A new fuzzy clustering-based recommendation method using grasshopper optimization algorithm and Map-Reduce

Collaborative filtering, an eminent approach of recommendation, finds the similarity in the data records to recommend the items. The performance of collaborative based recommendation depends on the effectiveness of clustering algorithm used for finding the similarity, especially in big data environment. This paper presents a novel fuzzy clustering based recommendation method using improved grasshopper optimization and MapReduce. The proposed method employs an improved variant of grasshopper optimization algorithm to find the optimal cluster centroids. Moreover, the proposed method runs in the MapReduce environment to handle big dataset. To experimentally validate the proposed variant, twenty-three benchmark functions are considered and compared against four other existing methods in terms of mean fitness value, standard deviation and Friedman test. Further, clustering efficacy of proposed method is also vindicated on four large-scale benchmark clustering dataset. Finally, to test the recommendation ability, the proposed method is tested on MovieLens dataset and results are validated using three performance parameters over the different number of clusters. The simulation clearly indicates that the proposed method can be effectively utilized to recommendations of items in a big data environment.