基于数据挖掘算法的地铁站内温度时序预测方法

Time Series Prediction of the Indoor Temperature in the Subway Station Based on Data Mining Techniques

提出了基于数据挖掘算法的地铁站内环境温度时序预测的方法.分别建立了支持向量数据(SVR)、BP神经网络(BPNN)、决策树CART三种预测模型,对比了一般的输入-输出建模方法与基于时间时间延迟方法的预测结果,以及不同时间延迟下三种数据挖掘模型的预测结果.结果表明:基于时间序列的预测模型效果比一般的输入-输出模型效果更好;当时间延迟为1时,三种模型的预测精度相对较高,且在此条件下,模型的预测效率也最高;基于SVR的时间序列预测模型精度比BPNN和CART更高.

Methods of time series prediction of the indoor temperature in subway station based on data mining techniques were proposed in this paper. The prediction models of support vector regression（SVR）, back propagation neural network（BPNN） and classification and regression tree（CART） were developed, respectively. In addition, results of time-series prediction based on time delay were comparative analyzed to the general input-output model. Further, the prediction models of three data mining algorithms were investigated based on different time delays. Results showed that the time-series based models had desirable goodness that general input-output model. Moreover, it is found that the data mining models had the best prediction accuracy as well as highest efficiency when the time delay was 1. In addition, the time-series prediction model based on SVR obtained better accuracy than BPNN and CART.