Mining the Change of Customer Behavior in Fuzzy Time-Interval Sequential Patterns with Aid of Similarity Computation Index (SCI) and Genetic Algorithm (GA)

Abstract

Sequential Pattern Mining is closely related to the concept of Data Mining. It is the process of discovering the frequent sequential patterns in a given database. Even though, various algorithms and techniques are used in the process of sequential pattern mining. One of the Mine Fuzz Change model was perform the Sequential Pattern Mining process by using SCI (Similarity Computation Index) and successfully perform the pattern classification process. But this model has the drawback in the SCI computation. Because, the SCI value is computed by using the raw data which is collected from different time interval and it creates the time complexity in the pattern mining process. To solve this drawback, in this paper, an optimized fuzzy time interval sequential pattern mining algorithm is proposed. The proposed method finds the customer behavior changes in the fuzzy time interval sequential patterns by exploiting the optimized FTI algorithm and the patterns are classified based on their support and SCI (Similarity Computation Index) value. Our proposed method performance is evaluated by conducting different experiments on the synthetic dataset. Moreover, our proposed method performance is compared with the existing Mine Fuzz Change model. The performance results shows that our proposed method reduces the time complexity and it will helps managers to understand the changing behaviors of their customers