Scribd
Upload
9 views5 pages

AIML

This document discusses using principal component analysis (PCA) for dimensionality reduction. It shows how to calculate the principal components and variance explained by each component for a sample dataset. Standardization, calculation of the covariance matrix, eigendecomposition, and transformation of the data onto the principal components are demonstrated.

Uploaded by

aditya.2133e
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
9 views5 pages

AIML

This document discusses using principal component analysis (PCA) for dimensionality reduction. It shows how to calculate the principal components and variance explained by each component for a sample dataset. Standardization, calculation of the covariance matrix, eigendecomposition, and transformation of the data onto the principal components are demonstrated.

Uploaded by

aditya.2133e
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

2/2/24, 1:30 PM Dimentionality Reduction Using PCA - Jupyter Notebook

In [5]: # import the liabaries


import numpy as np
import pandas as pd
import seaborn as sns
from matplotlib import pyplot as plt

In [7]: # creating the dataset


x= np.array ([[4,11],[8,4],[13,5],[7,14]])
x

Out[7]: array([[ 4, 11],


[ 8, 4],
[13, 5],
[ 7, 14]])

In [8]: x. shape

Out[8]: (4, 2)

In [10]: # scatter plot


plt.scatter(x[:,0],x[:,1])
plt. xlabel("x1")
plt. ylabel ("x2")
plt. show()

step-1:-standardization of data
In [11]: # calculation of mean values
xbar= x.mean(axis=0)
xbar

Out[11]: array([8. , 8.5])

localhost:8888/notebooks/Dimentionality Reduction Using PCA.ipynb 1/5


2/2/24, 1:30 PM Dimentionality Reduction Using PCA - Jupyter Notebook

In [12]: sigma=np.std(x,axis=0,ddof=1)
sigma

Out[12]: array([3.74165739, 4.79583152])

In [13]: z= (x-xbar)/sigma
z

Out[13]: array([[-1.06904497, 0.52128604],


[ 0. , -0.93831486],
[ 1.33630621, -0.72980045],
[-0.26726124, 1.14682928]])

In [14]: plt.scatter(z[:,0],z[:,1])
plt. xlabel("x1")
plt. ylabel ("x2")
plt. show()

In [15]: np.round(z.mean(axis=0),2)

Out[15]: array([-0., 0.])

In [16]: z.std (axis=0,ddof=1)

Out[16]: array([1., 1.])

step-2: calculation of covariance matrix


In [19]: cov_matrix= np.cov(x,rowvar=False)
cov_matrix

Out[19]: array([[ 14., -11.],


[-11., 23.]])

step 3:- calculation of Eigen values

localhost:8888/notebooks/Dimentionality Reduction Using PCA.ipynb 2/5


2/2/24, 1:30 PM Dimentionality Reduction Using PCA - Jupyter Notebook

In [20]: eigen_values,eigen_vector=np.linalg.eigh(cov_matrix)

In [21]: eigen_values

Out[21]: array([ 6.61513568, 30.38486432])

step-4: calculation of Eigen Vectors


In [24]: eigen_vector

Out[24]: array([[-0.83025082, -0.55738997],


[-0.55738997, 0.83025082]])

In [25]: sorted_index=np.argsort(eigen_values)[::-1]
sorted_index

Out[25]: array([1, 0], dtype=int64)

In [26]: sorted_eigenvector=eigen_vector[:,sorted_index]
sorted_eigenvector

Out[26]: array([[-0.55738997, -0.83025082],


[ 0.83025082, -0.55738997]])

In [29]: sorted_eigenvector= -1*sorted_eigenvector


sorted_eigenvector

Out[29]: array([[ 0.55738997, 0.83025082],


[-0.83025082, 0.55738997]])

step5- Calculation of Principle Component


In [30]: from sklearn.decomposition import PCA
pca=PCA(n_components=2)

In [32]: data_pca=pca.fit_transform(x)
data_pca

Out[32]: array([[-4.30518692, 1.92752836],


[ 3.73612869, 2.50825486],
[ 5.69282771, -2.20038921],
[-5.12376947, -2.23539401]])

In [33]: pca.explained_variance_

Out[33]: array([30.38486432, 6.61513568])

In [35]: explained_variance=(pca.explained_variance_ratio_)*100
explained_variance

Out[35]: array([82.12125493, 17.87874507])

localhost:8888/notebooks/Dimentionality Reduction Using PCA.ipynb 3/5


2/2/24, 1:30 PM Dimentionality Reduction Using PCA - Jupyter Notebook

In [36]: plt.bar(x=range(1,3),height=explained_variance)
plt.xlabel("Number of component")
plt.ylabel("variance percentage")
plt.title('variance of pc')
plt.show()

In [37]: plt.figure(figsize=(8,6))
plt.scatter(data_pca[:,0],data_pca[:,1])
plt.xlabel("pc1")
plt.ylabel("pc2")
plt.show()

In [ ]: ​

In [ ]: ​

localhost:8888/notebooks/Dimentionality Reduction Using PCA.ipynb 4/5


2/2/24, 1:30 PM Dimentionality Reduction Using PCA - Jupyter Notebook

In [ ]: ​

In [ ]: ​

localhost:8888/notebooks/Dimentionality Reduction Using PCA.ipynb 5/5

You might also like