problem1.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "339\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f023fbe6d30>]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#!/usr/bin/env python3\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from scipy.optimize import curve_fit\n",
    "import numpy as np\n",
    "import torch\n",
    "from torch import nn\n",
    "from torch.autograd import Variable\n",
    "#读取工作簿和工作簿中的工作表\n",
    "\n",
    "data_frame=pd.read_excel('Problem_C_Data_Wordle.xlsx')\n",
    "\n",
    "data = list(reversed(data_frame.loc[ 1: ,\"Unnamed: 4\"].tolist()))\n",
    "start=19\n",
    "end=-1\n",
    "\n",
    "ydata = data[start:end]\n",
    "length=len(ydata)\n",
    "print(length)\n",
    "xdata = [ i for i in range(length)]\n",
    "local_max_x=[]\n",
    "local_max_y=[]\n",
    "local_min_x=[]\n",
    "local_min_y=[]\n",
    "for i in range(length-1):\n",
    "    if(i==0):\n",
    "        if(data[i]>data[i+1] and data[i]>data[i+2]):\n",
    "            local_max_x.append(xdata[i])\n",
    "            local_max_y.append(ydata[i])\n",
    "        if(data[i]<data[i+1] and data[i]<data[i+2]):\n",
    "            local_min_x.append(xdata[i])\n",
    "            local_min_y.append(ydata[i])\n",
    "    if(i==1):\n",
    "        if(data[i]>data[i+1] and data[i]>data[i-1] and data[i]>data[i+2]):\n",
    "            local_max_x.append(xdata[i])\n",
    "            local_max_y.append(ydata[i])\n",
    "        if(data[i]<data[i+1] and data[i]<data[i-1] and data[i]<data[i+2]):\n",
    "            local_min_x.append(xdata[i])\n",
    "            local_min_y.append(ydata[i])\n",
    "    else:\n",
    "        if(data[i]>data[i+1] and data[i]>data[i-1] and data[i]>data[i+2] and data[i]>data[i-2]):\n",
    "            local_max_x.append(xdata[i])\n",
    "            local_max_y.append(ydata[i])\n",
    "        if(data[i]<data[i+1] and data[i]<data[i-1] and data[i]<data[i+2] and data[i]<data[i-2]):\n",
    "            local_min_x.append(xdata[i])\n",
    "            local_min_y.append(ydata[i])    \n",
    "x=np.linspace(xdata[0], xdata[-1], len(xdata))  \n",
    "# print(x)          \n",
    "interp_max = np.interp(x, local_max_x, local_max_y)\n",
    "interp_min = np.interp(x, local_min_x, local_min_y)\n",
    "for i in range(len(interp_max)):\n",
    "    if(interp_max[i]<interp_min[i]):\n",
    "        temp=interp_min[i]\n",
    "        interp_min[i]=interp_max[i]\n",
    "        interp_max[i]=temp\n",
    "# print(x)       \n",
    "\n",
    "#print(interp_max)\n",
    "plt.plot(x,interp_max,'g--')\n",
    "plt.plot(x,interp_min,'r--')\n",
    "#plt.savefig(\"regression\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[302348, 317096.0, 314406.0, 305094.0, 303953.5, 318000.5, 346488.5, 356785.5, 360042.0, 358927.5, 339688.5, 315358.0, 299623.0, 312232.0, 320804.0, 305101.0, 291828.0, 274355.5, 273678.0, 269496.0, 274678.5, 288778.5, 315862.0, 303620.5, 273782.5, 277816.5, 276018.5, 292543.5, 291966.0, 264125.0, 253290.5, 252135.0, 249388.0, 250753.5, 245615.5, 248720.5, 248661.0, 221874.0, 216812.5, 224245.0, 218595.0, 213034.0, 204636.0, 205341.5, 217616.5, 209199.0, 185742.5, 182421.0, 194130.5, 210355.5, 193463.5, 174450.5, 168070.5, 155649.0, 164311.5, 166898.5, 158473.0, 162925.5, 159631.5, 149852.0, 157487.5, 169582.0, 161383.0, 153604.5, 146679.0, 139933.5, 149863.5, 139805.5, 127091.5, 125503.5, 119606.0, 117808.5, 129459.5, 137684.0, 130225.5, 118034.5, 112367.5, 119081.0, 118351.5, 121719.5, 118989.0, 107334.0, 109532.0, 110641.0, 105453.0, 99981.0, 108593.5, 107397.0, 96507.0, 94500.0, 97350.5, 101060.0, 93970.5, 97813.0, 92321.5, 77824.5, 86650.5, 90730.0, 96783.5, 96864.5, 81135.5, 75375.0, 73488.0, 80725.0, 81672.0, 76929.0, 77559.5, 76629.0, 75405.0, 70170.0, 67732.0, 69535.5, 72327.5, 72426.5, 70402.0, 68349.0, 67361.5, 67170.0, 64905.5, 63051.5, 62955.5, 63517.0, 61957.5, 58454.0, 58904.0, 61868.5, 63004.5, 62259.5, 63354.5, 61847.0, 57500.5, 57608.0, 58734.5, 60008.5, 60523.0, 57698.0, 53667.0, 54321.0, 55243.0, 56885.0, 57978.5, 54709.5, 54047.5, 50935.0, 51282.0, 52921.5, 51913.0, 50493.5, 50378.0, 51864.0, 48353.0, 48269.5, 49218.0, 47649.0, 46478.5, 44928.5, 45730.0, 44506.5, 41125.5, 41565.5, 43611.5, 45961.0, 45375.5, 43106.5, 44950.0, 44439.5, 41165.0, 43727.5, 46578.0, 43397.5, 39891.5, 39001.5, 39190.0, 41092.5, 41120.5, 40952.0, 40661.5, 41092.5, 39934.0, 38291.0, 39520.5, 39199.5, 38777.5, 39517.0, 39220.5, 37572.0, 38301.5, 37956.0, 35785.5, 36645.0, 37805.0, 37289.5, 38095.5, 37532.0, 35869.5, 35869.5, 36938.5, 37477.5, 35749.5, 34737.0, 33464.0, 33514.0, 35240.5, 35460.0, 35376.5, 34451.5, 36105.0, 36931.0, 35752.5, 34718.5, 33624.5, 35218.5, 35726.5, 32978.5, 30727.5, 32247.5, 33970.5, 34501.5, 33623.0, 33813.5, 33457.5, 31604.5, 32375.5, 32733.5, 31863.0, 31477.0, 32067.0, 30704.5, 28562.0, 28517.0, 29322.0, 30819.5, 32743.0, 35326.5, 35363.5, 33260.0, 34076.0, 33163.5, 31626.5, 33215.5, 32982.0, 32143.0, 30885.5, 30350.0, 31345.5, 31170.0, 30916.0, 30850.0, 29712.5, 29145.0, 31188.0, 32151.0, 31474.5, 31728.5, 30774.0, 27965.5, 27656.5, 27643.0, 27726.5, 28863.0, 28174.0, 28051.5, 29654.5, 30431.0, 30864.0, 29940.5, 28467.0, 28531.5, 28689.0, 28860.5, 29181.5, 29113.0, 28950.0, 29508.0, 28836.0, 27757.0, 26530.5, 24914.0, 25585.0, 27000.0, 27586.0, 28612.0, 28442.0, 28536.5, 30405.5, 28582.0, 26654.5, 28098.5, 28225.5, 26730.0, 25326.5, 24872.5, 25810.5, 27005.5, 26525.5, 26520.5, 28336.5, 26978.5, 24870.0, 24639.5, 25862.5, 27050.0, 27184.0, 25951.0, 25289.0, 25793.5, 25628.5, 24895.0, 24654.0, 24098.5, 23637.0, 24259.5, 24725.0, 24365.0, 23331.0, 24204.0, 23049.0, 22419.5, 22398.5, 21552.0, 22410.0, 23487.0, 22462.5, 21500.0, 22514.5, 22594.5, 22251.0, 24088.0, 25073.5, 23158.5, 21335.0, 21213.5, 21109.0, 17917.5, 17782.5, 20445.0, 20519.5, 20080.5, 20602.5]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "new_ydata=[ydata[0]]\n",
