DocumentCode :
680731
Title :
Flexible Pattern Matching with Gap-Length and One-Off Conditions
Author :
Dan Guo ; Taining Xiang ; Xuegang Hu ; Xindong Wu
Author_Institution :
Sch. of Comput. Sci. & Inf., Hefei Univ. of Technol., Hefei, China
fYear :
2013
fDate :
4-6 Nov. 2013
Firstpage :
446
Lastpage :
453
Abstract :
This paper focuses on pattern matching with wildcard, gap-length and one-off conditions. It is difficult to achieve optimal solutions. We propose an FNP algorithm based on Free-Node Optimum Pruning. Each Free-Node set is a set of nodes labeled by the same number which appear on different layers in a directed graph structure WON-Net. Compared on biological data and artificial data, experimental results show that (1) FNP has a significant advantage with its solutions, winning above 95% of biological data among similar algorithms. There are theorems on obtaining optimal solutions by FNP. (2)FNP demonstrates an evident advantage on running time, when |Net| is large and k is small. (|Net| denotes the number of independent substructures without losing solutions in WON-Net and k is the number of Free-Node sets.)
Keywords :
bioinformatics; directed graphs; pattern matching; FNP algorithm; WON-Net; artificial data; biological data; directed graph structure; flexible pattern matching; free-node optimum pruning; gap-length; one-off conditions; Algorithm design and analysis; Educational institutions; Indexes; Parallel processing; Pattern matching; Time complexity; gap-length; one-off; optimal solution; pattern matching; wildcard;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Tools with Artificial Intelligence (ICTAI), 2013 IEEE 25th International Conference on
Conference_Location :
Herndon, VA
ISSN :
1082-3409
Print_ISBN :
978-1-4799-2971-9
Type :
conf
DOI :
10.1109/ICTAI.2013.73
Filename :
6735284
Link To Document :
بازگشت