SGDLinearRegression.py

#!/usr/bin/env python3

import numpy as np
import pandas as pd
from sklearn.metrics import r2_score


data_train = pd.read_csv('/datasets/train_data_n.csv')
features_train = data_train.drop(['target'], axis=1)
target_train = data_train['target']

data_test = pd.read_csv('/datasets/test_data_n.csv')
features_test = data_test.drop(['target'], axis=1)
target_test = data_test['target']

class SGDLinearRegression(object):
    def __init__(self, step_size, epochs, batch_size):
        self.step_size=step_size
        self.epochs=epochs
        self.batch_size=batch_size

    def fit(self, train_features, train_target):
        X = np.concatenate((np.ones(train_features.shape[0],1), train_features), axis=1)
        y = train_target
        w = np.zeros(X.shape[1])

        for _ in range(self.epochs):
            batches_count = X.shape[0] // self.batch_size
            for i in range(batches_count):
                begin = i * self.batch_size
                end = (i + 1) * self.batch_size

                X_batch = X[begin:end, :]
                y_batch = y[begin:end]

                gradient = 2 * X_batch.T.dot(X_batch.dot(w) - y_batch) / X_batch.shape[0]

                w -= self.step_size * gradient

        self.w = w[1:]
        self.w0 = w[0]

    def predict(self,test_features):
        return test_features.dot(self.w) + self.w0

model = SGDLinearRegression(0.01, 10, 100)
model.fit(features_train, target_train)
pred_train = model.predict(features_train)
pred_test = model.predict(features_test)
print(r2_score(target_train, pred_train).round(5))
print(r2_score(target_test, pred_test).round(5))