TY - GEN
T1 - Software fault feature clustering algorithm based on sequence pattern
AU - Ren, Jiadong
AU - Hu, Changzhen
AU - Wang, Kunsheng
AU - Zhang, Dongmei
PY - 2009
Y1 - 2009
N2 - Software fault feature analysis has been the important part of software security property analysis and modeling. In this paper, a software fault feature clustering algorithm based on sequence pattern (SFFCSP) is proposed. In SFFCSP, Fault feature matrix is defined to store the relation between the fault feature and the existing sequence pattern. The optimal number of clusters is determined through computing the improved silhouette of fault feature matrix row vector, which corresponds to the software fault feature. In the agglomerative hierarchical clustering phase, entropy is considered as the similarity metric. In order to improve the time complexity of the software fault feature analysis, the fault features of the software to be analyzed are matched to each centroid of clustering results. Experimental results show that SFFCSP has better clustering accuracy and lower time complexity compared with the SEQOPTICS.
AB - Software fault feature analysis has been the important part of software security property analysis and modeling. In this paper, a software fault feature clustering algorithm based on sequence pattern (SFFCSP) is proposed. In SFFCSP, Fault feature matrix is defined to store the relation between the fault feature and the existing sequence pattern. The optimal number of clusters is determined through computing the improved silhouette of fault feature matrix row vector, which corresponds to the software fault feature. In the agglomerative hierarchical clustering phase, entropy is considered as the similarity metric. In order to improve the time complexity of the software fault feature analysis, the fault features of the software to be analyzed are matched to each centroid of clustering results. Experimental results show that SFFCSP has better clustering accuracy and lower time complexity compared with the SEQOPTICS.
KW - Entropy
KW - Improved silhouette
KW - Sequence pattern
KW - Software fault feature clustering
UR - http://www.scopus.com/inward/record.url?scp=71549143176&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-05250-7_46
DO - 10.1007/978-3-642-05250-7_46
M3 - Conference contribution
AN - SCOPUS:71549143176
SN - 3642052495
SN - 9783642052491
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 439
EP - 447
BT - Web Information Systems and Mining - International Conference, WISM 2009, Proceedings
T2 - International Conference on Web Information Systems and Mining, WISM 2009
Y2 - 7 November 2009 through 8 November 2009
ER -