Document Type

Article

Publication Date

10-31-2021

Publication Source

European Journal of Advances in Engineering and Technology

Abstract

This paper presents a detailed investigation into various regression techniques applied to predict building energy consumption. The dataset utilized in this study encompasses diverse attributes related to buildings, including size, location, usage type, construction materials, and energy consumption patterns, alongside weather-related data such as temperature, humidity, and precipitation. After preprocessing, which includes loading, inspecting, and imputing missing data, the dataset undergoes evaluation using regression techniques including Ridge Regression, Bayesian Ridge, Linear Regression, Orthogonal Matching Pursuit, Elastic Net, Huber Regression, Lasso Regression, and Passive Aggressive Regression. Evaluation metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), R-squared (R2), Root Mean Squared Logarithmic Error (RMSLE), and Mean Absolute Percentage Error (MAPE) are employed to assess each model's performance. The results offer valuable insights into predictive accuracy, computational efficiency, and robustness, aiding in the selection of the most suitable regression technique for building energy prediction tasks.

ISBN/ISSN

2394-658X

Comments

The document available for download is the published version, provided in compliance with the publisher's open-access policy. Permission documentation is on file. DOI: https://doi.org/10.5281/zenodo.11103735

Volume

8

Issue

10

Keywords

Building energy prediction, Regression techniques, Ridge Regression, Bayesian Ridge, Linear Regression, Orthogonal Matching Pursuit, Elastic Net, Huber Regression, Lasso Regression, Passive Aggressive Regression, Data preprocessing, Model evaluation, Computational efficiency, Predictive accuracy, Robustness, Sustainability


Share

COinS