基于聚类分析与SVR公共建筑能耗预测研究

Research on Energy Consumption Prediction of Public Building Based on Cluster Analysis and SVR

为解决能耗预测算法中存在的高维度、局部极值点等问题,提出基于聚类分析与支持向量回归(SVR)的公共建筑能耗预测。对数据进行相关性分析和降噪标准化处理,然后输入K-Means算法中进行聚类分析;分析后的数据用于模型的训练与验证;以训练模型为基础出开发人机交互预测界面,实现公共建筑能耗可视化预测。仿真结果表明,未聚类预测模型的判定系数为0.89725,聚类后的预测模型的判定系数为0.93645,聚类后预测比未聚类预测精确性提高3.9%。提出的算法在公共建筑能耗预测中预测结果精准,也可用于其它类型建筑能耗预测。

In order to solve the problems of high dimension and local extreme points in energy consumption prediction algorithm,a public building energy consumption prediction based on cluster analysis and support vector regression(SVR) is proposed.The data were processed by correlation analysis and noise reduction standardization,and then input into the k-means algorithm for clustering analysis;the data after analysis was used for training and verification of the model;the human-computer Interaction prediction interface was developed based on the training model to realize visualized prediction of energy consumption in public buildings.The simulation results show that the decision coefficient of the non-cluster prediction model is 0.89725,and the decision coefficient of the post-cluster prediction model is 0.93645.The accuracy of the post-cluster prediction is 3.9% higher than that of the non-cluster prediction model.The Algorithm presented in this paper is accurate in the prediction of energy consumption of public buildings,and can also be used in other types of buildings.