart_197_normalized                 1             100 \n",
      "  smart_198_normalized                 1             100 \n",
      "\n"
     ]
    }
   ],
   "source": [
    "# First, let's remove columns that have no valid data\n",
    "df = df.dropna(axis='columns', how='all') \n",
    "# Next, let's remove a few rows that had invalid data (most likely error reading SMART statistics)\n",
    "df = df.dropna(axis='rows', how='any')\n",
    "\n",
    "df_failed = df[df['failure']==1]\n",
    "df_normal = df[df['failure']==0]\n",
    "\n",
    "col_list_raw = remove_constant_values(df_failed,\"_raw\")\n",
    "col_list_normalized = remove_constant_values(df_failed,\"_normalized\")\n",
    "col_list = ['date', 'serial_number', 'model', 'capacity_bytes', 'failure'] \n",
    "col_list = col_list + col_list_raw + col_list_normalized\n",
    "\n",
    "df = df[col_list]\n",
    "\n",
    "#keep the same columns as the training dataframe \n",
    "df_t = df_t[col_list]\n",
    "\n",
    "#fill in the small number of null values in remaining columns \n",
    "df = df.fillna(0)\n",
    "df_t = df_t.fillna(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# collecting the number of failed hard disks to compare the distributions of the sensor data with normal disks and also to plot time series data \n",
    "\n",
    "serial_num_list_failed = df_failed.serial_number.value_counts().index.tolist()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Histogram of sensor data \n",
    "Once the data is curated, a typical histogram plot of the data can reveal key operating points, outliers and provide further suggestions for preparing the data. The raw SMART sensor data is plotted below. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "21"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 25 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "n = 5 \n",
    "m = len(col_list_raw)//n \n",
    "if len(col_list_raw)%n > 0 : m += 1 \n",
    "    \n",
    "fig, axs = plt.subplots(figsize=(20,20), ncols=n, nrows=m)\n",
    "axs = axs.flatten()\n",
    "\n",
    "for i,cur_col in enumerate(col_list_raw): \n",
    "    axs[i].hist(df_normal[cur_col],alpha=0.2,color=\"tab:blue\")\n",
    "    axs[i].hist(df_failed[cur_col],alpha=0.6,color=\"tab:orange\")\n",
    "    axs[i].set_yscale('log')\n",
    "    axs[i].set_title(cur_col)\n",
    "\n",
    "for i in range(len(col_list_raw),axs.shape[0]): \n",
    "    axs[i].axis('off')\n",
    "\n",
    "i = len(col_list_raw)\n",
    "axs[i].plot([],[],alpha=0.2,color=\"tab:blue\",label=\"Normal disks\",linewidth=15)\n",
    "axs[i].plot([],[],alpha=0.6,color=\"tab:orange\",label=\"Failed disks\",linewidth=15)\n",
    "axs[i].legend(fontsize=16)\n",
    "    \n",
    "len(col_list_raw)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Time series visualization \n",
    "\n",
    "Of all the disks with available data, several disks fail. Since the SMART sensor data is time series, it is useful to visualize the sensor behavior for the failed disks. In the plots below, the raw sensor data are displayed for 10 failed disks out of the 1061 failed disks. This can give us insights to the behavior of the disk until the time of failure. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "NOTE : \n",
      "    The large circles represent time of failure.\n",
      "    The lines show the data leading up to the point of faiilure. \n",
      "    Only serial numbers of failed disks are being used.\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x2268 with 21 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from pandas.plotting import register_matplotlib_converters\n",
    "register_matplotlib_converters()\n",
    "\n",
    "\n",
    "n_rows = len(col_list_raw)\n",
    "fig, axs = plt.subplots(ncols=1, nrows=n_rows, \n",
    "                            figsize=(10,1.5*n_rows),sharex=True)\n",
    "for cur_num in serial_num_list_failed[:10]: # [:1]: \n",
    "    plot_df(df[df['serial_number']==cur_num],col_list_raw,axs)\n",
    "\n",
    "print(\"NOTE : \\n    The large circles represent time of failure.\\n    The lines show the data leading up to the point of faiilure. \\n    Only serial numbers of failed disks are being used.\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving the train and test data sets...\n",
      "Completed saving data sets\n"
     ]
    }
   ],
   "source": [
    "# Save datasets for future use\n",
    "print('Saving the train and test data sets...')\n",
    "df.to_pickle(data_dir+pkl_train_file)\n",
    "df_t.to_pickle(data_dir+pkl_test_file)\n",
    "print('Completed saving data sets')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"5\"></a>\n",
    "## Training GPU XGBoost models with RAPIDS for Time Series"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this section, we will leverage the Backblaze Hard Drive SMART data to train an XGBoost model that will predict potential future failures.\n",
    "\n",
    "XGBoost will use the historical collection of SMART data to learn common feature patterns present in disk failures.  The model we create will be able to predict a failure based on the characteristics learned. To implement our XGBoost model we will be using  NVIDIA's [RAPIDS](https://rapids.ai/) which is an open source, GPU accelerated, data science platform.  RAPIDS includes cuML which is a \"scikit-learn-like\" library that contains GPU-accelerated versions of some of the most popular machine learning algorithms.\n",
    "\n",
    "In this lab, we will be using RAPIDS to train XGBoost models."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"6\"></a>\n",
    "### XGBoost\n",
    "\n",
    "XGBoost is widely known today as a \"go to\" model when working with structured data. In fact, there is a disproportionately large number of XGBoost-based winning entries of Kaggle competitions. In this lab, we introduce the model, but we do not intend to go through a rigorous mathematical formulation of XGBoost. For curious readers, we recommend reviewing of the original paper on [XGBoost](https://www.kdd.org/kdd2016/papers/files/rfp0697-chenAemb.pdf)\n",
    "\n",
    "Traditionally, in tree-based ensemble methods such as Random Forests, we train each tree independently and the predictions of multiple trees are summed to obtain the final score. The figure below classifies whether someone would like to play computer games or not. The left and right trees are classifying gamers based on different features, namely age and whether samples use computers on a daily basis. A weighted combination of such trees would result in a more accurate decision.\n",
    "\n",
    "<br>\n",
    "\n",
    "<img src=\"img/twocart.png\" style=\"margin-top:10px;\"/>\n",
    "<p style=\"text-align: center;color:gray\"> Figure 1. Tree classifier to detect computer gamers<br/>\n",
    "<i>Image credit: XGBoost: A Scalable Tree Boosting System, T. Chen and C. Guestrin</i></p>\n",
    "\n",
    "\n",
    "\n",
    "In the case of XGBoost, an implementation of Gradient Boosted Decision Trees, we repeatedly build new models and combine them into an ensemble. Unlike Random Forests, we build trees one at a time, where each new tree helps to correct errors made by previously trained tree. The incremental training procedure is shown in figure 2.\n",
    "\n",
    "<br/><br/>\n",
    "\n",
    "\n",
    "\n",
    "<img src=\"img/Ensemble Tree.jpg\" style=\"margin-top:10px;\"/>\n",
    "<p style=\"text-align: center;color:gray\"> Figure 2. XGBoost training pipeline</p>\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "The loss function controls the predictive power of the model, and the XGBoost regularization term ensures simplicity and manages overfitting. Since \"Boosting\" focuses iteratively building learners for the most difficult parts of your data, it provides an efficient algorithm to deal with unbalanced datasets by strengthening the impact of the positive class. With this brief introduction to the method, we are going to further curate the data and implement the model.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"7\"></a>\n",
    "## More on Data Preparation\n",
