An improved software defect prediction model based on grey incidence analysis and Naive Bayes algorithm

Article type: Research Article

Authors: Zhang, Shu | Wang, Yuhong^{; *}

Affiliations: School of Business, Jiangnan University, Wuxi, Jiangsu Province, China

Correspondence: [*] Corresponding author. Yuhong Wang, School of Business, Jiangnan University, Wuxi, Jiangsu Province 214122, China. E-mail: yuhongwang@jiangnan.edu.cn.

Abstract: This paper aims to improve the accuracy of software defect prediction by using a prediction model based on grey incidence analysis and Naive Bayes algorithm. The model employs the Naïve Bayes as the basic classifier of the software defect prediction model. The grey incidence analysis is used to analyze the relation between software modules and ideal modules. Then, the grey correlation degree is embedded into the Naive Bayes classification model as a feature attribute. According to the comparison and analysis of NASA’s public dataset, the prediction model in this paper improves the prediction accuracy.

Keywords: Naive Bayes, grey incidence analysis, software defect prediction

DOI: 10.3233/JIFS-213570

Journal: Journal of Intelligent & Fuzzy Systems, vol. 43, no. 5, pp. 6047-6060, 2022