Volume: 19 Issue-01 (January-June) 2024

---

DEEP LEARNED MAPREDUCE MEAN SHIFT CLUSTERING FOR BIG E-COMMERCE DATA ANALYTICS

ABSTRACT

Clustering of big data has attained greater significance recently. Few research works have been developed in existing for grouping similar data. But, clustering accuracy using conventional algorithm was not adequate when taking big e-commerce data as input. In addition, time complexity of big data clustering was also minimal. In order to resolve such limitations, Deep Learned MapReduce Mean Shift Clustering (DLMMSC) technique is introduced. The proposed DLMMSC technique contains three layers such as input, hidden and output layers to efficiently group the large volume of data in given dataset with higher accuracy and minimal time. The DLMMSC technique initially gets the big data as input in the input layer and sent it to the hidden layer. In DLMMSC technique, three hidden layers are used in order to deeply analyze the input big e-commerce data by applying the MapReduced Mean Shift Clustering concepts. Through a deep analysis, DLMMSC technique accurately groups the similar e-commerce data together into different clusters. At last, output layer produces the optimal clustering result of big e-commerce results. By the effectual clustering of data, DLMMSC technique increases the big e-commerce data analytics performance as compared to state-of-the-art works. The DLMMSC technique conducts the experimental evaluation using parameters such as clustering accuracy, time complexity and error rate and space complexity. The experimental result shows that the DLMMSC technique is able to improve the clustering accuracy and also minimizes the time complexity of big e-commerce data analytics when compared to state-of-the-art works.

FULL TEXT

PDF