mso1830 / k-means Goto Github PK

PROJECT TITLE: K-Means Clustering

PURPOSE OF PROJECT: Implement k-means clustering on a training data set 
                    to predict classification values of a test data set.
                    2nd Year BSc Computer Science, Machine Learning 
                    Coursework.
                    
AUTHORS: Karim Tabet, Snippets from Chris Thornton

VERSION or DATE: 1.0 25/11/2010

INSTRUCTIONS: 1) Compile both Java files 
                  (javac Data.java KMeansClustering.java)
              2) Run KMeansClustering (java KMeansClustering)
              3) After prompted, specify number of centroids to use.
------------------------------------------------------------------------
 The experiment intended to classify test data values using the data 
 mining strategy known as k-means clustering. 
 The 6 variables in the training data are used to plot the data and 
 centroids are placed to locate clumps of the 255, classified training 
 data. Once located, a classification value can be assigned to the 100, 
 test data. 
------------------------------------------------------------------------