    "\n",
    "Now that we have filtered the dataset based on the `model` column, this column is not required anymore and can be removed. Also, the `date` and `serial_number` do not possess any useful information for our training. Later in the course, where we treat the hard drive date as time sequences, we would need to further sort the data using the `date` column. Let's start by removing these unnecessary columns at this stage."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# RAPIDS\n",
    "import cudf \n",
    "import xgboost as xgb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dropping unneeded columns...\n"
     ]
    }
   ],
   "source": [
    "# drop unneccessary columns\n",
    "print(\"Dropping unneeded columns...\")\n",
    "df = df.drop(columns=['model', 'date', 'serial_number'])\n",
    "df_t = df_t.drop(columns=['model', 'date', 'serial_number'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For training/testing our model, we require labels. The `failure` column provides these class labels. We need to extract these labels as follows:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Separating Training Data Features and Labels\n",
      "Separating Test Data Features and Labels\n",
      "Completed\n"
     ]
    }
   ],
   "source": [
    "print(\"Separating Training Data Features and Labels\")\n",
    "# separate training set features from labels\n",
    "df_train_target = pd.DataFrame(df['failure'])\n",
    "\n",
    "print(\"Separating Test Data Features and Labels\")\n",
    "# separate test set features from labels\n",
    "df_test_target = pd.DataFrame(df_t['failure'])\n",
    "\n",
    "print(\"Completed\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Each SMART measure on the list of columns has a companion \"normalized\" twin column. The values of the normalized column are just the scaled version of the \"raw\" columns. We can choose to opt either of these columns for this lab, as XGBoost is not sensitive to the scale of the values of each column. Here, we choose to keep the \"raw\" version.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['capacity_bytes', 'failure', 'smart_1_raw', 'smart_4_raw',\n",
       "       'smart_5_raw', 'smart_7_raw', 'smart_9_raw', 'smart_12_raw',\n",
       "       'smart_183_raw', 'smart_184_raw', 'smart_187_raw', 'smart_188_raw',\n",
       "       'smart_189_raw', 'smart_190_raw', 'smart_192_raw', 'smart_193_raw',\n",
       "       'smart_194_raw', 'smart_197_raw', 'smart_198_raw', 'smart_199_raw',\n",
       "       'smart_240_raw', 'smart_241_raw', 'smart_242_raw'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cols = [c for c in df.columns if c.lower().find(\"normalized\")==-1]\n",
    "df=df[cols]\n",
    "df.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"e1\"></a>\n",
    "## Exercise 1: Remove \"zero-valued\" rows\n",
    "\n",
    "Another important drawback with the data is \"zero-valued\" rows. These rows have zero values for every entry of the dataset and do not possess any useful information for training the model. We can remove these rows from the training sets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['capacity_bytes', 'failure', 'smart_1_raw', 'smart_4_raw',\n",
       "       'smart_5_raw', 'smart_7_raw', 'smart_9_raw', 'smart_12_raw',\n",
       "       'smart_183_raw', 'smart_184_raw', 'smart_187_raw', 'smart_188_raw',\n",
       "       'smart_189_raw', 'smart_190_raw', 'smart_192_raw', 'smart_193_raw',\n",
       "       'smart_194_raw', 'smart_197_raw', 'smart_198_raw', 'smart_199_raw',\n",
       "       'smart_240_raw', 'smart_241_raw', 'smart_242_raw'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#remove zero valued rows\n",
    "# df = <<TO DO>>\n",
    "df = df.loc[:, (df != 0).any(axis=0)]\n",
    "df.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If you get stuck, click [here](#a1) for an answer."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that these columns are removed, we should apply the same to our test set. Note that for our test set, even if they contain non-zero values, since they are removed from the train set, they will not be useful. Also, we need to remove the `failure` column - which is our class labels! - from the training and testing sets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Completed\n"
     ]
    }
   ],
   "source": [
    "cols = [c for c in df_t.columns if c.lower() in df.columns]\n",
    "df_t=df_t[cols]\n",
    "\n",
    "df_train = df.drop(columns=['failure'])\n",
    "df_test = df_t.drop(columns=['failure'])\n",
    "print(\"Completed\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's look at a few records from the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>capacity_bytes</th>\n",
       "      <th>failure</th>\n",
       "      <th>smart_1_raw</th>\n",
       "      <th>smart_4_raw</th>\n",
       "      <th>smart_5_raw</th>\n",
       "      <th>smart_7_raw</th>\n",
       "      <th>smart_9_raw</th>\n",
       "      <th>smart_12_raw</th>\n",
       "      <th>smart_183_raw</th>\n",
       "      <th>smart_184_raw</th>\n",
       "      <th>...</th>\n",
       "      <th>smart_190_raw</th>\n",
       "      <th>smart_192_raw</th>\n",
       "      <th>smart_193_raw</th>\n",
       "      <th>smart_194_raw</th>\n",
       "      <th>smart_197_raw</th>\n",
       "      <th>smart_198_raw</th>\n",
       "      <th>smart_199_raw</th>\n",
       "      <th>smart_240_raw</th>\n",
       "      <th>smart_241_raw</th>\n",
       "      <th>smart_242_raw</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4000787030016</td>\n",
       "      <td>0</td>\n",
       "      <td>124084032.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>407077929.0</td>\n",
       "      <td>9938.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>24.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>33920.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9690.0</td>\n",
       "      <td>3.179327e+10</td>\n",
       "      <td>8.596073e+09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>4000787030016</td>\n",
       "      <td>0</td>\n",
       "      <td>22057080.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>789891767.0</td>\n",
       "      <td>17785.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>18.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>70545.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>17310.0</td>\n",
       "      <td>2.169678e+10</td>\n",
       "      <td>1.860347e+11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>4000787030016</td>\n",
       "      <td>0</td>\n",
       "      <td>119535272.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>338259228.0</td>\n",
       "      <td>8600.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>29.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9086.0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8514.0</td>\n",
       "      <td>2.755353e+10</td>\n",
       "      <td>1.631456e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>4000787030016</td>\n",
       "      <td>0</td>\n",
       "      <td>90592144.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>64202331.0</td>\n",
       "      <td>18105.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>4466.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>21.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>14150.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>564.0</td>\n",
       "      <td>18051.0</td>\n",
       "      <td>2.150714e+10</td>\n",
       "      <td>2.376942e+11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>4000787030016</td>\n",
       "      <td>0</td>\n",
       "      <td>240109536.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>664498121.0</td>\n",
       "      <td>14657.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>25.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>46665.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>14303.0</td>\n",
       "      <td>1.645696e+10</td>\n",
       "      <td>7.311286e+10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    capacity_bytes  failure  smart_1_raw  smart_4_raw  smart_5_raw  \\\n",
