KMeans is a windows console application to find clusters in data with any number of columns ( attributes ) using the KMeans algorithm ( https://en.wikipedia.org/wiki/K-means_clustering )
Usage: KMeans <data dimension> <data file path> <number of clusters>
The iris data set ( http://archive.ics.uci.edu/ml/datasets/Iris ) has 4 attributes
1. sepal length in cm 2. sepal width in cm 3. petal length in cm 4. petal width in cm
The application outputs
KMeans.exe 4 ..\data\bezdekIris.data 3
Total distance 103.982
Total distance 97.249
Total distance 97.2046Cluster 0 means 5.006 3.428 1.462 0.246
mins: 4.3, 2.3, 1, 0.1,
maxs: 5.8, 4.4, 1.9, 0.6,
sds: 0.348947, 0.375255, 0.171919, 0.104326,Cluster 1 means 5.90161 2.74839 4.39355 1.43387
mins: 4.9, 2, 3, 1,
maxs: 7, 3.4, 5.1, 2.4,
sds: 0.462633, 0.293885, 0.504774, 0.295091,Cluster 2 means 6.85 3.07368 5.74211 2.07105
mins: 6.1, 2.5, 4.9, 1.4,
maxs: 7.9, 3.8, 6.9, 2.5,
sds: 0.48761, 0.28625, 0.482118, 0.276165,
I have used this code to locate servers among clients to minimize the total distance between clients and their nearest server, when there was too many clients to calculate the absolute optimal locations directly.
The open source code for this can be seen at https://github.com/JamesBremner/KMeans
::raven::set 