基于LVQ的软件项目风险评估模型的研究

Software Project Risk Evaluation Model Based on Learning Vector Quantization

以16种风险为基础,建立了一个新的软件项目风险评估模型,把以往每个软件项目的16种风险看做一个16×1维列矢量,并做为LVQ神经网络的训练矢量,对其进行聚类分析,最终把项目风险水平分为:风险水平很低、风险水平中等、风险水平很高等三个类别,并对项目风险水平做出预测。

This paper builds a new software project risk evaluation model based on sixteen kinds of risks.The model considers the risks of every project in the past as a 16×1 dimension column vector and them as the training vectors of LVQ,then has a cluster analysis about the training vectors.At last project risk is divided into three levels：low risk, middle risk and high risk,the paper explains the application in software project risk prediction.