IJSRP, Volume 5, Issue 5, May 2015 Edition [ISSN 2250-3153]
Ms. V. V. Kamble, Mr. S. V. Kamble
Abstract:
Frequent pattern mining from sequential datasets is an important data mining method. It has various applications like discovery of motifs in DNA sequences, financial industry, the analysis of web log, customer shopping sequences and the investigation of scientific or medical processes etc. Motif mining requires efficient mining of approximate patterns that are contiguous. The main challenge in discovering frequently occurring patterns is to allow for some noise or mismatches in the matching process. Existing algorithms focus on mining subsequences but very few algorithms find approximate pattern mining. In this paper we have presented a new method for finding frequently occurring approximate sequences from sequential datasets. Proposed method uses suffix trees for discovering frequent pattern with fixed length, maximum distance & minimum support.