Relevant Tiws Pattern Mining With Reduced Search Space

Abstract

The sequential pattern mining finds all sequential patterns which occurs frequently for a given sequence database whose frequency is no less than the threshold. The weighted sequential pattern mining aims to find more interesting sequential patterns by considering the different significance of each data element in a sequence database. In the conventional weighted sequential pattern mining, pre-assigned weights of data elements are used to get the priorities which are derived from their quantitative information. Generally in sequential pattern mining, the generation order of data is considered to find sequential patterns. However, generation time and time-intervals between the data are also important in real world application domains. Therefore, time-intervals information of data elements can be helpful in finding more interesting sequential patterns. The proposed system presents a framework for finding time-interval weighted sequential (TiWS) patterns in a sequence database. In addition, the CloSpan algorithm is used for mining TiWS patterns in a sequence database which reduces the multiple database scans and the processing time. This is useful in application fields such as web access pattern analysis, customer purchase pattern analysis, DNA sequence analysis, etc.