ml.py

import pandas as pd
import numpy as np

data=pd.read_csv('train.csv')
phi_df=data[['T6']]
Y=data[['Output']]

phi=phi_df.as_matrix()

phi_t=phi.T
w=np.zeros(1)
w.shape=(1,1)

def rmse(predicted,true_vals):
    diff=predicted-true_vals
    n=len(predicted)
    err=(np.sum(diff**2)/n)**.5
    return err


t1=np.dot(phi_t,phi)
t2=np.dot(phi_t,Y)

err=[]
t=0.00000001
for i in range(50):
    del_w=np.dot(t1,w)-t2
    y1=np.dot(phi,w)
    err.append(rmse(y1,Y))
    w=w-t*del_w

print(w)

y1=np.dot(phi,w)

print(rmse(y1,Y))

import matplotlib.pyplot as plt

plt.scatter(phi,Y,color='blue')
plt.plot(phi,np.dot(phi,w),color='red')
plt.show()

ml2.py

import pandas as pd
import numpy as np

data=pd.read_csv('train.csv')
phi_df=data[['T6','T1']]
Y=data[['Output']]

phi=phi_df.as_matrix()

phi=np.hstack((phi,np.ones(len(phi)).reshape((len(phi),1))))
phi_t=phi.T
w=np.zeros(3)
w.shape=(3,1)

def rmse(predicted,true_vals):
    diff=predicted-true_vals
    n=len(predicted)
    err=(np.sum(diff**2)/n)**.5
    return err


t1=np.dot(phi_t,phi)
t2=np.dot(phi_t,Y)

err=[]
t=0.00000001
for i in range(50):
    del_w=np.dot(t1,w)-t2
    y1=np.dot(phi,w)
    err.append(rmse(y1,Y))
    w=w-t*del_w

print(w)

y1=np.dot(phi,w)

print(rmse(y1,Y))

import matplotlib.pyplot as plt


plt.plot(range(50),err,color='red')
plt.show()