    "for i in range(1,len(ydata)):\n",
    "    new_ydata.append((ydata[i]-ydata[i-1])*0.5+ydata[i-1])\n",
    "plt.plot(xdata,new_ydata,'b-')\n",
    "print(new_ydata)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def func(x, a, b, c,d):\n",
    "    return a * np.exp(-(x-d)*b) + c"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 4.49318800e+05  1.58649467e-02  2.12399254e+04 -1.51024581e+01] [ 4.35578283e+05  1.51983734e-02  1.87390311e+04 -1.90758032e+01]\n",
      "[23481.58159538 23446.29846186 23411.57067643 23377.38949807\n",
      " 23343.74632329 23310.63268405 23278.0402456  23245.96080434\n",
      " 23214.38628584 23183.30874272 23152.72035271 23122.61341666\n",
      " 23092.9803566  23063.81371383 23035.10614706 23006.85043054\n",
      " 22979.03945225 22951.66621212 22924.72382026 22898.20549522] [21295.10447828 21256.55004519 21218.57714639 21181.17701032\n",
      " 21144.34099776 21108.06059975 21072.32743571 21037.13325145\n",
      " 21002.46991731 20968.32942621 20934.70389188 20901.58554698\n",
      " 20868.96674135 20836.8399402  20805.19772241 20774.03277877\n",
      " 20743.33791035 20713.10602677 20683.33014464 20654.00338586]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "interp_max=np.array(interp_max)\n",
    "interp_min=np.array(interp_min)\n",
    "x=np.array(x)\n",
    "plt.plot(xdata,ydata,'b-')\n",
    "popt_min, pcov_min = curve_fit(func, x, interp_min,bounds = ([0,1e-4,1e+04,-150],[1e6,1,4e+04,150]))\n",
    "#popt数组中，三个值分别是待求参数a,b,c\n",
    "y_min = [func(i, popt_min[0],popt_min[1],popt_min[2],popt_min[3]) for i in xdata]\n",
    "plt.plot(xdata,y_min,'r--')\n",
    "popt_max, pcov_max = curve_fit(func, x, interp_max,maxfev=10000,bounds = ([0,1e-4,1e+04,-150],[1e6,1,4e+04,150]))\n",
    "#popt数组中，三个值分别是待求参数a,b,c\n",
    "y_max = [func(i, popt_max[0],popt_max[1],popt_max[2],popt_max[3]) for i in xdata]\n",
    "plt.plot(xdata,y_max,'g--')\n",
    "y_predict_max=[]\n",
    "y_predict_min=[]\n",
    "for k in range(len(ydata)):\n",
    "    pre_data=k\n",
    "    y_predict_max.append(func(pre_data, popt_max[0],popt_max[1],popt_max[2],popt_max[3]))\n",
    "    y_predict_min.append(func(pre_data, popt_min[0],popt_min[1],popt_min[2],popt_min[3]))\n",
    "\n",
    "print(popt_max,popt_min)\n",
    "print(np.array(y_predict_max[-20:]),np.array(y_predict_min[-20:]))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "957829556.8132689\n"
     ]
    }
   ],
   "source": [
    "gd=[21204,\n",
    "20380,\n",
    "22072,\n",
    "22252,\n",
    "22018,\n",
    "22394,\n",
    "22283,\n",
    "19725,\n",
    "21946,\n",
    "21510,\n",
    "20248,\n",
    "22248,\n",
    "22184,\n",
    "22606,\n",
    "20879,\n",
    "19542,\n",
    "24058,\n",
    "22147,\n",
    "21946,\n",
    "22775,\n",
    "22670,\n",
    "22818]\n",
    "exp_max=[20943.59,\n",
    "21982.324,\n",
    "23239.725,\n",
    "24064.318,\n",
    "24847.467,\n",
    "25650.793,\n",
    "26434.133,\n",
    "27232.865,\n",
    "27774.508,\n",
    "28405.03,\n",
    "28809.441,\n",
    "29170.197,\n",
    "29882.947,\n",
    "30674.066,\n",
    "31585.715,\n",
    "32521.205,\n",
    "33477.273,\n",
    "34468,\n",
    "35482.934,\n",
    "36506.64,\n",
    "37559.914,\n",
    "38630.816]\n",
    "exp_min=[17943.59,\n",
    "18982.324,\n",
    "20239.725,\n",
    "21064.318,\n",
    "21847.467,\n",
    "22650.793,\n",
    "23434.133,\n",
    "24232.865,\n",
    "24774.508,\n",
    "25405.03,\n",
    "25809.441,\n",
    "26170.197,\n",
    "26882.947,\n",
    "27674.066,\n",
    "28585.715,\n",
    "29521.205,\n",
    "30477.273,\n",
    "31468,\n",
    "32482.934,\n",
    "33506.64,\n",
    "34559.914,\n",
    "35630.816]\n",
    "mse=0\n",
    "for i in range(len(gd)):\n",
    "    if(gd[i]<exp_min[i] or gd[i]>exp_max[i]):\n",
    "        mse+=min((gd[i]-exp_min[i])**2,(gd[i]-exp_max[i])**2)\n",
    "print(mse)\n",
    "pinaw=0\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "(1+np.exp(1.5))*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.18584070796460178\n",
      "158048864.20593685\n",
      "7403.411481049138\n",
      "0.9789779751105627\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import mean_squared_error #均方误差\n",
    "from sklearn.metrics import mean_absolute_error #平方绝对误差\n",
    "from sklearn.metrics import r2_score#R square\n",
    "#调用\n",
    "acc=0\n",
    "total=0\n",
    "y_pred=[]\n",
    "y_test=ydata\n",
    "x_test=xdata\n",
    "for i in range(len(y_test)):\n",
    "    if(y_test[i]>y_predict_min[i] and y_test[i]<y_predict_max[i]):\n",
    "        acc+=1\n",
    "    total+=1\n",
    "    y_pred.append((y_predict_min[i]+y_predict_max[i])/2)\n",
    "print(acc/total)\n",
    "\n",
    "print(mean_squared_error(y_test,y_pred))\n",
    "print(mean_absolute_error(y_test,y_pred))\n",
