It is a significant work to analyze the software behavior. Finding high utility patterns from software executing network is useful to comprehend the software. Compared with frequent patterns mining, the concept of utility provides more reliable and available knowledge. However, it is difficult for users to determine a proper minimum threshold to predict the exact number of patterns mined by the threshold. To address the issue, a novel algorithm for mining top-k high utility traversal patterns

more »

... amed THTP-Miner) from software dynamic executing graphs is proposed, where k is the desired number of patterns. In THTP-Miner, software executing traversal sequences are extracted from software executing graphs firstly. Secondly, a structure local information list of pattern is designed to store the useful information. Thirdly, three effective strategies are introduced to resolve the issue of efficiency, including two methods for raising the threshold and one pruning for filtering unpromising patterns. Comprehensive experimental results show that THTP-Miner has excellent performance.