Sequential Pattern Mining for Library System using Prefix Spanning

Sequential pattern mining is an important datamining problem with broad applications. Most of thesequential pattern mining methods, such as GSP(Generalized Sequential Pattern) and AprioriAllexplore a candidate generation and test approach toreduce the number of candidates to be examined.These approaches may not be efficient in mininglarge sequence databases having numerous patternsand/or long patterns. The better algorithm forsequential pattern is based on pattern-growth, adivide-and-conquer algorithm that projects andpartitions databases based on the currently identifiedfrequent patterns and grow such patterns to longerones using the projected databases. This paperpresents mining sequential pattern from librarydatabase by prefixSpan algorithm, which exploresprefix projection in sequential pattern mining.PrefixSpan mines the complete set of patterns butgreatly reduces the efforts of candidate subsequencegeneration. Moreover, prefix-projection substantiallyreduces the size of projected databases and leads toefficient processing. The experimental results of oursystem are also discussed in this paper