运用软件成本估算技术可以在软件项目实施过程中有效控制进度、降低风险从而保证所开发软件的质量.本文对成本估算技术中加权类比估算模型的特征属性的权重，使用优化的粒子群算法(C&S-PSO)进行优化后，利用MMRE和Pred(0.25)两个标准与非加权类比模型、支持回归和模糊神经网络等模型进行估算的精度比较.另外，采用非参数自助法对优化的PSO加权类比估算模型稳定性进行评估.研究结果表明，从估算精度来看，采用本文的估算模型比上述估算模型估算精度高，同时该模型具有较好的稳定性.%Software cost estimation technology has become an important measure of software project management aspects of the control software schedule, reducing software risks and ensure software quality. In this paper, we use the model of C&S-PSO and weighted analogy estimates characteristic attributes to the weights. According to both MMRE and Pred (0.25) standards, we compare un-weighted analogy model, SRV, FNN and CART to estimate accuracy. Additionally, we use the non-parametric bootstrap method for raw data sampling, analysis PSO weighted analogy estimation model of stability. The results show that from the view of estimation accuracy the model of PSO has higher precision in weighted analogy estimates, and the model has also better stability.
展开▼