       "4    4000787030016        0  124084032.0          8.0          0.0   \n",
       "7    4000787030016        0   22057080.0         21.0          0.0   \n",
       "8    4000787030016        0  119535272.0          7.0          0.0   \n",
       "12   4000787030016        0   90592144.0         12.0          0.0   \n",
       "13   4000787030016        0  240109536.0         20.0          0.0   \n",
       "\n",
       "    smart_7_raw  smart_9_raw  smart_12_raw  smart_183_raw  smart_184_raw  ...  \\\n",
       "4   407077929.0       9938.0           8.0            0.0            0.0  ...   \n",
       "7   789891767.0      17785.0          21.0            0.0            0.0  ...   \n",
       "8   338259228.0       8600.0           7.0            0.0            0.0  ...   \n",
       "12   64202331.0      18105.0          12.0         4466.0            0.0  ...   \n",
       "13  664498121.0      14657.0          20.0            0.0            0.0  ...   \n",
       "\n",
       "    smart_190_raw  smart_192_raw  smart_193_raw  smart_194_raw  smart_197_raw  \\\n",
       "4            24.0            0.0        33920.0           24.0            0.0   \n",
       "7            18.0            0.0        70545.0           18.0            0.0   \n",
       "8            29.0            0.0         9086.0           29.0            0.0   \n",
       "12           21.0            0.0        14150.0           21.0            0.0   \n",
       "13           25.0            0.0        46665.0           25.0            0.0   \n",
       "\n",
       "    smart_198_raw  smart_199_raw  smart_240_raw  smart_241_raw  smart_242_raw  \n",
       "4             0.0            0.0         9690.0   3.179327e+10   8.596073e+09  \n",
       "7             0.0            0.0        17310.0   2.169678e+10   1.860347e+11  \n",
       "8             0.0            0.0         8514.0   2.755353e+10   1.631456e+10  \n",
       "12            0.0          564.0        18051.0   2.150714e+10   2.376942e+11  \n",
       "13            0.0            0.0        14303.0   1.645696e+10   7.311286e+10  \n",
       "\n",
       "[5 rows x 23 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"8\"></a>\n",
    "## Define XGBoost Model Parameters"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Similar to other machine learning models, there are hyperparameters to be set for the XGBoost. Below is the list of parameters of the model:\n",
    "\n",
    "- <b>max_tree_depth</b>:  Defines the maximum depth of the tree.  Deeper trees are more complex and able to learn more relations amongst the data, however, also tend to over-fit as the relations may be specific to the training set.\n",
    "\n",
    "- <b>tree_method</b>:  Used to control GPU vs CPU operation.  'hist' = CPU histogram method.  'gpu_hist' = GPU histogram method.\n",
    "\n",
    "- <b>subsample</b>:  Denotes the fraction of observations to be randomly sampled for each tree.  Lower values make the algorithm more conservative and prevent overfitting however, too small of a value might lead to underfitting.\n",
    "\n",
    "- <b>regularization</b>:  Applies an L1-regularization term to the weights.  Increasing this value makes the model more conservative and can speed up performance on very high-dimensional data sets.\n",
    "\n",
    "- <b>gamma</b>:  Gamma specifies the minimum loss reduction required to split a node.\n",
    "\n",
    "- <b>pos_weight</b>:  Controls the balance of positive and negative weights to deal with unbalanced classes.\n",
    "\n",
    "- <b>early_stop</b>:  Used to control over-fitting.  Attempts to find the inflection point when the performance on the test set starts to decrease while the performance on the training set continues to increase."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "MAX_TREE_DEPTH = 8\n",
    "TREE_METHOD = 'hist'\n",
    "ITERATIONS = 85\n",
    "SUBSAMPLE = 0.6\n",
    "REGULARIZATION = 1.3\n",
    "GAMMA = 0.3\n",
    "POS_WEIGHT = 1\n",
    "EARLY_STOP = 10"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"9\"></a>\n",
    "## Train XGBoost on CPUs\n",
    "\n",
    "To demonstrate the benefits of GPU computing and the RAPIDS platform, we will first train a XGBoost model on the CPUs in our training enviroment. Let's take a look at the specs of the CPUs running on your cloud instance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model name:            Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz\n",
      "CPU(s):                48\n",
      "On-line CPU(s) list:   0-47\n",
      "NUMA node0 CPU(s):     0-47\n"
     ]
    }
   ],
   "source": [
    "# Show information on the CPUs\n",
    "!lscpu | grep 'Model name:'\n",
    "!lscpu | grep 'CPU(s)'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, we need to convert our data into XGBoost DMatrix format. DMatrix is optimized for memory efficiency and training speed in XGBoost. Next, we need to create a dictionary of hyperparameters discussed above and finally call the `train` method to start the training process. We also measure the total time to compare it with the GPU version."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Converting data sets into XGBoost DMatrix format...\n",
      "Completed converting data sets\n",
      "Starting CPU XGBoost Training...\n",
      "[05:28:12] WARNING: /conda/conda-bld/xgboost_1591204597597/work/src/learner.cc:480: \n",
      "Parameters: { silent } might not be used.\n",
      "\n",
      "  This may not be accurate due to some parameters are only used in language bindings but\n",
      "  passed down to XGBoost core.  Or some parameters are not used but slip through this\n",
      "  verification. Please open an issue if you find above cases.\n",
      "\n",
      "\n",
      "[0]\ttrain-rmse:0.47570\teval-rmse:0.47497\n",
      "Multiple eval metrics have been passed: 'eval-rmse' will be used for early stopping.\n",
      "\n",
      "Will train until eval-rmse hasn't improved in 10 rounds.\n",
      "[1]\ttrain-rmse:0.45096\teval-rmse:0.45146\n",
      "[2]\ttrain-rmse:0.42852\teval-rmse:0.42858\n",
      "[3]\ttrain-rmse:0.40691\teval-rmse:0.40738\n",
      "[4]\ttrain-rmse:0.38703\teval-rmse:0.38696\n",
      "[5]\ttrain-rmse:0.36739\teval-rmse:0.36762\n",
      "[6]\ttrain-rmse:0.34873\teval-rmse:0.34928\n",
      "[7]\ttrain-rmse:0.33146\teval-rmse:0.33169\n",
      "[8]\ttrain-rmse:0.31547\teval-rmse:0.31510\n",
      "[9]\ttrain-rmse:0.29973\teval-rmse:0.29952\n",
      "[10]\ttrain-rmse:0.28434\teval-rmse:0.28449\n",
      "[11]\ttrain-rmse:0.27039\teval-rmse:0.27032\n",
      "[12]\ttrain-rmse:0.25638\teval-rmse:0.25681\n",
      "[13]\ttrain-rmse:0.24413\teval-rmse:0.24402\n",
      "[14]\ttrain-rmse:0.23185\teval-rmse:0.23181\n",
      "[15]\ttrain-rmse:0.22039\teval-rmse:0.22012\n",
      "[16]\ttrain-rmse:0.20954\teval-rmse:0.20916\n",
      "[17]\ttrain-rmse:0.19872\teval-rmse:0.19879\n",
      "[18]\ttrain-rmse:0.18880\teval-rmse:0.18886\n",
      "[19]\ttrain-rmse:0.17960\teval-rmse:0.17938\n",
      "[20]\ttrain-rmse:0.17028\teval-rmse:0.17051\n",
      "[21]\ttrain-rmse:0.16212\teval-rmse:0.16189\n",
      "[22]\ttrain-rmse:0.15379\teval-rmse:0.15393\n",
      "[23]\ttrain-rmse:0.14628\teval-rmse:0.14622\n",
      "[24]\ttrain-rmse:0.13891\teval-rmse:0.13892\n",
      "[25]\ttrain-rmse:0.13201\teval-rmse:0.13202\n",
      "[26]\ttrain-rmse:0.12531\teval-rmse:0.12545\n",
      "[27]\ttrain-rmse:0.11926\teval-rmse:0.11921\n",
      "[28]\ttrain-rmse:0.11323\teval-rmse:0.11325\n",
      "[29]\ttrain-rmse:0.10752\teval-rmse:0.10765\n",
      "[30]\ttrain-rmse:0.10235\teval-rmse:0.10225\n",
      "[31]\ttrain-rmse:0.09719\teval-rmse:0.09722\n",
      "[32]\ttrain-rmse:0.09233\teval-rmse:0.09240\n",
      "[33]\ttrain-rmse:0.08794\teval-rmse:0.08782\n",
      "[34]\ttrain-rmse:0.08361\teval-rmse:0.08342\n",
      "[35]\ttrain-rmse:0.07934\teval-rmse:0.07929\n",
      "[36]\ttrain-rmse:0.07541\teval-rmse:0.07539\n",
      "[37]\ttrain-rmse:0.07177\teval-rmse:0.07166\n",
      "[38]\ttrain-rmse:0.06824\teval-rmse:0.06811\n",
