This library was originally implemented for the DIKU course Statistical Methods for Machine Learning. It contains the following algorithms:

Classification

  • K nearest neighbours
  • Linear discriminant analysis
  • Multilayer perceptron (with backpropagation)
  • Naive bayes (Also supporting different kernel estimator, like Epanechnikov and Gauassian kernels)
Attachments:
Download this file (mllib.jar)mllib.jar[Java machine learning library]80 kB
This is the source code for an implementation of a multilayer perceptron neural network in Java, using backpropagation for learning.

Main method included for solving XOR-problem to show example of use.
 
 
This algorithm has been enhanced and included in my more general ML library.
Attachments:
Access this URL (http://thiele.nu/attachments/article/4/MultiLayerPerceptron.java)MultiLayerPerceptron.java[ ]7 kB

This is a Java implementation of the AI algorithm k-nearest neighbours (KNN) used for classification.

Constructor: NearestNeighbour(ArrayList<DataEntry> dataset, int k).


Main method included to give an example of use.

 

This algorithm has been enhanced and included in my more general ML library.

Attachments:
Access this URL (http://thiele.nu/attachments/article/3/NearestNeighbour.java)NearestNeighbour.java[ ]3 kB