基于人工神经网络的软件质量预测模型研究-计算机软件与理论专业论文.docxVIP

软件模块风险性预测模型在模块高风险性和模块低风险性预测中的预测精 软件模块风险性预测模型在模块高风险性和模块低风险性预测中的预测精 度分别提高了1倍和4倍。另一种是基于PcA一小波神经网络的软件质量 预测模型，它通过主成分分析方法来剔除实验数据的多重共线性，减少了 数据的维数，但是却保留了原始数据97％的信息量，简化了神经网络的结 构，并采用遗传算法来优化小波神经网络，使小波神经网络的预测能力得 到提高，以此来提高软件质量的预测精度；实验结果表明此模型所采用的 方法获得了比基于BP神经网络的软件质量预测模型和基于广义回归神经 网络的软件质量预测模型更好的预测准确度。 关键词：软件可靠性软件质量软件质量预测模型学习矢量量化神经 网络遗传算法小波神经网络主成分分析方法 STUDY STUDY ON THE SOFTWARE OUAL I TY PRED I CT I ON MODEL BASED ON ARTIFlCIAL NEURAL NETWORK ABSTRACT With the development of computer system,computers are now used more and more widely in every aspect and the compmer solare systems are becoming more complicated．Therefore，the software reliability attracts lots of concem and attention from increasing number of people．How to test the software reliability?It is to appraise and predict the behaviors of the sothare based on the SOare reliability model． At present，a lot of soft-ware reliability models are built based on the hypotheses of mistake characters and malfunction random behaviors that are left in software．These hypotheses vary from each other,therefore the difference of malfunction behavior types expressed by various software reliability models is obvious；and the testing results are very different from each other．Thus， developing a umverSal prediction model has become one of the important tasks， which is one of the hotspots in software reliability prediction model．The sotb,vare reliability prediction model Can predict the software reliability and doesn’t need any premise or hypothesis using neural network． There are two main achievements in sofb口vare reliability prediction model using artificial neural network：firstly,people take the invalid software time as the the input of software reliability prediction model；secondly,people USe the software quality metrics as the mput of neural networks in the software quality prediction model．This paper proposes two soRware quality prediction models using software quality metrics as the input of neural networks：the first one is the software