An effective sequence clustering algorithm for checking software fault feature

Software security becomes increasingly important recently in the application of software. However, existing sequences clustering algorithms directly applied to software security area have got undesirable results. In order to improve the cluster quality and the time complexity of the software fault feature, we propose a new similarity method and a sequence clustering algorithm called SCA (Sequence Clustering Algorithm). The number of common sequence elements contained in software fault feature sequences is calculated to measure the relationship among sequences. And the similarity method also monitors the degree of normalization of fault feature sequences to get more accurate cluster results. Sequences are collected into clusters by this similarity metric. Experimental results on the synthetic data have shown that our algorithm has the higher cluster quality and lower time complexity.