      "[39]\ttrain-rmse:0.06488\teval-rmse:0.06477\n",
      "[40]\ttrain-rmse:0.06162\teval-rmse:0.06162\n",
      "[41]\ttrain-rmse:0.05858\teval-rmse:0.05858\n",
      "[42]\ttrain-rmse:0.05584\teval-rmse:0.05568\n",
      "[43]\ttrain-rmse:0.05308\teval-rmse:0.05299\n",
      "[44]\ttrain-rmse:0.05051\teval-rmse:0.05039\n",
      "[45]\ttrain-rmse:0.04801\teval-rmse:0.04793\n",
      "[46]\ttrain-rmse:0.04568\teval-rmse:0.04560\n",
      "[47]\ttrain-rmse:0.04355\teval-rmse:0.04343\n",
      "[48]\ttrain-rmse:0.04146\teval-rmse:0.04131\n",
      "[49]\ttrain-rmse:0.03948\teval-rmse:0.03933\n",
      "[50]\ttrain-rmse:0.03761\teval-rmse:0.03748\n",
      "[51]\ttrain-rmse:0.03581\teval-rmse:0.03568\n",
      "[52]\ttrain-rmse:0.03418\teval-rmse:0.03400\n",
      "[53]\ttrain-rmse:0.03257\teval-rmse:0.03240\n",
      "[54]\ttrain-rmse:0.03101\teval-rmse:0.03089\n",
      "[55]\ttrain-rmse:0.02963\teval-rmse:0.02945\n",
      "[56]\ttrain-rmse:0.02820\teval-rmse:0.02811\n",
      "[57]\ttrain-rmse:0.02699\teval-rmse:0.02682\n",
      "[58]\ttrain-rmse:0.02577\teval-rmse:0.02561\n",
      "[59]\ttrain-rmse:0.02464\teval-rmse:0.02447\n",
      "[60]\ttrain-rmse:0.02356\teval-rmse:0.02338\n",
      "[61]\ttrain-rmse:0.02254\teval-rmse:0.02237\n",
      "[62]\ttrain-rmse:0.02155\teval-rmse:0.02139\n",
      "[63]\ttrain-rmse:0.02065\teval-rmse:0.02048\n",
      "[64]\ttrain-rmse:0.01984\teval-rmse:0.01963\n",
      "[65]\ttrain-rmse:0.01905\teval-rmse:0.01882\n",
      "[66]\ttrain-rmse:0.01826\teval-rmse:0.01807\n",
      "[67]\ttrain-rmse:0.01759\teval-rmse:0.01735\n",
      "[68]\ttrain-rmse:0.01692\teval-rmse:0.01668\n",
      "[69]\ttrain-rmse:0.01626\teval-rmse:0.01606\n",
      "[70]\ttrain-rmse:0.01569\teval-rmse:0.01547\n",
      "[71]\ttrain-rmse:0.01516\teval-rmse:0.01493\n",
      "[72]\ttrain-rmse:0.01465\teval-rmse:0.01441\n",
      "[73]\ttrain-rmse:0.01416\teval-rmse:0.01392\n",
      "[74]\ttrain-rmse:0.01372\teval-rmse:0.01347\n",
      "[75]\ttrain-rmse:0.01330\teval-rmse:0.01306\n",
      "[76]\ttrain-rmse:0.01294\teval-rmse:0.01266\n",
      "[77]\ttrain-rmse:0.01256\teval-rmse:0.01230\n",
      "[78]\ttrain-rmse:0.01224\teval-rmse:0.01197\n",
      "[79]\ttrain-rmse:0.01193\teval-rmse:0.01165\n",
      "[80]\ttrain-rmse:0.01166\teval-rmse:0.01136\n",
      "[81]\ttrain-rmse:0.01138\teval-rmse:0.01110\n",
      "[82]\ttrain-rmse:0.01115\teval-rmse:0.01084\n",
      "[83]\ttrain-rmse:0.01091\teval-rmse:0.01062\n",
      "[84]\ttrain-rmse:0.01072\teval-rmse:0.01041\n",
      "CPU XGBoost training compeleted - elapsed time: 59.92710471153259 seconds\n"
     ]
    }
   ],
   "source": [
    "start_time = time.time()\n",
    "\n",
    "print('Converting data sets into XGBoost DMatrix format...')\n",
    "xgtrain = xgb.DMatrix(df_train, df_train_target)\n",
    "xgeval = xgb.DMatrix(df_test, df_test_target)\n",
    "print('Completed converting data sets')\n",
    "\n",
    "params = {'tree_method': TREE_METHOD, 'max_depth': MAX_TREE_DEPTH, 'alpha': REGULARIZATION,\n",
    "          'gamma': GAMMA, 'subsample': SUBSAMPLE, 'scale_pos_weight': POS_WEIGHT, 'learning_rate': 0.05, 'silent': 1}\n",
    "\n",
    "print('Starting CPU XGBoost Training...')\n",
    "bst = xgb.train(params, xgtrain, ITERATIONS, evals=[(xgtrain, \"train\"), (xgeval, \"eval\")],\n",
    "                early_stopping_rounds=EARLY_STOP)\n",
    "\n",
    "timetaken_cpu = time.time() - start_time\n",
    "print('CPU XGBoost training compeleted - elapsed time:', timetaken_cpu,'seconds')\n",
    "\n",
    "# free up memory\n",
    "del xgtrain\n",
    "del xgeval\n",
    "del bst"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"10\"></a>\n",
    "## Training with GPU without cuDF DataFrame\n",
    "\n",
    "We will now leverage GPUs to accelerate the XGBoost Training and compare the results to the CPU version.\n",
    "\n",
    "First, let's see which GPU we have in our lab system."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fri Jul  1 05:30:25 2022       \n",
      "+-----------------------------------------------------------------------------+\n",
      "| NVIDIA-SMI 440.33.01    Driver Version: 440.33.01    CUDA Version: 10.2     |\n",
      "|-------------------------------+----------------------+----------------------+\n",
      "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
      "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
      "|===============================+======================+======================|\n",
      "|   0  Tesla T4            On   | 00000000:00:1B.0 Off |                    0 |\n",
      "| N/A   28C    P8     9W /  70W |     11MiB / 15109MiB |      0%      Default |\n",
      "+-------------------------------+----------------------+----------------------+\n",
      "|   1  Tesla T4            On   | 00000000:00:1C.0 Off |                    0 |\n",
      "| N/A   31C    P8     9W /  70W |     11MiB / 15109MiB |      0%      Default |\n",
      "+-------------------------------+----------------------+----------------------+\n",
      "|   2  Tesla T4            On   | 00000000:00:1D.0 Off |                    0 |\n",
      "| N/A   30C    P8     9W /  70W |     11MiB / 15109MiB |      0%      Default |\n",
      "+-------------------------------+----------------------+----------------------+\n",
      "|   3  Tesla T4            On   | 00000000:00:1E.0 Off |                    0 |\n",
      "| N/A   29C    P8     9W /  70W |     11MiB / 15109MiB |      0%      Default |\n",
      "+-------------------------------+----------------------+----------------------+\n",
      "                                                                               \n",
      "+-----------------------------------------------------------------------------+\n",
      "| Processes:                                                       GPU Memory |\n",
      "|  GPU       PID   Type   Process name                             Usage      |\n",
      "|=============================================================================|\n",
      "|  No running processes found                                                 |\n",
      "+-----------------------------------------------------------------------------+\n"
     ]
    }
   ],
   "source": [
    "!nvidia-smi"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To configure the XGBoost training algorithm to use GPUs, we set params[tree_method] = 'gpu_hist'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Converting data sets into XGBoost DMatrix format...\n",
      "Completed converting data sets\n",
      "Starting GPU XGBoost Training...\n",
      "[05:30:56] WARNING: /conda/conda-bld/xgboost_1591204597597/work/src/learner.cc:480: \n",
      "Parameters: { silent } might not be used.\n",
      "\n",
      "  This may not be accurate due to some parameters are only used in language bindings but\n",
      "  passed down to XGBoost core.  Or some parameters are not used but slip through this\n",
      "  verification. Please open an issue if you find above cases.\n",
      "\n",
      "\n",
      "[0]\ttrain-rmse:0.47501\teval-rmse:0.47501\n",
      "Multiple eval metrics have been passed: 'eval-rmse' will be used for early stopping.\n",
      "\n",
      "Will train until eval-rmse hasn't improved in 10 rounds.\n",
      "[1]\ttrain-rmse:0.45127\teval-rmse:0.45126\n",
      "[2]\ttrain-rmse:0.42871\teval-rmse:0.42871\n",
      "[3]\ttrain-rmse:0.40729\teval-rmse:0.40728\n",
      "[4]\ttrain-rmse:0.38693\teval-rmse:0.38693\n",
      "[5]\ttrain-rmse:0.36760\teval-rmse:0.36759\n",
      "[6]\ttrain-rmse:0.34923\teval-rmse:0.34922\n",
      "[7]\ttrain-rmse:0.33178\teval-rmse:0.33177\n",
      "[8]\ttrain-rmse:0.31520\teval-rmse:0.31519\n",
      "[9]\ttrain-rmse:0.29945\teval-rmse:0.29944\n",
      "[10]\ttrain-rmse:0.28449\teval-rmse:0.28448\n",
      "[11]\ttrain-rmse:0.27028\teval-rmse:0.27027\n",
      "[12]\ttrain-rmse:0.25678\teval-rmse:0.25677\n",
      "[13]\ttrain-rmse:0.24396\teval-rmse:0.24395\n",
      "[14]\ttrain-rmse:0.23178\teval-rmse:0.23176\n",
      "[15]\ttrain-rmse:0.22021\teval-rmse:0.22019\n",
      "[16]\ttrain-rmse:0.20921\teval-rmse:0.20920\n",
