Identifying the influential features on the regional energy use intensity of residential buildings based on Random Forests

详细信息    查看全文

• 作者：Jun Ma ^jmaae@ust.hk ; Jack C.P. Cheng ; ^{cejcheng@ust.hk}
• 关键词：Energy use intensity (EUI) ; Multi-family residential buildings ; Random Forests (RF) ; Regional effect ; Variable importance
• 刊名：Applied Energy
• 出版年：2016
• 出版时间：1 December 2016
• 年：2016
• 卷：183
• 期：Complete
• 页码：193-201
• 全文大小：1917 K
• 卷排序：183

文摘

The influence of 171 different kinds of features was analyzed using Random Forests. Average energy use intensities of 1322 regions were set as the regression target. The model built by Random Forest has lower MSE than Lasso and SVM. An educational feature was found to be the most influential. The study not only identifies the influential features, but also matches the areas.