基于PSO-SVR的输变电工程造价权重预测模型建立及分析

Establishment and Analysis for Weight Prediction Model of Transmission and Transformation Project Cost Based on PSO-SVR

为了提高输变电工程造价的预测精度,降低以后输变电工程建设项目的结余率和控制项目经费超支的问题,提出了一种基于熵权法的PSO-SVR输变电工程预测模型。首先,使用熵权法对影响工程造价的因素做权重分析,选择权重系数较大的影响因素输入SVR模型,并使用粒子群优化算法(PSO)优化SVR中核函数宽度和惩罚因子,从而建立PSO-SVR预测模型。把宁夏某电网公司2012-2014年间的106组输变电工程造价样本数据为例子,分别使用PSO-SVR和SVR模型对实例样本数据进行测试和验证,实验结果表明,PSO-SVR模型预测精度更高,稳定性更好,可以为以后新建的输变电工程造价预测提供一定的参考价值和意义。

In order to improve the prediction accuracy of power transmission and transformation project cost, decrease the balance rate of power transmission and transformation construction projects and control the overexpenditure of project cost, this paper proposes a PSO-SVR forecasting model based on entropy weight. First, it uses the entropy weight method to do weighting analysis of the effect factors of the project cost, chooses the larger weight coefficient influence factors to input SVR model and uses the particle swarm optimization algorithm(PSO) to optimize the SVR kernel function in the width and the punishment factor. Thus, PSO-SVR forecasting model is established. Using the PSO-SVR and SVR model to test and verify the 106 samples of power transmission and transformation project cost data from 2012 to 2014 in ningxia., the experimental results show that the PSO-SVR model prediction accuracy is higher, better stability and it can provide new power transmission and transformation project cost forecast after certain reference value and significance.