TY - GEN
T1 - Energy Usage Prediction for Smart Home with Regression Based Ensemble Model
AU - Hoque, Mohammad Shamsul
AU - Jamil, Norziana
AU - Saharudin, Sharul Azim
AU - Amin, Nowshad
N1 - Publisher Copyright:
© 2020 IEEE.
PY - 2020/8/24
Y1 - 2020/8/24
N2 - Residential sectors using energy mainly though lighting and HV AC (Heating, Ventilation and Air-Conditioning) have become a significant consumer of world energy and it is expected to grow especially with the trend of increasing smart homes. To provide an optimum, accurate and reliable electricity distribution, load prediction is a prerequisite policy and operational implementation. Smart homes with the use of various sensors create big data that gives a favorable opportunity for developing data-driven energy usage prediction models. In this paper, a novel regression-based ensemble prediction model with inbuilt automated optimization for parameters is proposed to predict the demand of electricity. The model explains the 0.998 correlation between the features and their label, and achieved root mean squared error (RMSE) and Normalized Absolute Error as low as 5.508 and 0.0508 respectively. We have also proposed a novel data-driven classification of the energy usage by unsupervised learning through clustering.
AB - Residential sectors using energy mainly though lighting and HV AC (Heating, Ventilation and Air-Conditioning) have become a significant consumer of world energy and it is expected to grow especially with the trend of increasing smart homes. To provide an optimum, accurate and reliable electricity distribution, load prediction is a prerequisite policy and operational implementation. Smart homes with the use of various sensors create big data that gives a favorable opportunity for developing data-driven energy usage prediction models. In this paper, a novel regression-based ensemble prediction model with inbuilt automated optimization for parameters is proposed to predict the demand of electricity. The model explains the 0.998 correlation between the features and their label, and achieved root mean squared error (RMSE) and Normalized Absolute Error as low as 5.508 and 0.0508 respectively. We have also proposed a novel data-driven classification of the energy usage by unsupervised learning through clustering.
KW - Deep Neural Learning
KW - Ensemble model
KW - K-means Clustering
KW - Machine Learning
KW - Model Optimization
KW - Random Forest
KW - Rapidminer
KW - Unsupervised Learning
UR - https://www.scopus.com/pages/publications/85097650660
UR - https://www.scopus.com/inward/citedby.url?scp=85097650660&partnerID=8YFLogxK
U2 - 10.1109/ICIMU49871.2020.9243578
DO - 10.1109/ICIMU49871.2020.9243578
M3 - Conference contribution
AN - SCOPUS:85097650660
T3 - 2020 8th International Conference on Information Technology and Multimedia, ICIMU 2020
SP - 378
EP - 383
BT - 2020 8th International Conference on Information Technology and Multimedia, ICIMU 2020
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 8th International Conference on Information Technology and Multimedia, ICIMU 2020
Y2 - 24 August 2020 through 25 August 2020
ER -
