摘要
为进一步提高软件项目成本估测的准确度,减少在使用传统估测方法时产生的较大误差,提出了一种基于BP神经网络算法的软件项目成本估测模型。利用BP神经网络对软件项目成本进行估计,采用梯度下降自适应Ir的BP算法训练,仿真结果表明,基于算法的软件项目成本估测值与实际成本的偏差较小。
In order to improve the accuracy of software project cost estimation and reduce large gaps when in using traditional methods , a software project cost estimation model based on BP neural network algorithm is proposed .According to BP Neural Network ,the software project cost is estimated by utilizing the algorithm training of Gradient descent with adaptive Ir .The simulation result shows that the deviation value between estimation cost and actual cost is less based on this algorithm .
出处
《航空计算技术》
2014年第3期9-11,15,共4页
Aeronautical Computing Technique
基金
陕西省科技厅国际合作项目资助(2012KW-03-02)
陕西省科技厅工业攻关项目资助(2014k05-62)
西安市科技局科技计划项目资助(CXY1117)
西安市社科规划基金项目资助(14J28)
关键词
软件项目
成本估测模型
BP神经网络
改进算法训练
software project
cost estimation model
BP neural network
improved algorithm training
