PrefixSpan Algorithm for Finding Sequential Pattern with Various Constraints

Abstract

PrefixSpan (Prefix-projected Sequential pattern mining) algorithm is very well known algorithm for sequential data mining. It extracts the sequential patterns through pattern growth method. The algorithm performs very well for small datasets. As the size of datasets increases the overall time for finding the sequential patterns also get increased. The PrefixSpan algorithm is run on different datasets and results are drawn based on minimum support value. One new parameter maximum prefix length is also considered while running the algorithm. Through maximum prefix length parameter the length of prefix pattern is set which is helpful for running the algorithm on large datasets. The paper also shows the variation in time complexity and memory utilization while running the algorithm on different size of input sequential datasets.