Rewrite the implementation on this repo https://github.com/arnandprabaswara/simple_hebb_neuron/blob/master/Hebb_Rule.ipynb

hebb_rule_implementation_rewrite.py

"""
Author		: Muhammad Arifin
Date 		: 1 May 2020
Description	: Rewrite the implementation on this repo https://github.com/arnandprabaswara/simple_hebb_neuron/blob/master/Hebb_Rule.ipynb

"""
import numpy as np 
import pandas as pd 

# Inisiasi data
df = pd.DataFrame({  'x1':[0, 0, 0], 'x2':[0, 0, 0], 'x3':[1, 1, 1],
					 'x4':[1, 1, 1], 'x5':[0, 0, 0], 'x6':[0, 0, 0],
                   	 'x7':[0, 0, 0], 'x8':[1, 1, 1], 'x9':[0, 1, 1],
                   	'x10':[0, 1, 0],'x11':[1, 0, 0],'x12':[0, 0, 0], 
                   	'x13':[0, 0, 0],'x14':[1, 1, 1],'x15':[0, 0, 0],
                   	'x16':[0, 0, 0],'x17':[1, 1, 1],'x18':[0, 0, 0], 
                   	'x19':[0, 0, 0],'x20':[1, 1, 1],'x21':[1, 1, 1],
                   	'x22':[1, 1, 1],'x23':[1, 1, 1],'x24':[0, 0, 0], 
                   	'x25':[0, 0, 0],'x26':[1, 1, 1],'x27':[0, 0, 0],
                   	'x28':[0, 0, 0],'x29':[1, 1, 1],'x30':[0, 0, 0], 
                   	'x31':[0, 0, 0],'x32':[1, 1, 1],'x33':[0, 0, 0],
                   	'x34':[0, 0, 0],'x35':[1, 1, 1],'x36':[0, 0, 0],})

# Weights and bias initialization

# Weights values stored in a dictionary for easy retrieval in
# later processing steps.
weights = { 'w1':1, 'w2':1, 'w3':0, 'w4':0, 'w5':1, 'w6':1,
            'w7':1, 'w8':0, 'w9':1,'w10':1,'w11':0,'w12':1, 
           'w13':1,'w14':0,'w15':1,'w16':1,'w17':0,'w18':1,
           'w19':1,'w20':0,'w21':0,'w22':0,'w23':0,'w24':1,
           'w25':1,'w26':0,'w27':1,'w28':1,'w29':0,'w30':1,
           'w31':1,'w32':0,'w33':1,'w34':1,'w35':0,'w36':1,}

# Bias Value
b = -1

# Creating the Neuron and do calculations
yin1 = 0
yin1 += b

# Adding yin1 function with multiplication of feature vector and 
# their corresponding weigth
for i in range(1,37):
	yin1 += df.iloc[ : , i-1] * weights["w%d" %i]

# # Sigmoid function
# # https://en.wikipedia.org/wiki/Sigmoid_function
y = 1 / (1 + np.e**(-yin1))

# Updating the weights
new_weights = {} # new weights Dictionary object

for i in range(1,37):
	# Updating the weight
	wbuffer = 0 	# temporary variable buffer
	listbuffer = [] # temporary list buffer
	wbuffer += weights["w%d" %i] + ( df.iloc[ : , i - 1] * y)
	listbuffer = [w for w in wbuffer]
	new_weights["w%d" %i] = listbuffer

# updating the bias value
new_b = b + y

#Turning weights data and new_weights data into pandas dataframe
weights_for_df = {}
for key,val in weights.items():
	weights_for_df[key] = [val]


df_weights = pd.DataFrame(weights_for_df)
df_new_weights = pd.DataFrame.from_dict(new_weights)

# Print the old and updated weight values
print("==================== Initial Weights ====================\n")
print(df_weights,"\n")
print("==================== Updated Weights ====================\n")
print(df_new_weights)