基于相关向量机的电气化铁路用电量预测方法

为了提高用电量预测的精度，提出基于相关向量机回归的预测模型，在时间序列数据的基础上通过拟合训练得出其内在关系，进而可以计算得到较为准确的预测结果。相关向量机具有算法简洁和预测精度高等特点，易于编程使用。为了验证模型的有效性，本文选取2006年3月至2013年2月之间的电气化铁路用电量数据进行模型拟合训练，并预测分析2013年3月至7月的用电量情况。最后，通过对比分析表明相关向量机模型的预测结果比传统回归方法的预测结果更优。

Prediction approach of electricity demand for electrified railway based on relevance vector machine

In order to improve the prediction accuracy of electricity demand, a forecasting model was proposed based on relevance vector machine regression. On the basis of time series data, relative accurate results of pre- diction can be calculated by fitting simulation to derive the internal relationship. The relevance vector machine is simple, high prediction precision and is easy to use. In order to verify the validity of the model, the electricity demand data of electrified railway between 2006.3 to 2013.2 was adopted as training data, and the electricity demand from 2013.3 to 2013.7 was predicted and investigated. Finally, by comparing with the traditional re- gression results, the predicted results gained from relevance vector machine model are better.