    "print(r2_score(y_test,y_pred))\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fe992bcb940>]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import statsmodels.api as sm\n",
    "strengths=[]\n",
    "x_axis=[]\n",
    "for k in range(len(data)-8):\n",
    "    ytrend=data[k:]\n",
    "    res = sm.tsa.seasonal_decompose(ytrend, period=4)\n",
    "    if(k==0):\n",
    "        res.plot()\n",
    "        plt.savefig('decompose')\n",
    "        # print(list(res.resid)[2:-2])\n",
    "    x_axis.append(k)\n",
    "    data_no_trend=ytrend[2:-2]-res.trend[2:-2]\n",
    "    # print(data_no_trend)\n",
    "    deseasonal=ytrend[2:-2]-res.seasonal[2:-2]\n",
    "    # print(deseasonal)\n",
    "    trend_strength=max(0,1-res.resid[2:-2].var()/deseasonal.var())\n",
    "    # print(trend_strength)\n",
    "    strengths.append(trend_strength)\n",
    "\n",
    "plt.plot([i for i in range(len(data)-8)],strengths,'g-')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "batch_size=32\n",
    "time_step=30\n",
    "epoch=5000\n",
    "input_size=2\n",
    "output_size=1\n",
    "mid_dim=16\n",
    "mid_layers=1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# In[122]:\n",
    " \n",
    "class RegLSTM(nn.Module):\n",
    "    def __init__(self, inp_dim, out_dim, mid_dim, mid_layers,batch):\n",
    "        super(RegLSTM, self).__init__()\n",
    " \n",
    "        self.rnn = nn.LSTM(inp_dim, mid_dim, mid_layers,batch_first=batch)  # rnn\n",
    "        self.reg = nn.Sequential(\n",
    "            nn.Linear(mid_dim, mid_dim),\n",
    "            nn.Tanh(),\n",
    "            nn.Linear(mid_dim, out_dim),\n",
    "        )  # regression\n",
    " \n",
    "    def forward(self, x):\n",
    "        # print(x)\n",
    "        y = self.rnn(x)[0]  # y, (h, c) = self.rnn(x)\n",
    "        # print(y)\n",
    "        batch_size, seq_len, hid_dim = y.shape\n",
    "        y = y.reshape(-1, hid_dim)\n",
    "        \n",
    "        y = self.reg(y)\n",
    "        \n",
    "        y = y.reshape(batch_size, seq_len, -1)\n",
    "        return y\n",
    " \n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_data(ydata,xdata):\n",
    "    lstm_data=[np.log(ydata[i]) for i in range(len(ydata))]\n",
    "    lstm_data=np.diff(lstm_data)\n",
    "    \n",
    "    seq_number=np.array(lstm_data,dtype=np.float32)\n",
    "    #print(seq_number)\n",
    "    seq_number=seq_number[:,np.newaxis]\n",
    "    seq_year_month=np.array(xdata[:-1])[:,np.newaxis]\n",
    "    # print(seq_number)\n",
    "    # print(seq_year_month)\n",
    "    #print(seq_year_month.shape)\n",
    "    seq=np.concatenate((seq_number,seq_year_month),axis=1)\n",
    "    # print(seq.mean(axis=0),seq.std(axis=0))\n",
    "    seq=(seq-seq.mean(axis=0))/seq.std(axis=0)\n",
    "    return(seq)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1.1769853038448788, -1.2029865144892695, 0.7128892404167667, -0.6194566886866062, 1.8042869300673194, 0.43993424035532236, 0.4264823470435267, -0.029188450616522896, 0.1404893597504261, -1.2822063009618518, -0.22926699911250822, -0.7955923481933544, 1.8879114642600436, -1.0309188574548198, 0.07095410605498373, -0.9479482683738621, -0.2883326063904865, 0.4148670900244781, -0.598584173198767, 1.209514382263117, 0.16777248427098537, 2.0259533008098716, -2.872027828989604, 0.8196532940268955, -0.28695609015230233, 0.32083072981773403, 1.1934757008237364, -1.0585568121326212, -1.096897471571354, 0.3325563965388357, -0.257210773897373, 0.18750757242943797, 0.12333080925154702, -0.4285445886340112, 0.8967238700836684, -0.7191703314092235, -1.8195026503913323, 1.5004227512898078, -0.4959842419156356, 0.0916663588959097, -0.5171750865562468, -0.23163437944704782, 0.4939920596679566, 1.0278943314626894, -1.8242732876984507, -0.7003709730766529, 0.47326765475604227, 1.1403415312038672, 0.9234723794905643, -2.8641412231438217, 0.810964545625049, -1.5425483824577066, -0.007511751165854634, 1.416392194321792, -0.8501885125304168, -0.1567242415432914, 0.9709358685909066, -1.28810874630048, 0.03798062397109994, 1.2743204470812155, 0.6574772517548965, -1.6909279028652193, 0.7802445010740446, -1.7339279212252878, 0.8775891998314087, 0.9035187142178831, -2.466051788882529, 0.5613429334723992, -0.6792199141508648, -0.24954040514902232, 0.08226439536287876, 2.2045915078395106, -0.4968249974634878, -0.6220324228137297, -1.5322058339711353, 0.618754241172508, 0.9093579193809843, -0.875013495554496, 1.678734333237056, -2.0705972461103563, -0.050201095648519614, 0.6987602732888998, -0.27550562243678994, -0.670610215968427, -0.38092195667437606, 2.3834484494323047, -2.488483808836227, 0.320009736824734, -0.6369978455049865, 1.4831997064532672, -0.3913408624607936, -1.1493734395291813, 2.2046418914958372, -3.5574276311372786, 0.041777895729744406, 2.5203860152248567, -1.1722619527828704, 2.7452779823927687, -2.5399562733494307, -1.3020251465196735, -0.19203734272995718, -0.21614318592835124, 2.4705925426245408, -1.9866556423562685, 0.8548863242907702, -0.475628766983752, 0.3825965753184043, -0.5826737094778999, -0.924198455202112, 0.30409097717424755, 0.4895946957595178, 0.6093682872091529, -0.3934387336681739, -0.07990870017103055, -0.4321159211838173, 0.2811987720147069, -0.16489319131834018, -0.45649987347829535, -0.0298250111951474, 0.1777841402663429, 0.21245090275315098, -0.6193166568693682, -0.5527557206145867, 0.909480924997119, 0.4307163549765834, 0.17919761899722852, -0.27597319145433546, 