      "[17]\ttrain-rmse:0.19877\teval-rmse:0.19875\n",
      "[18]\ttrain-rmse:0.18885\teval-rmse:0.18884\n",
      "[19]\ttrain-rmse:0.17943\teval-rmse:0.17941\n",
      "[20]\ttrain-rmse:0.17048\teval-rmse:0.17046\n",
      "[21]\ttrain-rmse:0.16198\teval-rmse:0.16196\n",
      "[22]\ttrain-rmse:0.15390\teval-rmse:0.15388\n",
      "[23]\ttrain-rmse:0.14623\teval-rmse:0.14621\n",
      "[24]\ttrain-rmse:0.13895\teval-rmse:0.13893\n",
      "[25]\ttrain-rmse:0.13203\teval-rmse:0.13201\n",
      "[26]\ttrain-rmse:0.12546\teval-rmse:0.12543\n",
      "[27]\ttrain-rmse:0.11921\teval-rmse:0.11919\n",
      "[28]\ttrain-rmse:0.11329\teval-rmse:0.11326\n",
      "[29]\ttrain-rmse:0.10766\teval-rmse:0.10763\n",
      "[30]\ttrain-rmse:0.10231\teval-rmse:0.10228\n",
      "[31]\ttrain-rmse:0.09723\teval-rmse:0.09720\n",
      "[32]\ttrain-rmse:0.09241\teval-rmse:0.09237\n",
      "[33]\ttrain-rmse:0.08783\teval-rmse:0.08779\n",
      "[34]\ttrain-rmse:0.08348\teval-rmse:0.08344\n",
      "[35]\ttrain-rmse:0.07935\teval-rmse:0.07931\n",
      "[36]\ttrain-rmse:0.07543\teval-rmse:0.07539\n",
      "[37]\ttrain-rmse:0.07171\teval-rmse:0.07167\n",
      "[38]\ttrain-rmse:0.06818\teval-rmse:0.06814\n",
      "[39]\ttrain-rmse:0.06483\teval-rmse:0.06478\n",
      "[40]\ttrain-rmse:0.06164\teval-rmse:0.06159\n",
      "[41]\ttrain-rmse:0.05862\teval-rmse:0.05857\n",
      "[42]\ttrain-rmse:0.05576\teval-rmse:0.05570\n",
      "[43]\ttrain-rmse:0.05304\teval-rmse:0.05298\n",
      "[44]\ttrain-rmse:0.05046\teval-rmse:0.05040\n",
      "[45]\ttrain-rmse:0.04801\teval-rmse:0.04795\n",
      "[46]\ttrain-rmse:0.04569\teval-rmse:0.04562\n",
      "[47]\ttrain-rmse:0.04349\teval-rmse:0.04342\n",
      "[48]\ttrain-rmse:0.04140\teval-rmse:0.04133\n",
      "[49]\ttrain-rmse:0.03943\teval-rmse:0.03935\n",
      "[50]\ttrain-rmse:0.03755\teval-rmse:0.03747\n",
      "[51]\ttrain-rmse:0.03578\teval-rmse:0.03569\n",
      "[52]\ttrain-rmse:0.03410\teval-rmse:0.03401\n",
      "[53]\ttrain-rmse:0.03250\teval-rmse:0.03241\n",
      "[54]\ttrain-rmse:0.03100\teval-rmse:0.03090\n",
      "[55]\ttrain-rmse:0.02957\teval-rmse:0.02947\n",
      "[56]\ttrain-rmse:0.02822\teval-rmse:0.02811\n",
      "[57]\ttrain-rmse:0.02694\teval-rmse:0.02683\n",
      "[58]\ttrain-rmse:0.02574\teval-rmse:0.02562\n",
      "[59]\ttrain-rmse:0.02460\teval-rmse:0.02448\n",
      "[60]\ttrain-rmse:0.02353\teval-rmse:0.02339\n",
      "[61]\ttrain-rmse:0.02251\teval-rmse:0.02237\n",
      "[62]\ttrain-rmse:0.02156\teval-rmse:0.02141\n",
      "[63]\ttrain-rmse:0.02065\teval-rmse:0.02050\n",
      "[64]\ttrain-rmse:0.01981\teval-rmse:0.01965\n",
      "[65]\ttrain-rmse:0.01901\teval-rmse:0.01885\n",
      "[66]\ttrain-rmse:0.01826\teval-rmse:0.01809\n",
      "[67]\ttrain-rmse:0.01756\teval-rmse:0.01738\n",
      "[68]\ttrain-rmse:0.01689\teval-rmse:0.01671\n",
      "[69]\ttrain-rmse:0.01627\teval-rmse:0.01608\n",
      "[70]\ttrain-rmse:0.01569\teval-rmse:0.01550\n",
      "[71]\ttrain-rmse:0.01515\teval-rmse:0.01495\n",
      "[72]\ttrain-rmse:0.01465\teval-rmse:0.01443\n",
      "[73]\ttrain-rmse:0.01417\teval-rmse:0.01395\n",
      "[74]\ttrain-rmse:0.01373\teval-rmse:0.01350\n",
      "[75]\ttrain-rmse:0.01332\teval-rmse:0.01309\n",
      "[76]\ttrain-rmse:0.01294\teval-rmse:0.01270\n",
      "[77]\ttrain-rmse:0.01259\teval-rmse:0.01234\n",
      "[78]\ttrain-rmse:0.01226\teval-rmse:0.01200\n",
      "[79]\ttrain-rmse:0.01195\teval-rmse:0.01169\n",
      "[80]\ttrain-rmse:0.01167\teval-rmse:0.01140\n",
      "[81]\ttrain-rmse:0.01141\teval-rmse:0.01114\n",
      "[82]\ttrain-rmse:0.01117\teval-rmse:0.01089\n",
      "[83]\ttrain-rmse:0.01095\teval-rmse:0.01066\n",
      "[84]\ttrain-rmse:0.01075\teval-rmse:0.01046\n",
      "GPU XGBoost training compeleted - elapsed time: 20.977258920669556 seconds\n"
     ]
    }
   ],
   "source": [
    "# GPU, without using cuDF\n",
    "\n",
    "start_time = time.time()\n",
    "\n",
    "print('Converting data sets into XGBoost DMatrix format...')\n",
    "xgtrain = xgb.DMatrix(df_train, df_train_target)\n",
    "xgeval = xgb.DMatrix(df_test, df_test_target)\n",
    "print('Completed converting data sets')\n",
    "\n",
    "params = {'tree_method': \"gpu_\"+TREE_METHOD, 'max_depth': MAX_TREE_DEPTH, 'alpha': REGULARIZATION,\n",
    "          'gamma': GAMMA, 'subsample': SUBSAMPLE, 'scale_pos_weight': POS_WEIGHT, 'learning_rate': 0.05, 'silent': 1}\n",
    "\n",
    "print('Starting GPU XGBoost Training...')\n",
    "bst = xgb.train(params, xgtrain, ITERATIONS, evals=[(xgtrain, \"train\"), (xgeval, \"eval\")],\n",
    "                early_stopping_rounds=EARLY_STOP)\n",
    "\n",
    "timetaken_gpu_nocudf = time.time() - start_time\n",
    "print('GPU XGBoost training compeleted - elapsed time:', timetaken_gpu_nocudf,'seconds')\n",
    "\n",
    "# free up memory\n",
    "del xgtrain\n",
    "del xgeval\n",
    "del bst"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"11\"></a>\n",
    "## Training with GPU and cuDF\n",
    "\n",
    "In this section we will be using cuDF instead of pandas DataFrame. cuDF is a GPU DataFrame library for loading, joining, aggregating, filtering, and otherwise manipulating data, which is based on Apache Arrow columnar memory format. cuDF provides a pandas-like API for familiarity which allows us to easily accelerate workflows without going into the details of CUDA programming. For more information on cuDF, please refer to [RAPIDS Github Repositor](https://github.com/rapidsai/cudf).\n",
    "To use cuDF with the XGBoost algorithm, we load the Pandas Dataframe into a cuDF Dataframe (Python object type cudf.dataframe.dataframe.DataFrame) first.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load into cuDF Dataframe\n",
    "\n",
    "gdf_train = cudf.DataFrame.from_pandas(df_train)\n",
    "gdf_train_target = cudf.DataFrame.from_pandas(df_train_target)\n",
    "\n",
    "gdf_eval = cudf.DataFrame.from_pandas(df_test)\n",
    "gdf_eval_target = cudf.DataFrame.from_pandas(df_test_target)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, we are going to measure the XGBoost training time using the created cuDF and on GPU."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Converting cuDF data sets into XGBoost DMatrix format...\n",
      "Completed converting data sets\n",
      "Starting GPU XGBoost Training with cuDF Dataframes...\n",
      "[05:32:26] WARNING: /conda/conda-bld/xgboost_1591204597597/work/src/learner.cc:480: \n",
      "Parameters: { silent } might not be used.\n",
      "\n",
      "  This may not be accurate due to some parameters are only used in language bindings but\n",
      "  passed down to XGBoost core.  Or some parameters are not used but slip through this\n",
      "  verification. Please open an issue if you find above cases.\n",
      "\n",
      "\n",
      "[0]\ttrain-rmse:0.47501\teval-rmse:0.47501\n",
      "Multiple eval metrics have been passed: 'eval-rmse' will be used for early stopping.\n",
      "\n",
      "Will train until eval-rmse hasn't improved in 10 rounds.\n",
      "[1]\ttrain-rmse:0.45127\teval-rmse:0.45126\n",
      "[2]\ttrain-rmse:0.42871\teval-rmse:0.42871\n",
      "[3]\ttrain-rmse:0.40729\teval-rmse:0.40728\n",
      "[4]\ttrain-rmse:0.38693\teval-rmse:0.38693\n",
      "[5]\ttrain-rmse:0.36760\teval-rmse:0.36759\n",
      "[6]\ttrain-rmse:0.34923\teval-rmse:0.34922\n",
      "[7]\ttrain-rmse:0.33178\teval-rmse:0.33177\n",
      "[8]\ttrain-rmse:0.31520\teval-rmse:0.31519\n",
      "[9]\ttrain-rmse:0.29945\teval-rmse:0.29944\n",
      "[10]\ttrain-rmse:0.28449\teval-rmse:0.28448\n",
      "[11]\ttrain-rmse:0.27028\teval-rmse:0.27027\n",
      "[12]\ttrain-rmse:0.25678\teval-rmse:0.25677\n",
      "[13]\ttrain-rmse:0.24396\teval-rmse:0.24395\n",
      "[14]\ttrain-rmse:0.23178\teval-rmse:0.23176\n",
      "[15]\ttrain-rmse:0.22021\teval-rmse:0.22019\n",
      "[16]\ttrain-rmse:0.20921\teval-rmse:0.20920\n",
      "[17]\ttrain-rmse:0.19877\teval-rmse:0.19875\n",
      "[18]\ttrain-rmse:0.18885\teval-rmse:0.18884\n",
      "[19]\ttrain-rmse:0.17943\teval-rmse:0.17941\n",
      "[20]\ttrain-rmse:0.17048\teval-rmse:0.17046\n",
