naive_bayes.py

import numpy as np

class NaiveBayesBinaryClassifier:
    
    def fit(self, X, y):
        self.y_classes, y_counts = np.unique(y, return_counts=True)
        self.phi_y = 1.0 * y_counts/y_counts.sum()
        self.phi_x = [1.0 * X[y==k].mean(axis=0) for k in self.y_classes]
        return self
    
    def predict(self, X):
        return np.apply_along_axis(lambda x: self.compute_probs(x), 1, X)
    
    def compute_probs(self, x):
        probs = [self.compute_prob(x, y) for y in range(len(self.y_classes))]
        return self.y_classes[np.argmax(probs)]
    
    def compute_prob(self, x, y):
        res = 1
        for j in range(len(x)):
            Pxy = self.phi_x[y][j] # p(xj=1|y)
            res *= (Pxy**x[j])*((1-Pxy)**(1-x[j])) # p(xj=0|y)
        return res * self.phi_y[y]
    
    def evaluate(self, X, y):
        return (self.predict(X) == y).mean()