# coding: utf-8

# In[113]:

import scipy.io
from sklearn.preprocessing import StandardScaler
import numpy as np
import matplotlib.pyplot as plt

mat = scipy.io.loadmat('HW10/USPS.mat.mat')

# In[114]:

data,label=mat['A'],mat['L']

# In[115]:

print("Orignial Data Shape = ",data.shape)

# In[116]:

std_1 = StandardScaler()
std_fit=std_1.fit(data)
std_X=std_fit.transform(data)

# In[117]:

mean_x = np.mean(std_X, axis=0)
cov_matrix = (std_X - mean_x).T
cov_matrix=cov_matrix.dot((std_X - mean_x)) / (std_X.shape[0]-1)
#print('Covariance matrix \n%s' %cov_mat)

# In[118]:

# cov_matrix = np.cov(std_X.T)

eigen_values, eigen_vector = np.linalg.eig(cov_matrix)

# In[120]:

eigen_pairs = [(np.abs(eigen_values[i]), eigen_vector[:,i]) for i in range(len(eigen_values))]

eigen_pairs.sort()
eigen_pairs.reverse()

print('Eigenvalues in descending order:')
for val in eigen_pairs:
print(val[0])

# In[67]:

import chart_studio.plotly as py
tot = sum(eigen_values)
var_exp = [(i / tot)*100 for i in sorted(eigen_values, reverse=True)]
cum_var_exp = np.cumsum(var_exp)

trace1 = dict(
type='bar',
x=['PC %s' %i for i in range(1,30)],
y=var_exp,
name='Individual'
)

Ensemble Machine Learning Algorithms in Weka The 5 algorithms tha...