基于最长公共子序列距离的主旨模式挖掘算法.PDF

34 14 2008 7 Vol.34 No.14 Computer Engineering July 2008 · · 2008 A TP301.6 ( 116024) 60% Motif Pattern Mining Algorithm Based on LCSS Distance FENG Lin, YU Xiao-hang, SUN Tao, SHEN Xiao, PAN Xiao-wen (Institute of University Students’ Innovation, Dalian University of Technology, Dalian 116024) AbstractAccording to the problem that existing algorithms are apt to be interfered by noise, a motif mining algorithm based on the Longest Common Subsequence(LCSS) distance is introduced. The algorithm is pruned efficiently by using the heuristic strategy based on the distance between subsequences during the search. Minimum Description Length(MDL) principle is used to calculate the weights of the unequal-length candidate sequences based on the motif patterns selected. Experiment shows the speed of the algorithm prompts at least 60% compared with that of naive algorithm. Key wordsmotif pattern; noise interference; clustering analysis; Minimum Description Length(MDL) principle 1 T 2 d 2 D (P ,Q)R P , Q 2.2 x y n m LCSS (x , y ) LCSS (x ,y ) n m 0 if n 0 or m 0 ( ⎧ ⎪ ) ⎨1+LCSS (xn−1,y m−1) if xn y m