Mining Longest Common Subsequence and other Related Patterns using DNA Operations

Abstract

Longest Common Subsequence (LCS) and Shortest Common Subsequence (SCS) problems are to find subsequences in given sequences in which the subsequence is as long as possible and as short as possible subsequence respectively. These subsequences are not necessarily contiguous or unique. In this paper we have proposed two new approaches to find LCS and SCS, of N sequences parallely, using DNA operations. These approaches can be used to find LCS and SCS, of any window size, from any number of sequences, and from any type of input data. The proposed work can be applied to finding diverging patterns, constraint LCS, redescription mining, sequence alignment, speech recognition, find motifs in genetic data bases, pattern recognition, mine emerging patterns, contrast patterns in both scientific and commercial databases. Implementation results shown the correctness of the algorithms. Finally, the validity of the algorithms are checked and their time complexity is analyzed.