Developing efficient algorithms for incremental mining of sequential patterns

Abstract

Sequential pattern mining is used in spectrum of areas, as financial data analysis of banks and other financial institutions, retail industry, customer shopping history, goods transportation, consumption and services, telecommunication industry, biological data analysis, scientific applications, network intrusion detection. The sequential pattern mining algorithms, in general, result in enormous number of interesting and uninteresting patterns. Therefore, the miner may have difficulty in selecting the interesting/significant ones. The core idea behind this study is the incorporation of specific constraints into the mining process. One another important concern of sequential pattern mining methods is that they are based on assumption that the data is centralized and static. For dynamic data addition and deletion in the databases, these methods waste computational and input/output (I/O) resources. It is undesirable to mine sequential patterns from scratch each time when a small set of sequences grow, or when some new sequences are added, deleted or updated in the database.