Neural network forward pass with data generation and matrix implementation, one hidden layer

2022_march_21_simpleNN.py

# -*- coding: utf-8 -*-
"""
@author: Balaji
"""
print('\n 18 March Program for a simple neural network, doing a forward pass with one layer')
print('\n 21 March Added provision for data generation for use in training.')
print('\n 21 March Formating input data as a matrix for faster multiplication with weights')

import numpy as np
import  matplotlib.pyplot as plt
num_neurons     = 4;
num_pixels      = 2;
num_images      = 3;
np.random.seed(1); #For repeatability
# Input data set properties
slope           = np.tan(np.deg2rad(30));
intercept       = 50;

print('\n Creating input data:')
inputs  = np.zeros((num_images,num_pixels));
x = np.arange(0,num_pixels);
for dataSet_iter in range(num_images):
    inputs[dataSet_iter,:] = (slope*x+intercept)+ intercept*0.2*(np.random.randn(1,x.size) );

print('\n Plotting input data set :')    
plt.plot(x,inputs.transpose(),'o')
plt.xlabel('Pixel number')
plt.ylabel('Pixel value')
plt.title('Input data set')
plt.legend()
plt.show()

print('\n Each row is a data set')
weights = np.random.randn(num_neurons ,num_pixels);
bias    = np.random.randn(num_neurons ,1);
results = np.zeros_like(bias)

print('\n Input = ')
print(inputs)

print('\n Weights = ')
print(weights)

print('\n Bias = ')
print(bias)

# Output computation in matrix format
results = np.dot(weights,inputs.transpose())+bias;

# Output  computation in loop method
for input_iter in inputs:
    for weight_iter,bias_iter,result_iter in zip(weights,bias,results):
        print('\n Input data set:')
        print(input_iter);
        print('\n is multiplied with weights')
        print(weight_iter)
        print('\n and added with bias')
        print(bias_iter)
        print('\n to give the result of neuron as')
        result_iter = input_iter.dot(weight_iter) + bias_iter;
        print(result_iter)