      "[21]\ttrain-rmse:0.16198\teval-rmse:0.16196\n",
      "[22]\ttrain-rmse:0.15390\teval-rmse:0.15388\n",
      "[23]\ttrain-rmse:0.14623\teval-rmse:0.14621\n",
      "[24]\ttrain-rmse:0.13895\teval-rmse:0.13893\n",
      "[25]\ttrain-rmse:0.13203\teval-rmse:0.13201\n",
      "[26]\ttrain-rmse:0.12546\teval-rmse:0.12543\n",
      "[27]\ttrain-rmse:0.11921\teval-rmse:0.11919\n",
      "[28]\ttrain-rmse:0.11329\teval-rmse:0.11326\n",
      "[29]\ttrain-rmse:0.10766\teval-rmse:0.10763\n",
      "[30]\ttrain-rmse:0.10231\teval-rmse:0.10228\n",
      "[31]\ttrain-rmse:0.09723\teval-rmse:0.09720\n",
      "[32]\ttrain-rmse:0.09241\teval-rmse:0.09237\n",
      "[33]\ttrain-rmse:0.08783\teval-rmse:0.08779\n",
      "[34]\ttrain-rmse:0.08348\teval-rmse:0.08344\n",
      "[35]\ttrain-rmse:0.07935\teval-rmse:0.07931\n",
      "[36]\ttrain-rmse:0.07543\teval-rmse:0.07539\n",
      "[37]\ttrain-rmse:0.07171\teval-rmse:0.07167\n",
      "[38]\ttrain-rmse:0.06818\teval-rmse:0.06814\n",
      "[39]\ttrain-rmse:0.06483\teval-rmse:0.06478\n",
      "[40]\ttrain-rmse:0.06164\teval-rmse:0.06159\n",
      "[41]\ttrain-rmse:0.05862\teval-rmse:0.05857\n",
      "[42]\ttrain-rmse:0.05576\teval-rmse:0.05570\n",
      "[43]\ttrain-rmse:0.05304\teval-rmse:0.05298\n",
      "[44]\ttrain-rmse:0.05046\teval-rmse:0.05040\n",
      "[45]\ttrain-rmse:0.04801\teval-rmse:0.04795\n",
      "[46]\ttrain-rmse:0.04569\teval-rmse:0.04562\n",
      "[47]\ttrain-rmse:0.04349\teval-rmse:0.04342\n",
      "[48]\ttrain-rmse:0.04140\teval-rmse:0.04133\n",
      "[49]\ttrain-rmse:0.03943\teval-rmse:0.03935\n",
      "[50]\ttrain-rmse:0.03755\teval-rmse:0.03747\n",
      "[51]\ttrain-rmse:0.03578\teval-rmse:0.03569\n",
      "[52]\ttrain-rmse:0.03410\teval-rmse:0.03401\n",
      "[53]\ttrain-rmse:0.03250\teval-rmse:0.03241\n",
      "[54]\ttrain-rmse:0.03100\teval-rmse:0.03090\n",
      "[55]\ttrain-rmse:0.02957\teval-rmse:0.02947\n",
      "[56]\ttrain-rmse:0.02822\teval-rmse:0.02811\n",
      "[57]\ttrain-rmse:0.02694\teval-rmse:0.02683\n",
      "[58]\ttrain-rmse:0.02574\teval-rmse:0.02562\n",
      "[59]\ttrain-rmse:0.02460\teval-rmse:0.02448\n",
      "[60]\ttrain-rmse:0.02353\teval-rmse:0.02339\n",
      "[61]\ttrain-rmse:0.02251\teval-rmse:0.02237\n",
      "[62]\ttrain-rmse:0.02156\teval-rmse:0.02141\n",
      "[63]\ttrain-rmse:0.02065\teval-rmse:0.02050\n",
      "[64]\ttrain-rmse:0.01981\teval-rmse:0.01965\n",
      "[65]\ttrain-rmse:0.01901\teval-rmse:0.01885\n",
      "[66]\ttrain-rmse:0.01826\teval-rmse:0.01809\n",
      "[67]\ttrain-rmse:0.01756\teval-rmse:0.01738\n",
      "[68]\ttrain-rmse:0.01689\teval-rmse:0.01671\n",
      "[69]\ttrain-rmse:0.01627\teval-rmse:0.01608\n",
      "[70]\ttrain-rmse:0.01569\teval-rmse:0.01550\n",
      "[71]\ttrain-rmse:0.01515\teval-rmse:0.01495\n",
      "[72]\ttrain-rmse:0.01465\teval-rmse:0.01443\n",
      "[73]\ttrain-rmse:0.01417\teval-rmse:0.01395\n",
      "[74]\ttrain-rmse:0.01373\teval-rmse:0.01350\n",
      "[75]\ttrain-rmse:0.01332\teval-rmse:0.01309\n",
      "[76]\ttrain-rmse:0.01294\teval-rmse:0.01270\n",
      "[77]\ttrain-rmse:0.01259\teval-rmse:0.01234\n",
      "[78]\ttrain-rmse:0.01226\teval-rmse:0.01200\n",
      "[79]\ttrain-rmse:0.01195\teval-rmse:0.01169\n",
      "[80]\ttrain-rmse:0.01167\teval-rmse:0.01140\n",
      "[81]\ttrain-rmse:0.01141\teval-rmse:0.01114\n",
      "[82]\ttrain-rmse:0.01117\teval-rmse:0.01089\n",
      "[83]\ttrain-rmse:0.01095\teval-rmse:0.01066\n",
      "[84]\ttrain-rmse:0.01075\teval-rmse:0.01046\n",
      "GPU XGBoost Training with cuDF Dataframes compeleted - elapsed time: 13.423369646072388 seconds\n"
     ]
    }
   ],
   "source": [
    "# GPU, with cuDF\n",
    "\n",
    "start_time = time.time()\n",
    "\n",
    "print('Converting cuDF data sets into XGBoost DMatrix format...')\n",
    "xgtrain = xgb.DMatrix(gdf_train, gdf_train_target)\n",
    "xgeval = xgb.DMatrix(gdf_eval, gdf_eval_target)\n",
    "print('Completed converting data sets')\n",
    "\n",
    "params = {'tree_method': \"gpu_\"+TREE_METHOD, 'max_depth': MAX_TREE_DEPTH, 'alpha': REGULARIZATION,\n",
    "          'gamma': GAMMA, 'subsample': SUBSAMPLE, 'scale_pos_weight': POS_WEIGHT, 'learning_rate': 0.05, 'silent': 1}\n",
    "\n",
    "print('Starting GPU XGBoost Training with cuDF Dataframes...')\n",
    "bst = xgb.train(params, xgtrain, ITERATIONS, evals=[(xgtrain, \"train\"), (xgeval, \"eval\")],\n",
    "                early_stopping_rounds=EARLY_STOP)\n",
    "\n",
    "timetaken_gpu = time.time() - start_time\n",
    "print('GPU XGBoost Training with cuDF Dataframes compeleted - elapsed time:', timetaken_gpu,'seconds')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's visualize the training time for the three configurations described above. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU Time Taken:\n",
      " 59.9 seconds\n",
      "\n",
      "GPU (no cuDF) Time Taken:\n",
      " 21.0 seconds\n",
      "\n",
      "GPU (cuDF) Time Taken:\n",
      " 13.4 seconds\n",
      "\n",
      "Total speed-up with RAPIDS:\n",
      " 446.4 %\n"
     ]
    }
   ],
   "source": [
    "labels = 'CPU', 'GPU', 'GPU+cuDF'\n",
    "sizes = [timetaken_cpu, timetaken_gpu_nocudf, timetaken_gpu]\n",
    "\n",
    "plt.style.use('ggplot')\n",
    "\n",
    "x_pos = [i for i, _ in enumerate(labels)]\n",
    "plt.bar(x_pos, sizes)\n",
    "plt.xlabel(\"XGBoost Version\")\n",
    "plt.ylabel(\"Time (secs)\")\n",
    "plt.title(\"Training Time\")\n",
    "plt.xticks(x_pos, labels)\n",
    "\n",
    "plt.show()\n",
    "\n",
    "print(\"CPU Time Taken:\\n\", round(timetaken_cpu,1),\"seconds\")\n",
    "print(\"\\nGPU (no cuDF) Time Taken:\\n\", round(timetaken_gpu_nocudf,1),\"seconds\")\n",
    "print(\"\\nGPU (cuDF) Time Taken:\\n\", round(timetaken_gpu,1),\"seconds\")\n",
    "print(\"\\nTotal speed-up with RAPIDS:\\n\", round(timetaken_cpu/timetaken_gpu*100,1), \"%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As you can see, using the GPU + cuDF we see a significant speed-up. In the next section, we are going to explore the accuracy of the XGBoost model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"12\"></a>\n",
    "### XGBoost Model Accuracy\n",
    "\n",
    "Let's look at the model's accuracy on the evalutation set that we will use to compare against our Deep Learning models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting XGBoost prediction on test dataset...\n",
      "XGBoost prediction completed\n"
     ]
    }
   ],
   "source": [
    "print('Starting XGBoost prediction on test dataset...')\n",
    "\n",
    "# Use trained model to predict on test data set\n",
    "preds = bst.predict(xgeval)\n",
    "\n",
    "# Convert predictions to \"normal\" (0) or \"failed\"\n",
    "y_pred = []\n",
    "\n",
    "THRESHOLD = 0.5\n",
    "for pred in preds:\n",
    "    if pred<=THRESHOLD:\n",
    "        y_pred.append(0)\n",
    "    if pred>THRESHOLD:\n",
    "        y_pred.append(1)\n",
    "\n",
    "# Output array of classifications        \n",
    "y_pred = np.asarray(y_pred)\n",
    "        \n",
    "y_true = df_test_target.values.reshape(len(preds))\n",
    "\n",
    "print('XGBoost prediction completed')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy (Eval) 0.99993\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy (Eval)\", round(accuracy_score(y_true, y_pred),5))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"13\"></a>\n",
    "## Measuring Model Accuracy\n",
    "\n",
    "In this section, we are going to plot the confusion matrix. While accuracy rate provides us with an overall outlook on how well our model is performing, we need more insights into the accuracy. Especially, since our dataset is small, and we need to compromise some aspects of accuracy rate in favor of others. The `classification_report` provides us with the accuracy stats broken down into `precision`, `recall`, `f1-score` and `support`. Before introducing these measures, let's define some abbreviations:\n",
    "\n",
    "For brevity we are using the following abbreviations: \n",
    "\n",
    "\n",
