High Dimensional Data Clustering Using Cuckoo Search Optimization Algorithm

The amount of data available over Internet and World Wide Web is increasing exponentially. Retrieving data that is more close to user’s query effectively and efficiently is a challenging task in Information Retrieval (IR) system. Clustering of Documents is one of the solutions to this. Clustering is the process of partitioning a set of objects in such a way that the objects in same cluster are more similar. The number of possible ways in which the documents can be clustered is enormous and this makes the problem to be a combinatorial optimization problem. Nature inspired algorithms are commanding tools to attack this type of problem. In this paper, an attempt has been made to use Cuckoo Search Optimization (CSO) algorithm to solve the problem of document clustering. The CSO algorithm is experimented with standard benchmark dataset, Classic4 dataset. The quality of solutions generated by CSO algorithm in terms of DB Index was compared with K-means algorithm and Ant Colony Optimization (ACO) algorithm. The results reveal that CSO algorithm is a viable to achieve world class solutions to high dimensional data clustering. Keywords—document clustering, optimization, cuckoo search, ant colony, meta heuristic