Research on Program Code Mining Method
Liu Baozhong, Liu Jianbin · 2019
With the increasing amount of program source code data in the development of the open source community, there are reusable program patterns in the vast amount of code data. The existing program pattern mainly relies on manual extraction and there are problems such as high labor cost and a small number of acquired modes. This paper proposes a program code mining method for this problem. Automated mining of program pattern by this method saves labor costs and time costs and can more accurately mine more program patterns. In the mining process, the pre-processing is performed to convert the code data into a digital sequence and then the SPADE algorithm is used for automatic mining to obtain a reusable program pattern. After comparison with PrefixSpan and GSP mining algorithms, the results show that the method is effective.