Data clustering in sparse subspaces using the generalization of K-means algorithm

Dictionary training can be considered as finding suitable representatives for a data group, data clustering also has a similar goal. Clustering of data based on sub-spaces that have a low rank can be interpreted as dictionary training for thin representation. In this article, an algorithm known as K-subspace is presented for clustering subspaces, and then by generalizing the K-means algorithm in higher data, we will reduce the computational load of the algorithm. Finally, the simulation will be done with synthetic data in five subspaces to check the speed and accuracy of the proposed algorithm.

Keywords:

dictionary training , K-Means algorithm , clustering