0.8562264426806773, -1.261146714432808, -0.21757334338684836, 0.44385150448341376, 0.19350167476010277, 0.4870920179340675, -0.1021360256623564, -0.8472732099333286, -0.6511529165789329, 1.1066796023336436, -0.5167303204866076, 1.3567029372771229, -0.7088110173266677, -0.45378788630815836, 0.35809554617424727, -1.6190272980261373, 1.94944867220304, -0.9831182321989561, 0.7337141954704014, -1.2252059675342943, 1.3579379449357285, -0.46910486530485074, -1.0009614303501755, 1.1457632469664438, -0.4921537375750544, -0.07325804667741004, -0.3265506076865767, -0.28023921038283395, 0.8660064315838533, -1.3465245998656878, -0.27096546864027116, 0.6974100347335409, 0.6085270538852494, 0.7935504772363331, -0.9205847308862933, -0.07089238199323673, 1.2047461772704418, -1.3028783676790479, -0.2595687941647244, 1.7918032499236258, -0.0745470973225536, -1.4411069712307751, -0.29270964309964886, -0.04734438576900391, 0.3421763738989892, 0.9327301157354061, -0.7329257645332918, 0.8236007666316894, -0.8123037383130641, 1.2311909079770391, -1.760347720708841, 1.0190263736073928, -0.0809225735276781, 0.07471260058632252, -0.144969951693294, 0.7604206171505902, -0.7586196050238811, -0.044254553752059396, 0.6692460486231809, -0.7036203714362084, -0.47959291831615747, 1.197174034934834, -0.2636444410538451, 0.12949935684833175, 0.5480509431181598, -0.7219425617960196, -0.13814935108649176, 0.32148206887831116, 0.5434040063741599, -0.018152996552905292, -0.9209764203800439, 0.45351995439568815, -1.1722166943605496, 1.388554268649399, 0.0020059061094792816, 0.3251227223897524, -0.19738734865064178, -0.23764380318699635, 1.4754154315508512, -0.7383549509759444, 0.1800428048321025, -0.694121227982865, 0.14402560833609637, 1.0977054044581969, -0.5680591481450947, -1.1380297330440567, -0.2980559136802955, 1.5640680809081957, -0.12147774668160527, 0.6605243336737899, -1.1022469997159874, 1.4105948597144655, -1.4904172819405959, 0.3967739064129259, 0.34916727995101166, 0.09202254675322531, -0.5459233059090597, 0.45098880546099473, 0.16802075951022527, -1.0236771624100145, -0.45951889691271747, 0.6069056122694582, 0.2308381358072283, 1.092906882403353, 0.5197280313194395, 1.4016645303673716, -1.1924851706168973, -0.01910811431378091, 0.7583660063929522, -1.2362389458492822, 0.34936688175501823, 0.9622470333206873, -0.9504502512523287, 0.5516694451944921, -1.338205476390856, 1.1342041069163378, -0.1765287103853856, 0.23006811283207218, -0.23947390263467683, 0.373619141596064, -1.094806115491947, 0.8464123145069214, 0.9144524019825588, -0.0076979209476897015, -0.3080728360707647, 0.6749630734196909, -1.2342824963531074, -0.7930375493441183, 0.7254488869526284, -0.553822464234737, 0.8054950522573353, 0.3243506145736449, -0.7172838109532947, 0.8022119055348784, 0.6803676761127911, 0.11312209913868417, 0.3976719082185695, -0.943688764573337, -0.027110439690914347, 0.2628920007910989, 0.04940047329977481, 0.2722522960087861, 0.16957760906143515, -0.04129559979755046, 0.0940827891518538, 0.5303040819925714, -0.9011602677417844, 0.21600177780112118, -1.1175106415316722, -0.13484351000159478, 0.9241417003043475, 0.5252686698846208, 0.1626721000493218, 0.859666430250543, -0.820486565310696, 1.0781437761527581, 0.5998305307640216, -1.9416682493592456, 0.5803387011962695, 0.8267738076913987, -0.5351207893875782, -0.551436103160702, -0.5221335767199268, 0.2908949329379607, 0.7472935898366331, 0.4971083322459258, -0.7433953145811546, 0.9224836436693468, 0.8090636484533927, -1.8397808559054534, 0.2051189949030096, -0.24101192889531742, 1.513049416775421, -0.2419711729039094, 0.5386371044070793, -1.4868158929337398, 1.099207271758586, -0.439553460975498, 0.4761204121245875, -0.9919104408330174, 0.9575006322896292, -1.3330109214236128, 1.087677927560411, -0.27987397230865463, 0.8955197875658424, -1.0692377998003086, 0.2695748390169072, 0.7614254670549948, -1.7840037160377766, 1.3623663208533199, -1.202156139711988, 0.519091155842963, 0.573505982451437, 0.7014016799059934, -1.6123036859577633, 0.8250867370198238, 0.44228148894442465, -0.1751290437258032, 0.0025878210522283134, 1.956235176973449, -0.7796913531251417, -0.8941830868926922, -0.8323780637558096, 0.887269029500543, -0.823472752500133, -3.0023164895145418, 3.0293872507349686, 0.5867395538248497, -0.316914603111609, -0.0006537754415793041, 0.7726563315236229]\n"
     ]
    }
   ],
   "source": [
    "data_train=load_data(ydata,xdata)\n",
    "figure=[]\n",
    "for i in range(len(data_train)):\n",
    "    \n",
    "    figure.append(data_train[i][0])\n",
    "print(figure)\n",
    "train_size=int(len(data_train)-1)\n",
    "#print(data_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_sample=np.zeros((train_size-time_step+1,time_step,input_size))\n",
    "label_sample=np.zeros((train_size-time_step+1,time_step,output_size))\n",
    "for i in range(train_size-time_step+1):\n",
    "    data_sample[i]=data_train[i:i+time_step,:]\n",
    "    label_sample[i]=data_train[i+1:i+1+time_step,0:1:]\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    " \n",
    "device=torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "net=RegLSTM(input_size,output_size,mid_dim,mid_layers,True).to(device)\n",
    "criterion=nn.MSELoss()\n",
    "optimizer=torch.optim.Adam(net.parameters(),lr=1e-2)\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:    0, Loss: 0.78875\n",
      "Epoch:  100, Loss: 0.37145\n",
