A Fast Mining and Updating Algorithm for Frequent Patterns on Biological Single Sequence

Traditional Mining Frequent Pattems algorithms will construct lots of projected databases and generate lots of patterns with short length in the process of mining which cause the low efficiency of mining. In order to overcome the shortcomings of traditional algorithms, a fast and efficient algorithm SSPM was proposed. We used longer pattems for mining, which avoided producing lots of patterns with short length. We also used prefix tree of primary patterns for frequent patterns mining and pattern growth which could avoid producing plenty of irrelevant patterns. The experimental results show that SSPM not only improves the performance but also achieves effective mining results.