关键词: 软件质量预测模型, 软件度量, 神经网络, 粒子群算法

Abstract: The modeling technology of software which can find the nonlinear relationship between metric data and quality factors is the key technology in the software quality evaluation system. BP neural network is a kind of modeling method for the nonlinear relationship between metric data and quality factors, but there are some problems, such as slow convergence speed and easily getting into local minimum. So it proposes that using the optimized BP network based on PSO to establish the prediction model of software quality, which solves the problem of slow convergence speed and easily getting into local minimum well. In the basis of the evolutionary BP neural network, through the experiment with 28 groups of data, and by comparing with the result of BP model, the model is validated.

Key words: prediction model of software quality, software metrics, neural network, Particle Swarm Optimization（PSO）