PCA-ARIMA模型在国际水电工程项目风险预测中的应用

为有效预测少量数据下的国际水电工程项目风险问题，选取国际水电工程有关经济、社会和政治维度的16个风险指标，利用主成分分析法(PCA)对1984～2012年的29个训练数据进行降维，提取综合风险指标输入ARIMA模型进行风险预测拟合，并与2013～2017年的验证数据集做误差分析。结果表明，模型有效改善了传统风险分析法样本需求量高的缺点，风险预测结果与验证数据集在4年内的对比平均误差率低于2%,可见模型准确、可行。

Application of PCA-ARIMA Model in Risk Prediction of International Hydropower Projects

To effectively prediction of the risks of international hydropower projects with a small amount of data, this study selected 16 risk indicators of international hydropower projects from the three dimensions of economy, society and politics. The PCA was used to reduce the correlation of 29 training data during 1984 to 2012. And then the comprehensive risk indicators were extracted as the input of the ARIMA model for risk prediction. Finally, error analysis of the validation data sets from 2013 to 2017 was carried out. The results show that the model effectively improves the disadvantage of the large sample demand of the traditional risk analysis method. The average error rate of the comparison between risk prediction result and validation data set in 4 years is less than 2%, indicating that the model is accurate and feasible.