This is a project from the course Software Project from my third year in Tel Aviv University. This code performs a spectral k-means clustering algorithm to a list of coordinates:
In order to compile and run the code, run the following commands after cloning the repository:
gcc -ansi -Wall -Wextra -Werror -pedantic-errors spkmeans.c -lm -o spkmeans
python setup.py build_ext --inplace
If you want to run the code, create a txt/csv file with points in equal amount of coordinates in each line, separated by commas only, and run the following: The program (both on C and python) gets the following arguments:
- k - number of clusters (If k = 0, will run the eigengap heuristics algorithm to find it)
- goal - one of the following
- spk - run full spectral clustering
- wam - calculate the weighted adjacency matrix and print it
- ddg - calculate the diagonal degree matrix and print it
- lnorm - calculate lnorm matrix and print it
- jacobi - in this case, the file should be a matrix to calculate its eigenvalues and eigenvectors using Jacobi's algorithm
- file_name - the file where the data is in
Running on Python, use spkmeans.py, and running on C, use ./spkmeans