轻轨项目分项工程投资风险评价模型设计 ——基于BP神经网络的投资风险分析

以津滨轻轨一期工程的问卷调研数据为基础结合该工程及其它相似工程的投资统计数据，估计了各分项工程成本超支百分比分布及其特定风险因素诱发的成本超支百分比分布，利用BP神经网络原理，以路基工程为例构建了分项工程的投资风险BP神经网络评估种子模型，为二期工程及其它欲建的城市轨道工程的投资风险评估提供了有效的定量评估依据，为样本扩充后成熟评估模型的构建提供了升级模板。

Design of investment risk valuation model on subentry project of light track construction investment risk analysis based on BP neural net

The questionnaire survey cost datum from the first part of the Jin-Bin Light Track project and the statistic cost datum from this project and projects alike were processed to estimate the percents distribution of cost overspending on the subentry project and this kind of distribution caused by each risk factor.BP Neural Net Valuation Seed Model on the investment risk was constructed when the roadbed project was taken as an example with the neural net theory.It was helpful to valuate the investment risk of the second part project of the Jin-Bin Light Track and other city track projects which would begin as a effective quantified valuation model.It also prepared a promotion template for advanced valuation model with sufficient training sample.