Comprehensive Evaluation of Software Quality Based on LM-BP Neural Network

Anbang Wang,Lihong Guo,Yuan Chen,Junjie Wang,Yuanzhang Song
DOI: https://doi.org/10.1109/DSA.2017.37
2017-10-01
Abstract:In view of the shortcomings of traditional software quality evaluation methods, such as subjective and lack of self-learning ability, a comprehensive evaluation method of software quality based on Levenberg Marquardt Back Propagation (LM-BP) neural network is proposed. Comprehensive evaluation system of software quality is established based on ISO/IEC 9126 software quality model, and LM-BP algorithm is used to solve the problem of standard BP algorithm, such as slow convergence rate. A comprehensive evaluation model of software quality based on LM-BP neural network is established, which provides a new method and idea for comprehensive evaluation of software quality. The experimental results show that the comprehensive evaluation of software quality based on LM-BP neural network can obtain the comprehensive evaluation result of software quality quickly and accurately, and has good objectivity and practicability.
Computer Science
What problem does this paper attempt to address?