Skip to content

A neural network implementation in python using matrix operations with gradient descent as backpropogation algorithm. Neural network trained with a math formula and compared with target values and Keras Model optimized by sgd (Stochastic Gradient Descent)

Notifications You must be signed in to change notification settings

DevMilk/PythonNeuralNetworkNumpy

Repository files navigation

Visualization of DataSet, x and y as inputs, z as target outputs:

alt text

Error over per Iteration when Learning Rate= 0.1 and Error Rate=0.1:

alt text

Error over per Iteration when Learning Rate= 0.05 and Error Rate=0.01

alt text

Error over per Iteration when Learning Rate= 0.01 and Error Rate=0.05

alt text

Comparison between target values vs predicted values (Orange=target, Blue= Predicted)

When Learning Rate= 0.1 and Error Rate=0.1:

alt text

When Learning Rate= 0.05 and Error Rate=0.01:

alt text

When Learning Rate= 0.01 and Error Rate=0.05:

alt text

Comparison Between Keras Neural Network model optimized with sgd algorithm and neural network model i created by numpy

alt text

About

A neural network implementation in python using matrix operations with gradient descent as backpropogation algorithm. Neural network trained with a math formula and compared with target values and Keras Model optimized by sgd (Stochastic Gradient Descent)

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages