Approximation Algorithm for Constrained k-Center Clustering: A Local Search Approach

Dependencies

• Java version: 1.8.0

Data sets

• cnae (example)
• skin
• covertype

Running the Constrained Clustering Program

To execute the algorithm:

1. Compile the Java class

   javac addConstraints/OutPut.java
   

2. Run the program

   java addConstraints.OutPut
   

Note: The algorithm’s clustering parameters (e.g. number of clusters, constraint ratios, repetition ratios) are currently hard-coded in OutPut.java. You can modify those settings directly in the source before recompiling if you need to change the configuration.

// Example snippet from OutPut.java showing the settings
int[] k_vary = {10}; //number of clusters
double[] c = {0.1}; //constraint ratios
double[] c1 = {0.0}; //repetition ratios

The final results will be saved in /Results/cnae/.

Plot the output

• Measure the clustering cost and runtime of the constrained clustering problem.