期刊
NEUROCOMPUTING
卷 333, 期 -, 页码 319-328出版社
ELSEVIER
DOI: 10.1016/j.neucom.2018.12.033
关键词
Power consumption; Room classification; Echo state networks; Neural networks
资金
- National Natural Science Foundation of China [61533017, 61603387, 61722312, 61773075, 61601458]
- National Key Research and Development Program of China [2016QY01W0103]
Office buildings commonly contain such room types as office rooms, server rooms, storage rooms, meeting rooms, etc., while the power consumption inside the rooms generally comes from appliances, lights and air-conditioners. Based on the features of power consumption in different rooms, the aim of this study is to classify the rooms into different types by proposing an echo state network (ESN) based approach. Given the data on power consumption, the proposed approach is divided into two steps, where the first step is to establish three ESNs to model the three categories of power consumption, and the second step is to establish a fourth ESN to determine the type of a room by using the outputs of the first three ESNs. The practical performance of the proposed approach is displayed by a detailed experimental study, where the proposed approach achieves high classification accuracies and shows great superiority to several classical algorithms. (C) 2019 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据