Skip to content

Latest commit

 

History

History
11 lines (6 loc) · 763 Bytes

README.md

File metadata and controls

11 lines (6 loc) · 763 Bytes

Normalized Spectral Clustering Algorithm

2022, Software Project Course

Tel Aviv University

• An algorithm which groups datapoints that are retrieved from matrices into clusters by using centroids (mean value of the datapoints in the cluster) for each cluster.

• The main algorithm uses mathematical algorithms: creation of adjacency matrix, diagonalization to find eigenvectors, arranging the eigenvectors in a matrix as columns and renormalizing. A clustering algorithm (K-means) runs on the last matrix. The algorithm adjusts the centroids each iteration and repeats itself until convergence.

• Using Python C-API to implement C code for the main computational work of the algorithm in the Python code. working in a pair, coordinated using GitHub.