Mining e-shoppers´ purchase rules based on k-trees sequential pattern

With the rapid development of online shopping, electronic commerce has offered a new channel for instant on-line shopping. It is necessary for company to on-line one-to-one market to e-shoppers. Therefore, the ability to predict e-shoppers´ purchase behavior basing on data mining has become a key source of competitive advantage for company. Frequently occurring sequential patterns, identified in sequences in a large dataset, comprise lists of subsequences. Sequential pattern mining is crucial to data mining domains. In this paper, the concept of sequence close level is proposed for counting the distance between a pair of items of k-trees pattern in a transaction sequence. This paper presents a novel data mining approach for exploring hierarchical tree structures, which includes algorithm 1 and algorithm 2. The approache not only counts the frequency of occurring patterns, but also addresses the distance between a pair of items of k-trees pattern in a transaction sequence. The better result is achieved by applying the new approache to a given database for e-shoppers.