q1_q2_q3_test.py

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from linearRegression.linearRegression import LinearRegression
from metrics import *
import sys

np.random.seed(42)

N = 30
P = 5
X = pd.DataFrame(np.random.randn(N, P))
y = pd.Series(np.random.randn(N))


for fit_intercept in [True, False]:
    LR = LinearRegression(fit_intercept=fit_intercept)
    # LR.fit_vectorised(X, y) # here you can use fit_non_vectorised / fit_autograd methods
    # LR.fit_non_vectorised(X, y, lr=0.0001)
    # LR.fit_normal(X,y)

    if (sys.argv[1]=="normal"):
        LR.fit_normal(X,y)
    elif (sys.argv[1]=="vectorised"):
        LR.fit_vectorised(X,y,batch_size=1,lr=0.001,n_iter=10000)
    elif (sys.argv[1]=="non_vectorised"):
        LR.fit_non_vectorised(X,y,batch_size=5,lr=0.0001,n_iter=1000)
    elif (sys.argv[1]=="autograd"):
        LR.fit_autograd(X,y,batch_size=1,lr=0.001,n_iter=10000)
    else:
        LR.fit_normal(X,y)
    
    
    y_hat = LR.predict(X)

    print('RMSE: ', rmse(y_hat, y))
    print('MAE: ', mae(y_hat, y))