Simple two layer Perceptron model used on Iris data to classify Iris-setosa and Iris-versicolor. Neural network uses sigmoid function as the activation function. Backpropagation is performed with the derivative of the sigmoid.
The program performs the following steps;
- Parser function uses unseeded random sample to split a single source csv file into training and test data.
- Perceptron neural network trains over training data.
- Output R Data.frame compares predicted values against observed values.