Random extreme learning machines to predict electric load in buildings

The study of energy efficiency in buildings is an active field of research. Modeling and prediction of power-related magnitudes allow us to analyse the electrical consumption. This can lead to environmental and economical benefits. In this study we compare different techniques to predict active powe...

Full description

Saved in:
Bibliographic Details
Published in:Progress in artificial intelligence 2016-05, Vol.5 (2), p.129-135
Main Authors: Vergara, Gonzalo, Alonso-Barba, Juan I., Soria-Olivas, Emilio, Gámez, José A., Domínguez, Manuel
Format: Article
Language:English
Subjects:
Artificial Intelligence
Computational Intelligence
Computer Imaging
Computer Science
Control
Data Mining and Knowledge Discovery
Mechatronics
Natural Language Processing (NLP)
Pattern Recognition and Graphics
Regular Paper
Robotics
Vision
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The study of energy efficiency in buildings is an active field of research. Modeling and prediction of power-related magnitudes allow us to analyse the electrical consumption. This can lead to environmental and economical benefits. In this study we compare different techniques to predict active power consumed by buildings of the University of León (Spain). The original dataset contains time, environmental and electric data for 30 buildings. In our study we follow a two-step procedure: first, we grouped the buildings in terms of their electric load using principal component analysis (PCA) and k-medoids based clustering. From this clustering we selected three prototype buildings. Second, we have applied neural network-based machine learning techniques to carry out the prediction task. In particular, we used well-known multi layer perceptron (MLP) and extreme learning machine (ELM) algorithms, as well as a new model proposed in this paper which constitutes its major contribution: random extreme learning machines (RELM). Our analysis shows that RELM advantages MLP and ELM in labour days, which is the most interested case because of it highest electric demand.
ISSN:2192-6352
2192-6360
DOI:10.1007/s13748-015-0077-6