      "Epoch:  200, Loss: 0.17277\n",
      "Epoch:  300, Loss: 0.13112\n",
      "Epoch:  400, Loss: 0.11016\n",
      "Epoch:  500, Loss: 0.10686\n",
      "Epoch:  600, Loss: 0.09359\n",
      "Epoch:  700, Loss: 0.09072\n",
      "Epoch:  800, Loss: 0.08742\n",
      "Epoch:  900, Loss: 0.07745\n",
      "Epoch: 1000, Loss: 0.06893\n",
      "Epoch: 1100, Loss: 0.07159\n",
      "Epoch: 1200, Loss: 0.06442\n",
      "Epoch: 1300, Loss: 0.06793\n",
      "Epoch: 1400, Loss: 0.06173\n",
      "Epoch: 1500, Loss: 0.06218\n",
      "Epoch: 1600, Loss: 0.06265\n",
      "Epoch: 1700, Loss: 0.05432\n",
      "Epoch: 1800, Loss: 0.05526\n",
      "Epoch: 1900, Loss: 0.05508\n",
      "Epoch: 2000, Loss: 0.15416\n",
      "Epoch: 2100, Loss: 0.04975\n",
      "Epoch: 2200, Loss: 0.05009\n",
      "Epoch: 2300, Loss: 0.04524\n",
      "Epoch: 2400, Loss: 0.04550\n",
      "Epoch: 2500, Loss: 0.04326\n",
      "Epoch: 2600, Loss: 0.04484\n",
      "Epoch: 2700, Loss: 0.05139\n",
      "Epoch: 2800, Loss: 0.04209\n",
      "Epoch: 2900, Loss: 0.04330\n",
      "Epoch: 3000, Loss: 0.04056\n",
      "Epoch: 3100, Loss: 0.04408\n",
      "Epoch: 3200, Loss: 0.03890\n",
      "Epoch: 3300, Loss: 0.04482\n",
      "Epoch: 3400, Loss: 0.04077\n",
      "Epoch: 3500, Loss: 0.04644\n",
      "Epoch: 3600, Loss: 0.03888\n",
      "Epoch: 3700, Loss: 0.03434\n",
      "Epoch: 3800, Loss: 0.03413\n",
      "Epoch: 3900, Loss: 0.04186\n",
      "Epoch: 4000, Loss: 0.04261\n",
      "Epoch: 4100, Loss: 0.03592\n",
      "Epoch: 4200, Loss: 0.03386\n",
      "Epoch: 4300, Loss: 0.03733\n",
      "Epoch: 4400, Loss: 0.03533\n",
      "Epoch: 4500, Loss: 0.04022\n",
      "Epoch: 4600, Loss: 0.04527\n",
      "Epoch: 4700, Loss: 0.03544\n",
      "Epoch: 4800, Loss: 0.03209\n",
      "Epoch: 4900, Loss: 0.03479\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f4688139160>]"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "losses=[]\n",
    "for i in range(epoch):\n",
    "    for j in range(int((train_size-time_step+1)/batch_size)-1):\n",
    "        train_X=data_sample[j*batch_size:(j+1)*batch_size,:,:]\n",
    "        train_Y=label_sample[j*batch_size:(j+1)*batch_size,:,:]\n",
    "        var_x=torch.tensor(train_X,dtype=torch.float32,device=device)\n",
    "        var_y=torch.tensor(train_Y,dtype=torch.float32,device=device)\n",
    "        # print(var_x,var_y)\n",
    "        out = net(var_x)\n",
    "        loss=criterion(out,var_y)\n",
    "        #loss = criterion(out[:,-1,:], var_y[:,-1,:])\n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "    train_X=data_sample[(j+1)*batch_size:,:,:]\n",
    "    train_Y=label_sample[(j+1)*batch_size:,:,:]\n",
    "    var_x=torch.tensor(train_X,dtype=torch.float32,device=device)\n",
    "    var_y=torch.tensor(train_Y,dtype=torch.float32,device=device)\n",
    "    out = net(var_x)\n",
    "    loss = criterion(out, var_y)\n",
    "    optimizer.zero_grad()\n",
    "    loss.backward()\n",
    "    optimizer.step()\n",
    "    losses.append(loss)\n",
    "    if i%100==0:\n",
    "        print('Epoch: {:4}, Loss: {:.5f}'.format(i, loss.item()))\n",
    "epoch=[i for i in range(len(losses))]\n",
    "plt.plot(epoch,losses,'g-')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.052618515813335875\n",
      "-0.05273211396263733\n",
      "-0.052844681145218786\n",
      "-0.05295622048322279\n",
      "-0.05306673526312444\n",
      "-0.053176228930517766\n",
      "-0.05328470508496075\n",
      "-0.053392167474905605\n",
      "-0.053498619992694785\n",
      "-0.05360406666965112\n",
      "-0.05370851167124613\n",
      "-0.05381195929233051\n",
      "-0.05391441395246943\n",
      "-0.05401588019133641\n",
      "-0.054116362664208495\n",
      "-0.05421586613751295\n",
      "-0.05431439548448935\n",
      "-0.05441195568090684\n",
      "-0.05450855180087011\n",
      "-0.05460418901270003\n",
      "-0.05469887257491557\n",
      "-0.05479260783225648\n",
      "-0.05488540021182153\n",
      "-0.054977255219277765\n",
      "-0.05506817843511591\n",
      "-0.05515817551104796\n",
      "-0.05524725216640114\n",
      "-0.05533541418467536\n",
      "-0.0554226674101006\n",
      "-0.05550901774430983\n",
      "-0.05559447114310068\n",
      "-0.05567903361321491\n",
      "-0.05576271120926357\n",
      "-0.05584551003065741\n",
      "-0.0559274362186617\n",
      "-0.056008495953493664\n",
      "-0.056088695451492754\n",
      "-0.05616804096237615\n",
      "-0.05624653876654051\n",
      "-0.05632419517245424\n",
      "-0.05640101651409779\n",
      "-0.05647700914847675\n",
      "-0.05655217945321667\n",
      "-0.056626533824186254\n",
      "-0.05670007867321303\n",
      "-0.05677282042586107\n",
      "-0.056844765519247886\n",
      "-0.05691592039994653\n",
      "-0.0569862915219268\n",
      "-0.05705588534457107\n",
      "-0.05712470833074157\n",
      "-0.05719276694490105\n",
      "-0.057260067651290214\n",
      "-0.05732661691216556\n",
      "-0.05739242118608168\n",
      "-0.0574574869262392\n",
      "-0.05752182057886479\n",
      "-0.05758542858167104\n",
      "-0.05764831736232523\n",
      "-0.057710493337015834\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f4688093ac8>]"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "net=net.eval()\n",
    "lstm_out=list()\n",
    "reals=list()\n",
