import numpy as np class LogisticRegression: def __init__(self, x, y, m, n): self.x = x self.y = y self.m = m self.n = n self.w = None self.b = None def initialize_parameters(self): self.w = np.zeros(self.n) self.b = 0 def sigmoid(self, z): return 1 / (1 + np.exp(-z)) def compute_cost(self): z = np.dot(self.x, self.w) + self.b f = self.sigmoid(z) f = np.clip(f, 1e-15, 1 - 1e-15) # prevents log(0) crash loss = (-self.y * np.log(f)) - ((1 - self.y) * np.log(1 - f)) return np.sum(loss) / self.m def compute_grads(self): z = np.dot(self.x, self.w) + self.b f = self.sigmoid(z) error = f - self.y dj_dw = np.dot(self.x.T, error) / self.m dj_db = np.sum(error) / self.m return dj_dw, dj_db def gradient_descent(self, alpha=0.1, epochs=500, cost_display_step=100): self.initialize_parameters() for i in range(1, epochs + 1): dj_dw, dj_db = self.compute_grads() self.w = self.w - (alpha * dj_dw) self.b = self.b - (alpha * dj_db) def predict(self, x_test): # FIXED: Changed self.x to x_test z = np.dot(x_test, self.w) + self.b return self.sigmoid(z)