Bagging Algorithm Implementation

bagging.py

import copy
import numpy as np


class Bagging:
    num_base = 0
    classifiers = []
    minvs = []
    maxvs = []

    def __init__(self, classifier, num_base=0):
        self.num_base = num_base
        self.classifiers.clear()
        for i in range(0, self.num_base):
            self.classifiers.append(copy.copy(classifier))

    def fit(self, X, y):
        for i in range(0, self.num_base):
            sample_X, sample_y = self.boostrap(X, y)
            self.minvs.append(np.max(sample_X, axis=0))
            self.maxvs.append(np.min(sample_X, axis=0))
            self.classifiers[i].fit(sample_X, sample_y)

    def predict(self, x):
        arr = np.array([w.predict(x) for w in self.classifiers])
        (values, counts) = np.unique(arr, return_counts=True)
        ind = np.argmax(counts)
        return values[ind]

    @staticmethod
    def boostrap(X, y):
        assert len(X) == len(y)
        m = len(X)
        sample_X = np.empty([0,2])
        sample_y = np.array([])
        for i in range(0, m):
            sample_X = np.vstack([sample_X, X[np.random.randint(0, m)]])
            sample_y = np.append(sample_y, y[np.random.randint(0, m)])
        return sample_X, sample_y