Simple perceptron implementation that learns to predict the species of flower from width and length measurements.
This repository hosts two Java implementations of a perceptron - no specialised libraries are used. Net1.java uses simple feedback training to predict whether each test data point is of the species Iris-Setosa or not, while Net2.java implements Adaline error correction to try to separate the Iris-Virginica and Iris-Versicolor species.
The data used is from the Iris Data Set from the UCI Machine Learning Repository. The training data used can be found in train.txt and the test data used is in test.txt.