An improved parallel minimum spanning tree based clustering algorithm for microarrays data analysis

Solving duster identification problem on large amount of data ig known to be time consuming. Àlmoit au the state of art clustering techniques focuses on sequential algorithms which suffer from me problem of long runtime. So, parallel algorithms are needed. One of the attempts is a parallel minimum spanning tree (MST)-based clustering technique, called CLUMP, which identifies dense clusters in a noisy background. Although, CLUMP is efficient algorithm for clustering large data set, the MST construction is considered the time consuming phase of the algorithm. This paper presents and improved CLUMP algorithm CLUMP to enhance its speed. The experimental results showed that the proposed algorithm proved to be efficient than the original algorithm CLUMP in terms of complexity and runtime.