A hybrid clustering method based on k-means algorithm

We propose a novel hybrid algorithm that effectively combines K-means clustering and hierarchical and uses triangle inequality to accelerate the clustering speed. The HTK clustering algorithm can produce the same results as the standard K-means clustering algorithm. The proposed algorithm is superior to the standard K-means clustering algorithm in terms of running time and memory usage, thus improving the clustering speed and time complexity of the algorithm. The proposed clustering methods are tested on sci-kit learn datasets, and they are more favorable than the random restart K-means algorithm.