    "new_mean=np.array([-7.86209723e-03  ,1.68500000e+02] )\n",
    "new_std=np.array([8.57686214e-02 ,9.75717685e+01])\n",
    "for i in range(60):\n",
    "    test_x=data_train[-time_step:,:].reshape(1,time_step,input_size)\n",
    "    test_x=torch.tensor(test_x,dtype=torch.float32,device=device)\n",
    "    tem=net(test_x).cpu().data.numpy()\n",
    "    lstm_out.append(tem[0][-1])\n",
    "    \n",
    "\n",
    "    for item in data_train:\n",
    "        item[0]=item[0]*new_std[0]+new_mean[0]\n",
    "        item[1]=item[1]*new_std[1]+new_mean[1]\n",
    "    date=train_size+i\n",
    "    y_yester_max=func(date-1, popt_max[0],popt_max[1],popt_max[2],popt_max[3])\n",
    "    y_yester_min=func(date-1, popt_min[0],popt_min[1],popt_min[2],popt_min[3])\n",
    "    y_predict_max=func(date, popt_max[0],popt_max[1],popt_max[2],popt_max[3])\n",
    "    y_predict_min=func(date, popt_min[0],popt_min[1],popt_min[2],popt_min[3])\n",
    "    # print(y_predict_min,y_predict_max,y_predict_min,y_predict_max)\n",
    "    real=(np.log(y_predict_min)+np.log(y_predict_max)-np.log(y_yester_max)-np.log(y_yester_max))/2\n",
    "    print(real)\n",
    "    # print(lstm_out)\n",
    "    reals.append(real)\n",
    "    xdata=np.append(xdata,(xdata[-1]+1))\n",
    "    ydata=np.append(ydata,(real))\n",
    "    new=np.array([[int(real),train_size+1+i]])\n",
    "    data_train=np.concatenate((data_train,new),axis=0)\n",
    "    new_mean=data_train.mean(axis=0)\n",
    "    new_std=data_train.std(axis=0)\n",
    "    data_train=(data_train-data_train.mean(axis=0))/data_train.std(axis=0)\n",
    "    \n",
    "    \n",
    "\n",
    "    #print(data)\n",
    "y_test=np.array(reals).reshape((-1,1))\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "plt.plot(y_test, 'r', label='pred')\n",
    "\n",
    " \n",
    " \n",
    "\n",
    " \n",
    " \n",
    " \n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "331844.0008513118\n",
      "296968.0001500166\n",
      "313220.00015299395\n",
      "294686.99999333004\n",
      "341314.00230755337\n",
      "351663.00242208334\n",
      "361908.00267192913\n",
      "358176.00256283925\n",
      "359679.00263535866\n",
      "319698.0032430243\n",
      "311018.00321315834\n",
      "288228.002570476\n",
      "336236.0048893296\n",
      "305372.0051763544\n",
      "304830.00516895123\n",
      "278826.00457348395\n",
      "269885.00486623467\n",
      "277471.00507650635\n",
      "261521.0049039238\n",
      "287836.0044221004\n",
      "289721.00445731665\n",
      "342003.00744258834\n",
      "265238.0079145513\n",
      "282327.0079549873\n",
      "273306.0081229139\n",
      "278731.00821246725\n",
      "306356.0099145133\n",
      "277576.00875478203\n",
      "250674.00745682837\n",
      "255907.00745508543\n",
      "248363.007445674\n",
      "250413.0074098885\n",
      "251094.0074429503\n",
      "240137.00745385553\n",
      "257304.00836584484\n",
      "240018.00770436935\n",
      "203730.0068549528\n",
      "229895.0073934424\n",
      "218595.00667968977\n",
      "218595.00667996507\n",
      "207473.00644486453\n",
      "201799.0061863884\n",
      "208884.0062484991\n",
      "226349.00689489476\n",
      "192049.006994997\n",
      "179436.0064193087\n",
      "185406.00670587854\n",
      "202855.007497291\n",
      "217856.00798482978\n",
      "169071.0074809602\n",
      "179830.0077262424\n",
      "156311.00707694853\n",
      "154987.0070517721\n",
      "173636.00739215087\n",
      "160161.00694913723\n",
      "156785.0067809516\n",
      "169066.00774024185\n",
      "150197.006403895\n",
      "149507.0063464537\n",
      "165468.00744881615\n",
      "173696.00749442485\n",
      "149070.00615528785\n",
      "158139.00639572216\n",
      "135219.00576515912\n",
      "144648.0061708866\n",
      "155079.006251334\n",
      "124532.00427807505\n",
      "129651.00422951864\n",
      "121356.00364659543\n",
      "117856.00362182033\n",
      "117761.00362136173\n",
      "141158.00343079004\n",
      "134210.00334752654\n",
      "126241.00293707554\n",
      "109828.00184148467\n",
      "114907.00205924959\n",
      "123255.00262343061\n",
      "113448.00235140145\n",
      "129991.00231702901\n",
      "107987.00184679902\n",
      "106681.00186863614\n",
      "112383.00205032782\n",
      "108899.00184025022\n",
      "102007.00150237144\n",
      "97955.0014941731\n",
      "119232.00241687045\n",
      "95562.00199357983\n",
      "97452.00194934178\n",
      "91548.00182377761\n",
      "103153.00193084117\n",
      "98967.00167619581\n",
      "88974.00118665407\n",
      "106652.00067944684\n",
      "77991.00060384987\n",
      "77658.00060284551\n",
      "95643.00051801006\n",
      "85817.0003636179\n",
      "107750.00086186438\n",
      "85979.00097864798\n",
      "76292.00070050027\n",
      "74458.00069231066\n",
      "72518.00073267237\n",
      "88932.00030641249\n",
      "74412.00031070899\n",
      "79446.0001591877\n",
      "75673.0000730913\n",
      "77585.00005966221\n",
      "73224.99996871078\n",
      "67115.00011504855\n",
      "68349.00012516009\n",
      "70722.00005047947\n",
      "73933.00016290043\n",
      "70920.00025864386\n",
      "69884.00027668849\n",
      "66814.00033905418\n",
      "67909.00031483533\n",
      "66431.00028900284\n",
      "63380.00019290957\n",
      "62723.00021652704\n",
