论文部分内容阅读
针对轨道客车运营企业普遍存在的成本统计信息少、呈片断化等特征,提出一种基于改进的主成分分析法(PCA)及粒子群最小二乘支持向量机(PSOLSSVM)的轨道客车运营能耗成本估算模型。首先分析确定运营能耗相关的车辆、线路及运营条件参数,再利用改进的PCA提取出运营能耗成本的特征参数,以这些特征参数为输入,以能耗成本为输出,构建基于PSO-LSSVM的成本估算模型。PSOLSSVM模型利用PSO优选LSSVM的参数,提高了模型的训练速度和预测能力。实例研究结果表明,相较于传统基于PCA及LSSVM模型,所构建估算模型的特征参数减少了27%,预测准确度提高了88%。
Aiming at the fact that there are few cost statistic information and fragmentation in railway bus operation enterprises, a new method based on PCA and PSOLSSVM is proposed. Cost estimation model. Firstly, the parameters of vehicles, lines and operating conditions related to operating energy consumption are analyzed. Then, the improved PCA is used to extract the characteristic parameters of operating energy costs. With these parameters as input, the energy cost is used as output to build the PSO-LSSVM Cost estimation model. The PSOLSSVM model uses the parameters of PSO-optimized LSSVM to improve the training speed and predictive ability of the model. The results of the case study show that compared with the traditional PCA and LSSVM models, the estimated parameters of the proposed model are reduced by 27% and the prediction accuracy is improved by 88%.