Using neural network models in the quality management system for the software defect prediction

International Workshop "Advanced Technologies in Material Science, Mechanical and Automation Engineering – MIP: Engineering – 2019"

材料科学;机械制造;原子能学

Danilov, A.D.^1 ; Samotsvet, D.A.^1 ; Mugatina, V.M.^1

Voronezh State Technical University, Moscow Ave. 14, Voronezh

394026, Russia^1

关键词: Code parameters;    Fault prediction;    Independent variables;    Neural network model;    Quality management systems;    Regression testing;    Software defect prediction;    System development;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/537/4/042038/pdf
DOI  :  10.1088/1757-899X/537/4/042038

学科分类：材料科学（综合）

来源: IOP

【 摘 要 】

Reasonable distribution of resources for regression testing execution of software is considered to be the most important task. Finding the best solution for it may significantly reduce expenses on the whole system development. Neural network model may be used for testing management, as it has fault-prediction ability in each program module. Code parameters are independent variables and presence of errors is a dependent value in such model. Neural network can learn on real data - real testing product. Testing results received from different environment may be integrated easily in the knowledge base. This allows neural network to learn during each testing iteration. The module that potentially contains an error is tested at the first place and more thoroughly. Presented method may predict testing results and distribute resources accordingly.