Q3.py

'''
COMP 551 ASSIGNMENT 1 QUESTION 3
Name: Shatil Rahman
ID: 260606042
'''
import numpy as np
import matplotlib.pyplot as plt

def train(X,Y,Lambda):
    #Least squares training
    Xt_X = np.dot(X.T,X)
    Xt_X_inv = np.linalg.inv(np.add(Xt_X, np.eye(Xt_X.shape[0]) * Lambda))
    
    Xt_Y = np.dot(X.T, Y)
    W = np.dot(Xt_X_inv, Xt_Y)
    return W

def MSE(W,X,Y):
    #Calculate mean square error
    y_hat = np.dot(X,W)
    error = Y - y_hat 
    MSE = ((error**2).sum())/(error.size)
    return MSE
 
'''   
def k_fold_validate(W, val_sets):
    MSE_list = []
    for example in val_sets:
        MSE_list.append(MSE(W,example[0], example[1]))
    
    return MSE_list
''' 



############Load Data #####################
#train1 = np.loadtxt('Datasets/CandC-train1.csv', dtype=float, delimiter=',', ndmin=2)
#X_train1 = train1[:,:-1]
#Y_train1 = train1[:,-1].reshape(train1.shape[0],1)

examples_train = []
examples_test = []

for i in range(1,6):
    s_train = "Datasets/CandC-train" + str(i) + ".csv"
    s_test = "Datasets/CandC-test" + str(i) + ".csv"
    
    X = np.loadtxt(s_train, dtype=float, delimiter=',', ndmin=2)
    o = np.ones(X.shape[0]).reshape(X.shape[0],1)
    X = np.concatenate((o,X), axis=1)
    examples_train.append([X[:,:-1], X[:,-1].reshape(X.shape[0],1)])
    
    X2 = np.loadtxt(s_test, dtype=float, delimiter=',', ndmin=2)
    o = np.ones(X2.shape[0]).reshape(X2.shape[0],1)
    X2 = np.concatenate((o,X2), axis=1)
    examples_test.append([X2[:,:-1], X2[:,-1].reshape(X2.shape[0],1)])
    
    
##########################Training the models #################################
    

Models = []

for i in range(0,5):
    example = examples_train[i]
    test_example = examples_test[i]
    
    Models.append(train(example[0], example[1], 0.0))
    MSE_test = MSE(Models[i], test_example[0], test_example[1])
    print "MSE for Set 1: " + str(MSE_test)