基于PCA-BP-GA的BOT项目风险预测研究

摘要

Due to the complexity and long-term of BOT project, before the implementation of BOT project it needs to accurately estimates the size of the risks. Because of many influence factors of the BOT project risk problem, firstly, using the principal component analysis (pca) for dimension reduction, and then using genetic algorithm to find the optimal weights and threshold of BP neural network threshold, at last establishing PCA-GA-BP prediction model of BOT project risk. At the same time, using previous data as a sample to predict the risk of BOT project, in the end using the sewage plant in Lanzhou example verification to show that this model scientific has guidance for practical engineering.%由于BOT项目本身的长期性和复杂性，所以在BOT项目实施前需要准确科学的预测出所面临的风险大小。针对BOT项目风险影响因素众多的问题，先利用主成分分析法进行降维，然后利用遗传算法找出BP神经网络的最优全值阈值，建立了PCA-GA-BP BOT项目风险预测模型。同时将以往的BOT项目数据作为学习样本，对BOT项目风险进行预测，并利用某地污水厂的例子进行验证，说明此模型对实际工程的科学指导性。