Heuristics Miner for E-Commerce Visitor Access Pattern Representation

摘要

Click stream data from an e-commerce website can form a certain pattern that describes visitor behavior. This pattern can be used to determinealternative access sequence to surf the website. σ-Algorithm and Genetic Mining are two of the most common methods for pattern recognitionthat use frequent sequence item set approach. This study used heuristic miner algorithm, an advanced form of these methods, to discover thepattern of visitor behavior in e-commerce website. σ-Algorithm assumes that an activity in a website recorded in the data log is a completesequence from start to finish, without any tolerance for incomplete data or data with noise. On the other hand, Genetic Mining is a method thattolerates incomplete data or data with noise, so it can generate a more detailed e-commerce visitor access pattern. In this study, the samesequence of events was obtained from six-generated patterns. The resulting pattern describes the sequence of how visitors access the ecommercewebsite. This sequence can be used to enhance the e-commerce website based on visitor behavior.