A New Approach to Clustering Clickstream Data using Fuzzy Inference Engine

Nowadays, one of the most fundamental concerns for businesses is the Recognition of customers requirements and interests. When the volume, variety, and production speed of data increase, big data tools and platforms become a necessity, given that their power and high accuracy in data analysis and processing. Spark platform have proven that it can be a better choice than other platforms in terms of power and speed in processing performance. In this study, to identify the customers behavior patterns in the web environment, the data set a software online store including information of customer s clickstream data during a three-year period, were analyzed. We proposed a fuzzy inference engine which combines two variables and outputs new single variable which can be used as the input for the clustering model. Then, the k- medoids clustering algorithm have been employed for the fuzzy engine output and the result of clustering was compared with the condition in which only one variable is considered for clustering. The results indicate the superiority of the proposed model compared to the previous ones.

Keywords:

Online store , customers’ behavior patterns , big data , spark platform , fuzzy inference engine , k-medoids clustering algorithm