Classifier

classify.py

'''
Created on Jul 15, 2019

@author: Jesus Doménech
@email jdomenec@ucm.es
'''

# TO INSTALL: run following commands
# pip3 install -U sklearn numpy
# - if you want to plot you will also need matplotlib
# pip3 install -U matplotlib

# TO RUN: go to the bottom of the file

import numpy as np
from sklearn.svm import SVC, LinearSVC
BAD_CLASS = -1000
GOOD_CLASS = 1000


def split(good_points, bad_points):
    bad_y = [BAD_CLASS] * len(bad_points)
    good_y = [GOOD_CLASS] * len(good_points)
    X = np.array(bad_points + good_points)
    Y = np.array(bad_y + good_y)
    # Choose which classifier you want to use.
    ###  1 - Perceptron
    # clf = Perceptron()
    ###  2 - linear
    # clf = LinearSVC()
    ###  3 - SVM
    clf = SVC(kernel="linear")
    clf.fit(X, Y)
    line = []
    try:
        line.append(clf.intercept_[0])
        for i in range(len(clf.coef_[0])):
            line.append(clf.coef_[0][i])

        # You can call plot to print the line and the good and bad points
        # it only works if you are using dimension 2
        plot(clf, X, Y)
        return line
    except TypeError:
        return None


def plot(classifier, X, Y):
    if len(X) == 0 or len(X[0]) != 2:
        return
    import matplotlib.pyplot as plt
    w = classifier.coef_[0]
    if w[1] == 0:
        yy = np.linspace(min([Xi[1] for Xi in X]) - 2, max([Xi[1] for Xi in X]) + 2)
        # yy_down = yy_up = yy
        a = 0
        xx = [- (classifier.intercept_[0]) / w[0]] * len(yy)
    else:
        xx = np.linspace(min([Xi[0] for Xi in X]) - 2, max([Xi[0] for Xi in X]) + 2)
        # xx_down = xx_up = xx
        a = -w[0] / w[1]
        yy = a * xx - (classifier.intercept_[0]) / w[1]
    plt.plot(xx, yy, 'k-')
    plt.scatter(X[:, 0], X[:, 1], c=Y)
    plt.axis('tight')
    plt.show()


class Perceptron(object):
    """Perceptron classifier.

    Parameters
    ------------
    eta : float
        Learning rate (between 0.0 and 1.0)
    n_iter : int
        Passes over the training dataset.

    Attributes
    -----------
    w_ : 1d-array
        Weights after fitting.
    errors_ : list
        Number of misclassifications in every epoch.

    """
    def __init__(self, eta=0.2, n_iter=100):
        self.eta = eta
        self.n_iter = n_iter

    def fit(self, X, y):
        """Fit training data.

        Parameters
        ----------
        X : {array-like}, shape = [n_samples, n_features]
            Training vectors, where n_samples is the number of samples and
            n_features is the number of features.
        y : array-like, shape = [n_samples]
            Target values.

        Returns
        -------
        self : object

        """

        self.w_ = np.zeros(1 + X.shape[1])
        self.errors_ = []

        for __ in range(self.n_iter):
            errors = 0
            for xi, target in zip(X, y):
                update = self.eta * (target - self.predict(xi))
                self.w_[1:] += update * xi
                self.w_[0] += update
                errors += int(update != 0.0)
            self.errors_.append(errors)
        self.coef_ = [list(self.w_[1:])]
        self.intercept_ = [-self.w_[0]]
        return self

    def net_input(self, X):
        """Calculate net input"""
        return np.dot(X, self.w_[1:]) + self.w_[0]

    def predict(self, X):
        """Return class label after unit step"""
        return np.where(self.net_input(X) >= 0.0, 1, -1)


# DON'T REMOVE THIS LINE
if __name__ == '__main__':
    # This lines will be run we you execute the command:
    # python3 classify.py

    # 1 - Generate good and bad points
    goods = [[i, -10] for i in range(1000)]
    bads = [[i, 10] for i in range(1000)]
    # 2 - Call split
    line = split(goods, bads)
    # 3 - Analyze the output
    # The first position is the independent term.
    # The rest are the coeffs
    print(line)