Unsupervized_ML/PCA.py at main · StevetheGreek97/Unsupervized_ML · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
"""Importing dependencies"""
import numpy as np
import matplotlib.pyplot as plt
import plotly.express as px
import pandas as pd


class PCA():
    """
    This class performs Principle Component Analysis (PCA) on scaled data"""

    def __init__(self,X):
        """
        Initialaze
        """

        self.X = X
        self.covariance_matrix = self.cov()
        self.vals = self.eig()[0]
        self.vecs = self.eig()[1]
        self.projections = self.get_projections()

    def get_vals(self):
        """
        Returns:
                Eigenvalues
        """
        return self.vals

    def get_vecs(self):
        """
        Returns:
                Eigenvectors
        """
        return self.vecs


    def cov(self):
        """
        arr: self.X

        Returns:
                Covariance matrix
        """
        self.covariance_matrix = np.dot(self.X.T, self.X) / (self.X.shape[0] - 1)
        return self.covariance_matrix


    def eig(self):
        """
        arr:
                self.covariance_matrix

        returns:
                sorted eigenvalues highest to lowest
                sortrd eigenvectors by eigenvalues
        """
        self.vals, self.vecs = np.linalg.eig(self.covariance_matrix)

        # Sort eigenvectors and values by eigenvalues
        order = np.argsort(self.vals)[::-1]
        self.vals =self. vals[order]
        self.vecs = self.vecs[:, order]

        return self.vals, self.vecs

    def scree_plot(self):
        """
        arr:
                self.vals
                self.vecs

        returns:
                scree plot
        """
        # Scree plot  -> Kaiser rule:
        plt.bar(np.arange(len(self.vals)) + 1, self.vals, label = 'PC')

        plt.axhline(1, label = 'eigenvalues > 1')
        plt.title("Kaiser rule")
        plt.ylabel('Eigenvalues')
        plt.xlabel('No. of components')
        plt.legend()
        plt.show()

    def get_projections(self):
        """
        arr:
                self.X
                self.vecs
                self.vals

        returns:
               self.projections as pandas dataframe
        """

        pca_dict= {}
        for n_vec in range(len(self.vecs)):

            pca_dict['PC{}'.format(n_vec + 1)] = np.dot(self.X,self.vecs[:, n_vec])
            self.projections = pd.DataFrame(pca_dict)

        return self.projections

    def plot_PCA(self, x = 'PC1', y = 'PC2', color = 'PC3'):
        """
        arr:
                self.projections
        returns:
                PCA plot
        """

        fig = px.scatter(self.projections, x = x, y=y, color = color)

        fig.update_layout(
            width = 600,
            height = 600,
            title = "Priciple Component Analysis"
        )
        fig.show()

def main():
    """
    main function
    """
    print('PCA module')

if __name__ == "__main__":
    main()