Exploring Feature Selection Techniques in Predicting Indian Household Electricity Consumption

Abu Bakar Siddique Mahi (), Farhana Sultana Eshita (), Nishat Tasnim (), Aloke Kumar Saha () and Shah Murtaza Rashid Al Masud ()
Additional contact information
Abu Bakar Siddique Mahi: University of Asia Pacific
Farhana Sultana Eshita: University of Asia Pacific
Nishat Tasnim: University of Asia Pacific
Aloke Kumar Saha: University of Asia Pacific
Shah Murtaza Rashid Al Masud: University of Asia Pacific

A chapter in Machine Learning Technologies on Energy Economics and Finance, 2025, pp 219-241 from Springer

Abstract: Abstract A precise prediction of household energy usage is critical for optimizing energy management, forecasting demand, and formulating sustainable energy policies. In this study, we explored three feature selection techniques: LASSO, Random Forest feature importance, and Mutual Information Regression to determine the key factors that impact household energy consumption. After we identified the most relevant features which is Random Forest feature importance, we proceeded to assess the performance of eight distinct machine learning models. These models encompassed: Generalized Additive Model (GAM), Least Angle Regression (LARS), Huber Regressor, Random Sample Consensus Regressor (RANSAC), Extreme Gradient Boosting (XGBoost), Partial Least Squares Regression (PLS), Extremely Randomized Trees (ExtraTrees), and Light Gradient Boosting Machine (LightGBM). Our findings demonstrated that the ExtraTrees regressor surpassed all other models, achieving an outstanding accuracy of 99.99% on the test set. This demonstrates the effectiveness of feature selection and machine learning model in capturing the intricate connections between the input characteristics and the intended variable. The feature selection analysis uncovered the important factors. These valuable insights provide valuable guidance to policymakers and energy service providers as they work toward developing targeted interventions and customized energy efficiency programs. This study highlights the success of a thorough feature engineering and model selection approach in accurately predicting household energy consumption. The findings of this research have important implications for the development of smart energy management systems and the promotion of energy-efficient behaviors at the household level.

Keywords: Energy consumption; Lasso; Random forest feature importance; Mutual information gain; Machine learning (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:isochp:978-3-031-95099-5_10

Ordering information: This item can be ordered from
http://www.springer.com/9783031950995

DOI: 10.1007/978-3-031-95099-5_10

Access Statistics for this chapter

More chapters in International Series in Operations Research & Management Science from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().