      "63188.0002314957\n",
      "63846.00022327753\n",
      "60069.00015656452\n",
      "56839.00005098535\n",
      "60969.00002117309\n",
      "62767.99999417095\n",
      "63240.99999310104\n",
      "61277.999978887914\n",
      "65430.9999718965\n",
      "58262.999768286354\n",
      "56737.99979183925\n",
      "58477.99975873601\n",
      "58990.99974989424\n",
      "61025.99978207457\n",
      "60019.999756434736\n",
      "55375.999815378054\n",
      "51957.99991888807\n",
      "56684.00005580458\n",
      "53801.99996209261\n",
      "59967.99983371027\n",
      "55988.999997460385\n",
      "53429.999904220975\n",
      "54664.99985707667\n",
      "47204.99964580998\n",
      "55358.99955677998\n",
      "50483.99962040422\n",
      "53341.99953754293\n",
      "47644.99942338782\n",
      "53110.99924842297\n",
      "50616.99928551167\n",
      "46088.99931466615\n",
      "50449.99924629325\n",
      "47985.99929667113\n",
      "47311.99929179476\n",
      "45644.99924483265\n",
      "44211.999206713524\n",
      "47247.99919842991\n",
      "41764.99933121147\n",
      "40485.99939000021\n",
      "42644.999424335256\n",
      "44577.999364960364\n",
      "47343.999388999015\n",
      "43406.99929894351\n",
      "42805.9992933838\n",
      "47093.99907364383\n",
      "41784.99927173981\n",
      "40544.999301173026\n",
      "46909.99888149233\n",
      "46245.99887937483\n",
      "40548.99897462411\n",
      "39233.999076479384\n",
      "38768.99907453783\n",
      "39610.999029919985\n",
      "42573.99887601844\n",
      "39666.99898840469\n",
      "42236.99886917831\n",
      "39085.99889559334\n",
      "43098.99883611541\n",
      "36768.99906974614\n",
      "39812.99886297942\n",
      "39227.99887150893\n",
      "39170.9988715646\n",
      "38383.998927458335\n",
      "40649.99890034626\n",
      "37790.99905461583\n",
      "37352.99906945401\n",
      "39249.998988745734\n",
      "36661.999108738164\n",
      "34908.99915172732\n",
      "38380.99904679902\n",
      "37228.999095269624\n",
      "37349.99909102326\n",
      "38840.998994627684\n",
      "36222.99899236626\n",
      "35515.99902632927\n",
      "36222.9989924575\n",
      "37653.99899595735\n",
      "37300.9989920173\n",
      "34197.999006364334\n",
      "35275.99900454666\n",
      "31651.99914452938\n",
      "35375.99910624736\n",
      "35104.99912073096\n",
      "35814.99913367632\n",
      "34937.999154114674\n",
      "33964.99920573293\n",
      "38244.99911313554\n",
      "35616.999078877016\n",
      "35887.99907628999\n",
      "33548.99923798273\n",
      "33699.999234523624\n",
      "36736.99908439009\n",
      "34715.999084731106\n",
      "31240.999255517112\n",
      "30213.999265924333\n",
      "34280.99896602364\n",
      "33659.99898395863\n",
      "35342.998936723445\n",
      "31902.999066951023\n",
      "35723.99883645717\n",
      "31190.999213193834\n",
      "32017.99921597993\n",
      "32732.99920908442\n",
      "32733.999209064186\n",
      "30991.999246540603\n",
      "31961.99923519825\n",
      "32171.999223766776\n",
      "29236.99936459034\n",
      "27886.99940773632\n",
      "29146.999361277063\n",
      "29496.999345755394\n",
      "32141.99917538517\n",
      "33343.999128381605\n",
      "37308.9989847309\n",
      "33417.999126784176\n",
      "33101.9991218454\n",
      "35049.99905939642\n",
      "31276.99925431893\n",
      "31975.999247924137\n",
      "34454.999219387355\n",
      "31508.999302240183\n",
      "32776.99927262248\n",
      "28993.999408468015\n",
      "31705.999353585226\n",
      "30984.999357430486\n",
      "31354.999351475413\n",
      "30476.99935431189\n",
      "31222.999323881722\n",
      "28201.999300413103\n",
      "30087.99914435622\n",
      "32287.99913412491\n",
      "32013.999133448968\n",
      "30934.999115354833\n",
      "32521.999091193302\n",
      "29025.999130689794\n",
      "26904.99926284995\n",
      "28407.99925009792\n",
      "26877.999318120797\n",
      "28574.999314806144\n",
      "29150.999300635405\n",
      "27196.999444769084\n",
      "28905.999382511498\n",
      "30402.999310042847\n",
      "30458.999309671755\n",
      "31268.99928097358\n",
      "28611.999330083454\n",
      "28321.999341898718\n",
      "28740.999340757564\n",
      "28636.999343590203\n",
      "29083.999335857705\n",
      "29278.999325633886\n",
      "28946.999330488463\n",
      "28952.999330531213\n",
      "30062.99926440342\n",
      "27608.999384217197\n",
      "27904.999370817186\n",
      "25155.99951342823\n",
      "24671.99952557196\n",
      "26497.999525392534\n",
      "27501.99955129734\n",
      "27669.999549480337\n",
      "29553.999597827955\n",
      "27329.999620646086\n",
      "29742.999523906954\n",
      "31067.999505737764\n",
      "26095.99954530337\n",
      "27212.999490311813\n",
      "28983.999433185298\n",
      "27466.999459515162\n",
      "25992.999527628614\n",
      "24659.99952657736\n",
      "25084.99951692104\n",
      "26535.999490114173\n",
      "27474.999484936663\n",
      "25575.999520927795\n",
      "27464.999433759775\n",
      "29207.999367563745\n",
      "24748.999347204037\n",
      "24990.999348097503\n",
      "24287.99938409588\n",
      "27436.999280520355\n",
      "26662.999293736062\n",
      "27704.99923517052\n",
      "24196.99926282945\n",
      "26380.999184716085\n",
