NN-Sample-01

Multiclass Classification

Description

A simple classification problem of data with 5 classes using a neural network with only one layer and sigmoid activation. Network performance is shown below for 4 and 10 epochs.

Sample Output after 4 Epochs:

Sample Output after 10 Epochs:

• The data points and their labels are loaded from Points-35.dbs
• Neural Network topology, optimizer, number of layers and layer parameters are saved in NN-Model-Points.snn

Setup

Create an Empty Project using Visual Studio and copy all the files from NN-Sample-01 to the project.

And follow Project Setup to run the program.

Required Header Files

• SepMath.h
• NeuralNetwork.h
• Layer.h
• Dir.h
• Util.h