Analyzing Biogas Production in Livestock Farms Using Explainable Machine Learning

Md. Mahedi Hassan, Mahira Shamim, Mahmudul Hasan, Md Amir Hamja, Kanij Fatema and Sudipto Roy Pritom
Additional contact information
Md. Mahedi Hassan: World University of Bangladesh
Mahira Shamim: University of Chittagong
Mahmudul Hasan: Hajee Mohammad Danesh Science and Technology University
Md Amir Hamja: Hajee Mohammad Danesh Science and Technology University
Kanij Fatema: Hajee Mohammad Danesh Science and Technology University
Sudipto Roy Pritom: American International University of Bangladesh

A chapter in Machine Learning Technologies on Energy Economics and Finance, 2025, pp 199-225 from Springer

Abstract: Abstract Biogas is gaining popularity as a renewable energy source due to its ability to reduce greenhouse gas emissions. Investors from various sectors are increasingly contributing to the production of renewable energy. The environmental damage caused by fossil fuels and the depletion of natural resources have drawn public attention to renewable energy as a sustainable solution for future energy production. In this study, we propose a methodology for estimating daily biogas production, emphasizing the importance of reliable and personalized data for model effectiveness. We employed ten machine learning algorithms to estimate biogas production, including Ridge Regression, Lasso Regression, k-Nearest Neighbors, ElasticNet, Classification and Regression Tree, Random Forest, XGBoost, LightGBM, Gradient Boosting Machine (GBM), and CatBoost. Additionally, we utilized explainable artificial intelligence (XAI) tools to discuss the global and local interpretation of the features. XAI models were used to identify significant parameters that influence biogas production. We evaluated model stability using 80:20, 70:30, and 50:50 training-testing ratios. XGBoost and LightGBM consistently outperformed the other models, with XGBoost excelling at the 80:20 and 50:50 ratios (root-mean-square error (RMSE): 0.091 and R 2 $$^{2}$$ : 0.847) and LightGBM performing best at the 70:30 ratio (RMSE: 0.075 and R 2 $$^{2}$$ : 0.895) for accurate and efficient biogas predictions. Waste efficiency and total waste (kg/day) were found to significantly impact biogas predictions, and prioritizing these variables enhances accuracy and productivity in biogas generation. This study will assist domain experts and investors in recommending a focus on these factors within prediction systems and management policies to increase biogas production.

Keywords: Biogas; Renewable energy; Machine learning; Explainable AI (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:978-3-031-94862-6_9

Ordering information: This item can be ordered from
http://www.springer.com/9783031948626

DOI: 10.1007/978-3-031-94862-6_9

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 ().