Automatic Classification of Software Bug Reports Based on LDA and Word2Vec
Yang Tang, Hanqing Zhou, Hang Su · 2022 2nd International Conference on Computer Science, Electronic Information Engineering and Intelligent Control Technology (CEI) · 2022
Obtaining the types of software faults is of great significance for fault location and repair. Usually, manual classification methods are expensive. In this paper, we propose an automatic classification framework for Bug Reports based on LDA and Word2Vec. The feature representation of common classification methods has the problems of data sparsity and high dimensionality, so this paper adopts the feature representation method that combines LDA and Word2vec, which can represent words as low-dimensional word vectors with semantic relationships. Furthermore, to improve the quality of the classification model, we introduce the Self-Attention mechanism. The results show that the framework can automatically classify Bug Reports well.