Amelioration Design of Customer-Behavior Analyzing Engine

Abstract

This article proposes an amelioration design of Customer-Behavior analyzing engine. This engine is constructed on clicks stream analyzing method which expands ExLF log file formats and distinguishes between users with Cookie recognition mechanism and embedded Session Variable. This method also distinguishes between user-visit affairs by the time window model, and presents detailed database tables from the data source. Through clustering and CLV value analysis, it can recognize core customers. And then, the article combines clicks-stream dialog files with commercial website's inner data, and mines multidimensional associative rules by improved Apriori algorithm in order to analyze customer behavior and interests models.