Tram Energy Consumption Prediction Based on Improved Gaussian Process Regression Model

To overcome the problem of low energy consumption prediction accuracy caused by the diversity of influencing factors for tram energy consumption, a tram energy consumption prediction theory based on improved Gaussian process regression model come up, in combination with gray theory. Due to the fact that the standard Gaussian process regression model directly selects the covariance function, which can easily lead to a problem that the fitting degree with the physical process of the research object is not high enough, a multi kernel covariance function is proposed, in order to improve the traditional Gaussian process regression algorithm. The calculation results of an actual example indicate that the accuracy of this prediction method is improved compared with conventional means.