    "- __True Positive (TP)__:  test correctly classifies the input to the class (hard drive failed, and we classified it as fail)\n",
    "- __False Positive (FP)__:  test incorrectly classifies the input to the class (hard drive is normal, but we classified it as fail)\n",
    "- __False Negative (FN)__:  test incorrectly misses to classify the input to the class (hard drive failed but we classified as normal)\n",
    "- __True Negative (TN)__:  test correctly misses to classify the input to the class (hard drive is normal, and we classified as normal)\n",
    "\n",
    "The first measure is focused on identifying positive cases and is called __recall__. We define recall as the ability of the model to identify all true positive samples of the dataset. In mathematical terms, recall is the ratio of true positives over true positives plus false negatives. By other means, recall tells us, among all the test samples belonging to the output class, how many of them are identified correctly by the model:\n",
    "\n",
    "\\begin{equation*}\n",
    "recall = \\frac{TP}{TP+FN}\n",
    "\\end{equation*}\n",
    "\n",
    "The next measure, is called __precision__ and is the ability of the model to identify the relevant samples only, and is defined as the ratio of true positives over true positives plus false positives:\n",
    "\n",
    "\\begin{equation*}\n",
    "precision = \\frac{TP}{TP+FP}\n",
    "\\end{equation*}\n",
    "\n",
    "\n",
    "Selecting a proper threshold, usually stems from a good balance between the precision and recall values. A well-known measure that provides such a balance is `F1 score`, which is a harmonic mean of precision and recall, and defined as:\n",
    "\n",
    "\\begin{equation*}\n",
    "{F_1 \\: score} = 2*\\frac{precision*recall}{precision+recall}\n",
    "\\end{equation*}\n",
    "\n",
    "F1 Score is generally considered a better overall measurement than accuracy when an uneven class distribution exists (large number of True Negatives) as in the case of predicting a hard drive failure.\n",
    "\n",
    "One very important measure is the recall rate on defective samples. We are interested in models that retrieve as many defective hard drives as possible. We try to predict at least half of the defective hard drives. Let's measure how our model is performing:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "      normal       1.00      1.00      1.00   3196318\n",
      "        fail       0.78      0.03      0.06       234\n",
      "\n",
      "    accuracy                           1.00   3196552\n",
      "   macro avg       0.89      0.51      0.53   3196552\n",
      "weighted avg       1.00      1.00      1.00   3196552\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_true, y_pred, target_names=[\"normal\", \"fail\"]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"14\"></a>\n",
    "## Interpreting the results\n",
    "\n",
    "When trying to predict hard drive failure, we can interpret the classification report above in this way:\n",
    "\n",
    "- Precision is the percentage of times the model is correct when predicting the given class.\n",
    "- Recall is the percentage of times we predicted that class against the total number of instances of the class in the data set.\n",
    "\n",
    "Therefore, the classification report above is telling us that the model predicted a failure for ~67% of the 234 fail sequences, and was correct 3% of the time. As for normal hard drives the recall rate is 1.00 and we are accurate ~100% of the time. \n",
    "\n",
    "<b>note</b>:  your precision / recall percentages may vary slightly depending on the weights learned during training"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## <b>Group Discussion</b>\n",
    "\n",
    "What do the results above tell us and any ideas how to improve?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Summary of training data:\n",
      "- Number of disks :  12237697\n",
      "- Number of failed disks:  1061\n",
      "- Percentage of failed disks: 0.0087 %\n"
     ]
    }
   ],
   "source": [
    "print('Summary of training data:')\n",
    "train_all_count = df.shape[0]\n",
    "train_fail_count = df[df['failure'] >= 1].shape[0]\n",
    "print ('- Number of disks : ', train_all_count)\n",
    "print ('- Number of failed disks: ', train_fail_count)\n",
    "print ('- Percentage of failed disks: %.4f' %(train_fail_count/train_all_count*100),'%' )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Summary of test data:\n",
      "- Number of disks :  3196552\n",
      "- Number of failed disks:  234\n",
      "- Percentage of failed disks: 0.0073 %\n"
     ]
    }
   ],
   "source": [
    "print('Summary of test data:')\n",
    "test_all_count = df_t.shape[0]\n",
    "test_fail_count = df_t[df_t['failure'] >= 1].shape[0]\n",
    "print ('- Number of disks : ', test_all_count)\n",
    "print ('- Number of failed disks: ', test_fail_count)\n",
    "print ('- Percentage of failed disks: %.4f' %(test_fail_count/test_all_count*100),'%' )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Our training and test sets have large class imbalances between normal and failed disks.  During the training process, our model is trying to find the optimal function for the lowest error loss.  Naively, if the model classifies every hard drive as normal, then the model is right more than 99% of the time.\n",
    "\n",
    "Thus, we need to come up with a strategy to address this imbalance. Some strategies to resolve class imbalance data include:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Up / Down Sample Classes\n",
    "- Weighted Class Cost Optimization\n",
    "- Evaluate various error metrics\n",
    "\n",
    "\n",
    "In the next section, we are going to resample the \"normal\" class to match the number of samples within both \"normal\" and \"failure\" classes of the dataset."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"15\"></a>\n",
    "## Dealing with Imbalanced Data\n",
    "\n",
    "When the data samples do not equally represent their classes, it is referred to as \"imbalanced data\" problem. In our example, we have __3196318__ samples of the normal hard drives, while this number for the \"failure\" class is only __234__. This clearly results in a model that is biased towards the \"normal\" samples, and the gradients of the \"failure\" class tend to vanish during the training epochs.\n",
    "\n",
    "\n",
    "```\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "      normal       1.00      1.00      1.00   3196318\n",
    "        fail       0.67      0.03      0.05       234\n",
    "```\n",
    "\n",
    "\n",
    "To mitigate the issue, in the next section we are going to resample the \"normal\" class to be roughly equal to the number of samples in the \"failure\" class (by a factor of 1.2)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sampled Dataset Sizes:\n",
      "- Train: Number of Normal Disks : 1273\n",
      "- Train: Number of Failed Disks : 1061\n",
      "- Test:  Number of Normal Disks : 281\n",
      "- Test:  Number of Failed Disks : 234\n"
     ]
    }
   ],
   "source": [
    "# Sampling training and test data sets for smaller, more balanced test\n",
    "num_normal_delta = round(train_fail_count * 0.2)\n",
    "num_normal_test_delta = round(test_fail_count * 0.2)\n",
    "\n",
    "# Sample training data\n",
    "df_tmp = df[df['failure'] > 0]\n",
    "sample_train_count_failed = df_tmp.shape[0]\n",
    "df_tmp = df_tmp.append(df[df['failure'] == 0].sample(n=(df_tmp.shape[0]+num_normal_delta)))\n",
    "sample_train_count_normal = df_tmp.shape[0]-sample_train_count_failed\n",
    "\n",
    "# separate training set features from labels\n",
    "df_train = df_tmp.drop(columns=['failure'])\n",
    "df_train_target = pd.DataFrame(df_tmp['failure'])\n",
    "\n",
    "# Sample test data\n",
    "df_tmp = df_t[df_t['failure'] > 0]\n",
    "sample_test_count_failed = df_tmp.shape[0]\n",
    "df_tmp = df_tmp.append(df_t[df_t['failure'] == 0].sample(n=(df_tmp.shape[0]+num_normal_test_delta)))\n",
    "sample_test_count_normal = df_tmp.shape[0]-sample_test_count_failed\n",
    "\n",
