Article: Building planning energy consumption benchmark evaluation system based on regression analysis and data mining Journal: International Journal of Materials and Product Technology (IJMPT) 2024 Vol.69 No.3/4 pp.219 - 243 Abstract: First, this paper presents EMLBECPM for smart building planning and operations issues that use WSN to determine the best places to put various kinds of sensors and gateways and maximise energy usage within the bounds of connection, resources, security, and clustering coverage. Secondly, it uses support vector regression (SVR) and a nonlinear auto-regressive model with exogenous input (NARX) to predict power use for the next day, week, and month with a granularity of 15 minutes. The proposed autoregressive feature, temperature, and other situational relevant contextual data to heterogeneous time series exhibit distinct patterns on weekdays, weekends, and holidays. The outcomes demonstrate that the suggested technique is superior in optimising the building energy, accuracy, robustness, and generalisability. The simulation outcomes demonstrate the suggested energy optimisation ratio of 98.9%, efficiency ratio of 97.5%, energy monitoring ratio of 96.5%, prediction ratio of 95.4%, and error rate of 10.2% compared to other existing models. Inderscience Publishers - linking academia, business and industry through research

Title: Building planning energy consumption benchmark evaluation system based on regression analysis and data mining

Authors: Jiayi Li; Xiaobin Ma; Yu Zhang

Addresses: College of Architecture and Civil Engineering, Chongqing Metropolitan College of Science and Technology, Yongchuan 402100, Chongqing, China ' China Northwest Architecture Design and Research Institute Co., Ltd, Xi'an 710018, Shaanxi, China ' Communication Repair Shop beside the Beidilangqingju Community in Jiulongpo District, Jiulongpo 401329, Chongqing, China

Abstract: First, this paper presents EMLBECPM for smart building planning and operations issues that use WSN to determine the best places to put various kinds of sensors and gateways and maximise energy usage within the bounds of connection, resources, security, and clustering coverage. Secondly, it uses support vector regression (SVR) and a nonlinear auto-regressive model with exogenous input (NARX) to predict power use for the next day, week, and month with a granularity of 15 minutes. The proposed autoregressive feature, temperature, and other situational relevant contextual data to heterogeneous time series exhibit distinct patterns on weekdays, weekends, and holidays. The outcomes demonstrate that the suggested technique is superior in optimising the building energy, accuracy, robustness, and generalisability. The simulation outcomes demonstrate the suggested energy optimisation ratio of 98.9%, efficiency ratio of 97.5%, energy monitoring ratio of 96.5%, prediction ratio of 95.4%, and error rate of 10.2% compared to other existing models.

Keywords: machine learning; support vector regression; SVR; nonlinear auto-regressive model with exogenous input; NARX; building; energy consumption.

DOI: 10.1504/IJMPT.2024.145772

International Journal of Materials and Product Technology, 2024 Vol.69 No.3/4, pp.219 - 243

Received: 04 Jun 2024
Accepted: 06 Nov 2024
Published online: 23 Apr 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Building planning energy consumption benchmark evaluation system based on regression analysis and data mining

Keep up-to-date