Software system module division method based on clustering and label propagation

Abstract

The invention discloses a software system module division method based on clustering and label propagation, which uses an outlier detection algorithm based on feature density to filter out outliers, and uses label propagation to distribute the outliers to a proper cluster, thereby overcoming the problem of semantic outliers in a traditional semantic clustering process. Different from an existing software clustering technology based on semantic or structural information, the method has the advantages that part of software entities are independently processed, firstly, outliers which interfere with the clustering process are filtered out, then, a division skeleton is constructed by using remaining high-quality semantic features, and then, the outlier entities are processed by using the structural information. According to the method, the software module division quality can be effectively enhanced, so that the software maintenance operation is optimized, and a complex software system can be understood more meticulously.

Description

technical field [0001] The invention relates to a method for re-dividing modules of an entire software system in a software maintenance process, in particular to a software system module division method based on clustering and label propagation. Background technique [0002] During the software lifecycle, the modular structure of software often deviates from the original design. The code structure of a software reflects the key design decisions taken in realizing the software's functional and non-functional requirements. During maintenance, some program modification operations (iteration of new requirements and fixing bugs) are usually performed by designated developers. However, these developers are likely to learn different design concepts from design documents. Some people even deviate from the preset code baseline to make coding easier. In fact, these maintenance activities do not always work as intended. As a result, the quality of software modularity has been decli...

AI Technical Summary

Problems solved by technology

However, semantic methods are based on the idea that entities using similar concept vocabularies should be classified into the same cluster, but they hardly consider semantic outliers in the clustering process

The presence of these outliers always reduces the accuracy of the clustering results

Images