    "# separate test set features from labels\n",
    "df_test = df_tmp.drop(columns=['failure'])\n",
    "df_test_target = pd.DataFrame(df_tmp['failure'])\n",
    "\n",
    "# Output name data set sizes\n",
    "print('Sampled Dataset Sizes:')\n",
    "print('- Train: Number of Normal Disks :',sample_train_count_normal)\n",
    "print('- Train: Number of Failed Disks :',sample_train_count_failed)\n",
    "print('- Test:  Number of Normal Disks :',sample_test_count_normal)\n",
    "print('- Test:  Number of Failed Disks :',sample_test_count_failed)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"e2\"></a>\n",
    "### Exercise 2: Implications of balancing data samples\n",
    "\n",
    "Before proceeding with training, take some time and think about the implications of rebalancing the data. we have down-sampled the \"normal\" data points from 3M to ~320, a 10-thousand-time reduction in the number of samples. \n",
    "* How do you expect this down-sampling to affect each class? \n",
    "* What would be the effect on the precision/recall and f1-score values for each class? \n",
    "\n",
    "Write your answers below before training the model:"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {},
   "source": [
    "\n",
    "#Your answer to the above question\n",
    "# will help increase catch the defects but decrease the correct guess ig\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we are going to build the cuDF data from the pandas DataFrame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load sampled data into cuDF Dataframe\n",
    "\n",
    "gdf_train = cudf.DataFrame.from_pandas(df_train)\n",
    "gdf_train_target = cudf.DataFrame.from_pandas(df_train_target)\n",
    "\n",
    "gdf_eval = cudf.DataFrame.from_pandas(df_test)\n",
    "gdf_eval_target = cudf.DataFrame.from_pandas(df_test_target)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And finally perform training on the resampled training data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Converting cuDF data sets into XGBoost DMatrix format...\n",
      "Completed converting data sets\n",
      "Starting GPU XGBoost Training with cuDF Dataframes...\n",
      "[05:42:52] WARNING: /conda/conda-bld/xgboost_1591204597597/work/src/learner.cc:480: \n",
      "Parameters: { silent } might not be used.\n",
      "\n",
      "  This may not be accurate due to some parameters are only used in language bindings but\n",
      "  passed down to XGBoost core.  Or some parameters are not used but slip through this\n",
      "  verification. Please open an issue if you find above cases.\n",
      "\n",
      "\n",
      "[0]\ttrain-rmse:0.48699\teval-rmse:0.48674\n",
      "Multiple eval metrics have been passed: 'eval-rmse' will be used for early stopping.\n",
      "\n",
      "Will train until eval-rmse hasn't improved in 10 rounds.\n",
      "[1]\ttrain-rmse:0.47459\teval-rmse:0.47412\n",
      "[2]\ttrain-rmse:0.46313\teval-rmse:0.46302\n",
      "[3]\ttrain-rmse:0.45193\teval-rmse:0.45255\n",
      "[4]\ttrain-rmse:0.44135\teval-rmse:0.44241\n",
      "[5]\ttrain-rmse:0.43197\teval-rmse:0.43321\n",
      "[6]\ttrain-rmse:0.42303\teval-rmse:0.42477\n",
      "[7]\ttrain-rmse:0.41477\teval-rmse:0.41808\n",
      "[8]\ttrain-rmse:0.40717\teval-rmse:0.41067\n",
      "[9]\ttrain-rmse:0.39995\teval-rmse:0.40455\n",
      "[10]\ttrain-rmse:0.39295\teval-rmse:0.39857\n",
      "[11]\ttrain-rmse:0.38726\teval-rmse:0.39331\n",
      "[12]\ttrain-rmse:0.38150\teval-rmse:0.38858\n",
      "[13]\ttrain-rmse:0.37650\teval-rmse:0.38345\n",
      "[14]\ttrain-rmse:0.37140\teval-rmse:0.37967\n",
      "[15]\ttrain-rmse:0.36680\teval-rmse:0.37609\n",
      "[16]\ttrain-rmse:0.36232\teval-rmse:0.37205\n",
      "[17]\ttrain-rmse:0.35834\teval-rmse:0.36892\n",
      "[18]\ttrain-rmse:0.35489\teval-rmse:0.36628\n",
      "[19]\ttrain-rmse:0.35209\teval-rmse:0.36352\n",
      "GPU XGBoost Training with cuDF Dataframes compeleted - elapsed time: 0.20801687240600586 seconds\n"
     ]
    }
   ],
   "source": [
    "# GPU, with cuDF\n",
    "\n",
    "start_time = time.time()\n",
    "\n",
    "print('Converting cuDF data sets into XGBoost DMatrix format...')\n",
    "xgtrain = xgb.DMatrix(gdf_train, gdf_train_target)\n",
    "xgeval = xgb.DMatrix(gdf_eval, gdf_eval_target)\n",
    "print('Completed converting data sets')\n",
    "\n",
    "params = {'tree_method': \"gpu_\"+TREE_METHOD, 'max_depth': MAX_TREE_DEPTH, 'alpha': REGULARIZATION,\n",
    "          'gamma': GAMMA, 'subsample': SUBSAMPLE, 'scale_pos_weight': POS_WEIGHT, 'learning_rate': 0.05, 'silent': 1}\n",
    "\n",
    "print('Starting GPU XGBoost Training with cuDF Dataframes...')\n",
    "bst = xgb.train(params, xgtrain, 20, evals=[(xgtrain, \"train\"), (xgeval, \"eval\")],\n",
    "                early_stopping_rounds=EARLY_STOP)\n",
    "\n",
    "timetaken_gpu = time.time() - start_time\n",
    "print('GPU XGBoost Training with cuDF Dataframes compeleted - elapsed time:', timetaken_gpu,'seconds')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that the model is trained, let's run it over the test set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting XGBoost prediction on test dataset...\n",
      "XGBoost prediction completed\n"
     ]
    }
   ],
   "source": [
    "print('Starting XGBoost prediction on test dataset...')\n",
    "\n",
    "# Use trained model to predict on test data set\n",
    "preds = bst.predict(xgeval)\n",
    "\n",
    "# Convert predictions to \"normal\" (0) or \"failed\"\n",
    "y_pred = []\n",
    "THRESHOLD = 0.5\n",
    "\n",
    "for pred in preds:\n",
    "    if pred<=THRESHOLD:\n",
    "        y_pred.append(0)\n",
    "    if pred>THRESHOLD:\n",
    "        y_pred.append(1)\n",
    "\n",
    "# Output array of classifications                \n",
    "y_pred = np.asarray(y_pred)\n",
    "        \n",
    "y_true = df_test_target.values.reshape(len(preds))\n",
    "print('XGBoost prediction completed')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And finally review accuracy with the new results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy (Eval) 0.85243\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      normal       0.81      0.96      0.88       281\n",
      "        fail       0.93      0.73      0.82       234\n",
      "\n",
      "    accuracy                           0.85       515\n",
      "   macro avg       0.87      0.84      0.85       515\n",
      "weighted avg       0.87      0.85      0.85       515\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy (Eval)\", round(accuracy_score(y_true, y_pred),5))\n",
    "\n",
    "print(classification_report(y_true, y_pred, target_names=[\"normal\", \"fail\"]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As you can see, the recall/precision values for the \"fail\" class are hugely improved. The cost we pay here is the reduction in the precision/recall values of the \"normal\" class. Could you explain some scenarios where each of these strategies would be applicable? \n",
    "Finally, Compare the obtained results with your answers to exercise 2. How closely were you able to predict the results?\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Summary\n",
    "\n",
    "In this lab, we learned how to work with a sample dataset and curate, normalize, and filter data to prepare it for our XGBoost model. We reviewed the XGBoost model and learned how to utilize the cuDF data structure to accelerate the training/testing phases for the XGBoost model. Moreover, we learned how to modify the data to mitigate the issue of imbalanced classes.\n",
    "In the next two labs, we will use deep-learning-based approaches for classification and prediction."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"All-answers\"></a>\n",
    "## Answers to selected exercises:\n",
    "---\n",
    "\n",
    "<a name=\"a1\"></a>\n",
    "**Exercise 1: More insights into data:**\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.loc[:, (df != 0).any(axis=0)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Click [here](#e1) to go back"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}