k-means-hadoop

k-means clustering with Apache Hadoop

Build

mvn clean install
hadoop jar target/k-means.jar \
            PATH_INPUT \
            PATH_OUTPUT \
            NUM_K \
            NUM_COORDINATES \
            THRESHOLD

Example

hadoop jar target/k-means.jar \
            input \
            output \
            7 \
            3 \
            0.0005

Usage

Generate the points (x;y , x;y;z etc.) for the data_set with the python script generatePoints.py (parameters, numPoints, range).

python generatePoints.py 3 5 50 # 5 3d points in -50..50 range

Output in input/data_set:

21.6041331095;8.46492932874;39.6766968839;
11.9148005581;47.8849166781;17.8483647205;
19.9629853313;2.6589522782;20.4473549181;
44.7893224782;39.1567505862;39.7058609459;
26.9080526686;24.6560481195;32.782580723;

Import as a Maven project The main class is KMeans.java and the args are:

• input folder
• output folder
• number of clusters K
• number of parameters
• threshold

Use plot2d.py or plot3d.py to plot the results

---

Example with k = 7, 1000 points

python generatePoints.py 2 1000 300

Example with k = 5, 1000 points

python generatePoints.py 3 1000 100