      "25205.999247535554\n",
      "26050.999232082042\n",
      "23738.999337998852\n",
      "25568.999212253802\n",
      "22627.99922266194\n",
      "24645.99918067123\n",
      "23872.999213312494\n",
      "25576.999245700707\n",
      "23152.999296626458\n",
      "23508.999292623903\n",
      "24898.999248618504\n",
      "21198.999248777454\n",
      "23639.999213518455\n",
      "21156.999306621077\n",
      "21946.999273444897\n",
      "22872.999254798284\n",
      "24100.999218084755\n",
      "20823.999409086293\n",
      "22175.999357956352\n",
      "22852.999329311617\n",
      "22335.99932870014\n",
      "22165.99933027573\n",
      "26009.999238220807\n",
      "24136.999217553748\n",
      "22179.999349535527\n",
      "20489.999469367485\n",
      "21936.999378114335\n",
      "20280.99946154847\n",
      "15553.999639264457\n",
      "20010.999313085184\n",
      "20878.999254781193\n",
      "20159.99928253373\n",
      "20000.999293208686\n",
      "18975.78554086584\n",
      "18001.0773195318\n",
      "17074.513720967803\n",
      "16193.836478637722\n",
      "15356.885961723872\n",
      "14561.597299219848\n",
      "13805.99663136308\n",
      "13088.197485640281\n",
      "12406.397274576517\n",
      "11758.873912501795\n",
      "11143.982548478609\n",
      "10560.152412568987\n",
      "10005.88377261907\n",
      "9479.744998744936\n",
      "8980.369732712454\n",
      "8506.454159418485\n",
      "8056.7543776981065\n",
      "7630.083867704926\n",
      "7225.311052136454\n",
      "6841.356948605143\n",
      "6477.192910486956\n",
      "6131.838453613948\n",
      "5804.359166213428\n",
      "5493.864699535538\n",
      "5199.506836652113\n",
      "4920.4776369518295\n",
      "4656.00765390189\n",
      "4405.364223691257\n",
      "4167.849822418342\n",
      "3942.8004895334443\n",
      "3729.5843152953425\n",
      "3527.599990051347\n",
      "3336.2754131998\n",
      "3155.0663597452212\n",
      "2983.4552024067893\n",
      "2820.949687292262\n",
      "2667.0817612005108\n",
      "2521.406448666956\n",
      "2383.5007769172703\n",
      "2252.96274694535\n",
      "2129.410348982079\n",
      "2012.4806206713467\n",
      "1901.8287463192908\n",
      "1797.1271956318353\n",
      "1698.0649004038096\n",
      "1604.3464676707174\n",
      "1515.6914278812508\n",
      "1431.8335166947888\n",
      "1352.51998905364\n",
      "1277.5109642243342\n",
      "1206.5788005460427\n",
      "1139.507498667061\n",
      "1076.0921320922232\n",
      "1016.1383039051259\n",
      "959.4616285691396\n",
      "905.8872377502677\n",
      "855.249309143108\n",
      "807.3906173183112\n",
      "762.1621056462001\n",
      "719.4224783863824\n"
     ]
    }
   ],
   "source": [
    "\n",
    "firstday=np.log(ydata[0])\n",
    "pred=[]\n",
    "pred.append(firstday)\n",
    "for item in data_train[:train_size]:\n",
    "    \n",
    "    item[0]=item[0]*new_std[0]+new_mean[0]\n",
    "    item[1]=item[1]*new_std[1]+new_mean[1]\n",
    "    firstday+=item[0]\n",
    "    pred.append(firstday)\n",
    "    print(np.exp(firstday))\n",
    "\n",
    "for item in reals:\n",
    "    date=len(data_train)+i\n",
    "    y_predict_max=func(date, popt_max[0],popt_max[1],popt_max[2],popt_max[3])\n",
    "    y_predict_min=(func(date, popt_min[0],popt_min[1],popt_min[2],popt_min[3]))\n",
    "    firstday+=item\n",
    "    pred.append(firstday)\n",
    "    print(np.exp(firstday))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "337\n"
     ]
    }
   ],
   "source": [
    "print(train_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "22409.63106677466 20109.7549206709 22391.220178737734 20089.079661423224\n"
     ]
    }
   ],
   "source": [
    "\n",
    "y_yester_max=func(360, popt_max[0],popt_max[1],popt_max[2],popt_max[3])\n",
    "y_yester_min=func(360, popt_min[0],popt_min[1],popt_min[2],popt_min[3])\n",
    "y_predict_max=func(361, popt_max[0],popt_max[1],popt_max[2],popt_max[3])\n",
    "y_predict_min=func(361, popt_min[0],popt_min[1],popt_min[2],popt_min[3])\n",
    "print(y_yester_max,y_yester_min,y_predict_max,y_predict_min)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "21804.02698244808\n",
      "21776.1044405374\n",
      "21748.611769191724\n",
      "21721.542347507842\n",
      "21694.889656604537\n",
      "21668.647278049768\n",
      "21642.80889231217\n",
      "21617.368277236408\n",
      "21592.319306542064\n",
      "21567.65594834566\n",
      "21543.37226370548\n",
      "21519.462405188828\n",
      "21495.920615461415\n",
      "21472.741225898448\n",
      "21449.91865521717\n",
      "21427.44740813051\n",
      "21405.322074021446\n",
      "21383.537325637837\n",
      "21362.08791780741\n",
      "21340.968686172502\n"
     ]
    }
   ],
   "source": [
    "for i in range(20):\n",
    "    y_predict_max=func(train_size+i, popt_max[0],popt_max[1],popt_max[2],popt_max[3])\n",
    "    y_predict_min=func(train_size+i, popt_min[0],popt_min[1],popt_min[2],popt_min[3])\n",
    "    print((y_predict_max+y_predict_min)/2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "bdec6ca5ce638a4776cd655f71204534c127e30145c88be198b9bebdb38fc0db"
  },
  "kernelspec": {
   "display_name": "Python 3.6.5 ('snow')",
   "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": "3.6.5"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}