Comparative Study Of Quality Measures Of Sequential Rules For The Clustering Of Web Data

To exploit large databases in the Web, data mining techniques have been applied. Among these techniques, the cluster analysis and the extraction of sequential patterns are considered to be the most important aspects in the process of exploring the web to find large groups. Web data that we handle are streams of sequential data where time plays a vital role in sequential patterns found to extract sequential rules. In this case, the ordering of events must be taken into account in the measure of calculation in order to measure the quality and interest of a rule. The purpose of this study is to construct a model of clustering based on the grouping of sequential rules by quality measures. We aim at the end of our study to detect a good measure of applicable data quality and provide a good partitioning through the measures evaluation of the clustering quality.