Clustering Evaluation

https://nlp.stanford.edu/IR-book/html/htmledition/evaluation-of-clustering-1.html

Python NLTK Resource 'tokenizers/punkt/english.pickle' not found

import nltk

nltk.download('punkt')

jupyter 500 internal error

Solution
sudo pip install --upgrade "ipython[all]"

How to Run Parallel Data Analysis in Python using Dask Dataframes

Here is the link

https://towardsdatascience.com/trying-out-dask-dataframes-in-python-for-fast-data-analysis-in-parallel-aa960c18a915

Python dataframe remove non ascii words

df.text.replace({r'[^\x00-\x7F]+':''}, regex=True, inplace=True)

text is the column name in dataframe df

Figures settings in matplotlib

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import os
from matplotlib.ticker import NullFormatter
#reading the data
os.chdir('E:/backup folder/paper')
test_data = pd.read_csv('keygenerationtime.csv')
print(test_data['Users'])
Users= test_data['Users']
Time=test_data['Time']
N = len(Users)
x2 = np.arange(N)
print(N)
plt.plot(x2, Time,linewidth=2,color='black')
plt.xticks(x2, Users)
plt.ylabel('Time(secs)',fontsize=14)
plt.xlabel('Number of Users',fontsize=14)
plt.subplots_adjust(bottom=0.2)
fig = plt.gcf()
fig.set_size_inches(5, 5)
fig.savefig('Fig3.png', dpi=200)
plt.show()

#----------------------------------------------

test_data = pd.read_csv('encryptiontime.csv')
print(test_data['File'])
File= test_data['File']
Time=test_data['Time']
N = len(File)
x2 = np.arange(N)
#print(N)
plt.bar(x2, Time,color='black',width=0.5)
plt.xticks(x2, File)
plt.ylabel('Time(secs)',fontsize=14)
plt.xlabel('File Size (KB)',fontsize=14)
plt.subplots_adjust(bottom=0.2)
fig = plt.gcf()
fig.set_size_inches(5, 5)
fig.savefig('Fig4.png', dpi=200)
plt.show()

#--------------------------------------------
test_data = pd.read_csv('decryptiontime.csv')
print(test_data['File'])
File= test_data['File']
Time=test_data['Time']
N = len(File)
x2 = np.arange(N)
#print(N)
plt.bar(x2, Time,color='black',width=0.5)
plt.xticks(x2, File)
plt.ylabel('Time(secs)',fontsize=14)
plt.xlabel('File Size (KB)',fontsize=14)
plt.subplots_adjust(bottom=0.2)
fig = plt.gcf()
fig.set_size_inches(5, 5)
fig.savefig('Fig5.png', dpi=200)
plt.show()

#------------------------------------------
test_data = pd.read_csv('turnaround.csv')
print(test_data)
File= test_data['File']
Ses_up=test_data['a']
up1=test_data['b']
up2=test_data['c']
Ses_down=test_data['d']
down1=test_data['e']
down2=test_data['f']

patterns = [ "/" , "\\" , "|" , "-" , "+" , "x", "o", "O", ".", "*" ]
width = 0.15

pos = list(range(len(Ses_up)))
fig, ax = plt.subplots(figsize=(5,5))
bar1=plt.bar(pos, Ses_up, width,
alpha=1,
color='black',
hatch="/", # this one defines the fill pattern
label=File[0])
plt.bar([p + width for p in pos], up1, width,
alpha=0.5,
color='w',
hatch="\\",
label=File[1])
plt.bar([p + width*2 for p in pos], up2, width,
alpha=0.5,
color='k',
hatch='-',
label=File[2])

plt.bar([p + width*3 for p in pos], Ses_down, width,
alpha=0.7,
color='black',hatch="//",
label=File[3])
plt.bar([p + width*4 for p in pos], down1, width,
alpha=0.5,
color='w',
hatch="...",
label=File[4])
plt.bar([p + width*5 for p in pos], down2, width,
alpha=0.5,
color='white',
hatch="///",
label=File[5])


ax.set_ylabel('Time (secs)',fontsize=14)
ax.set_xlabel('File Size (KB) ',fontsize=14)
#ax.set_title('Grouped bar plot')
ax.set_xticks([p + 1.5 * width for p in pos])
ax.set_xticklabels(File)
plt.legend(['SeSPHR T-up', '[14] T-up', '[27] T-up', 'SeSPHR T-down', '[14] T-down', '[27] T-down'], loc='upper left')
fig.savefig('Fig6.png', dpi=200)

plt.show()

Softmax vs Sigmoid function usage in last dense layer

If we require that sum of probabilities of all the classes should be equal to 1 then we use softmax but if we require separate and individual probabilities, such as in the case of multi-label classification we use sigmoid

https://www.depends-on-the-definition.com/classifying-genres-of-movies-by-looking-at-the-poster-a-neural-approach/