基于量子免疫克隆BP算法的软件缺陷预测模型

Software defect prediction model based on quantum immune clonal BP algorithm

针对现有的软件缺陷预测模型中所存在的不足,将量子免疫克隆算法和BP神经网络算法结合,应用到软件缺陷预测中,设计了基于量子免疫克隆BP算法的软件缺陷预测模型(SDPM-QICBP).在该模型中,将量子计算引入到传统进化算法中,特别是在计算量子旋转门的角度时,将传统的查表计算方式与Logistic映射公式相结合,设计了新的量子旋转角的计算公式.模型采用量子免疫克隆算法(QIC)对标准BP神经网络的阈值和权值优化改进,并基于相关数据集进行实验分析.仿真实验的结果表明,和标准BP神经网络算法和朴素贝叶斯算法(NB)相比,该模型准确度和精确度均较高,且迭代次数减少.

Aiming at the shortcomings of the existing software defect prediction models,quantum immune clonal algorithm(QCA)is combined with BP neural algorithm.It is applied to software defect prediction.The model is designed,that is,a software defect prediction model based on quantum immune clonal BP algorithm(SDPM-QICBP).In this model,quantum computing is introduced into the traditional evolutionary algorithm,especially when calculating the angle of quantum revolving door.This paper puts forward a new formula for calculating the quantum rotation angle by combining the standard look-up table calculation method with Logistic mapping formula.In this model,quantum immune clonal algorithm is used to improve and optimize the thresholds and weights of the traditional BP neural network,and experimental analysis is carried out based on relevant data sets.The simulation results show that this model has higher accuracy and precision,and less number of iteration,compared with the standard BP neural network and NP.