All Issue

2025 Vol.34, Issue 4 Preview Page

Original Articles

Comparison and Evaluation of Machine Learning Models for Predicting CH4 Emissions from Pig Housing

양돈시설 CH4 배출 예측에 적합한 머신러닝 비교 및 평가

Kyoung Won Min1
,
Rack Woo Kim2*
,
Dae Heon Park3
,
Jun Gyu Kim3
,
Su Bin Ahn4
,
Sun Hyoung Lee4
,
Chan Min Kim4
,
Hee Woong Seok4
민 경원1
,
김 락우2*
,
박 대헌3
,
김 준규3
,
안 수빈4
,
이 선형4
,
김 찬민4
,
석 희웅4
1Undergraduate Student, Department of Smart-Farm Engineering, College of Industrial Sciences, Kongju National University, Yesan 32439, Korea
2Associate Professor, Department of Agriculture Engineering, College of Industrial Sciences, Kongju National University, Yesan 32439, Korea
3Electronics and Telecommunications Research Institute (ETRI) JEJU AI Transformation Convergence Research Section, Jeju 63309, Korea
4Graduate Student, Department of Agricultural Engineering, College of Industrial Sciences, Kongju National University, Yesan 32439, Korea
1국립공주대학교 스마트함공학과 학부생
2국립공주대학교 스마트팜공학과 교수
3한국전자통신연구원 제주AX 융합연구실 연구원
4국립공주대학교 농공학과 대학원생
*Corresponding Author
31 October 2025. pp. 604-618
PDF XML Citation
Abstract

CH₄ is a major air pollutant emitted from pig housing and is influenced by various environmental and management factors. This study monitored CH4 concentrations using sensor-based measurements and analyzed their relationships through machine learning models. Outliers were detected using the Hampel filter, and missing values were corrected with hybrid imputation techniques to ensure data quality. Among Random Forest, LightGBM, and HistGB models, the HistGB model showed the highest predictive accuracy (R2 = 0.83) and lowest RMSE. The results indicate that feed intake, pH, and temperature significantly affect CH₄ emissions, while ventilation and humidity primarily influence NH3 variation. This study provides a data-driven approach for identifying key emission factors and improving gas management in intensive pig production systems.

Keywords
Machine Learning
methane
pig farming facilities
prediction

본 연구는 양돈시설에서 수집된 복합 환경 데이터를 활용하여 CH4 농도를 예측할 수 있는 머신러닝 기반 모델을 구축하고, 성능을 비교·분석하였다. 데이터 전처리 및 보간 과정을 통해 이상치를 제거하고 품질을 향상시킨 데이터를 기반으로 RandomForest, LightGBM, HistGB 모델을 학습한 결과, HistGB 모델이 가장 우수한 성능을 보였다 (RMSE=3.90, MAE=3.02, R2=0.835). 이는 히스토그램 기반 분할 방식이 데이터의 잡음을 완화하고 제한된 학습 데이터에서도 안정적인 성능을 확보했기 때문이다. RandomForest 역시 유사한 수준의 성능을 보였으나, 저농도 구간이나 급격한 농도 변화 구간에서는 변동성이 다소 크게 나타났다. 반면, LightGBM은 설명력이 상대적으로 낮았으며, R2가 0.773으로 HistGB (0.835) 대비 약 7.4%p 감소하였고, RMSE도 4.575로 가장 높은 예측 오차를 보였다. 상관관계 분석을 통해 CH4 농도가 분뇨의 산성도, 사료 섭취량, NH3 및 CO2 농도와 밀접한 관련성을 확인하였으며, 이는 분뇨 발효가 CH4 발생의 핵심 요인임을 뒷받침한다. 따라서 CH4 발생을 저감하기 위해서는 환경제어를 포함한 분뇨 관리와 급이 전략을 함께 고려해야 함을 나타낸다. 본 연구의 결과는 양돈시설에서 고가의 CH4 센서를 상시 운용하지 않고도, 기존 환경 데이터와 머신러닝 모델을 활용하여 CH4 농도를 간접적으로 추정할 수 있음을 보여준다. 이는 향후 스마트팜 시스템과 연계하여 온실가스 배출 모니터링 및 저감 전략 수립에 활용될 수 있으며, 지속가능한 축산업을 위한 데이터 기반 관리 체계 구축에 기여할 것으로 기대된다.

키워드
데이터 전처리
스마트팜 시스템
온실 가스 저감
환경 센서 모니터링
References
1

Aarnink A.J.A. 1997, Ammonia emission from houses for growing pigs as affected by pen design, indoor climate and behaviour. Ph.D. Thesis, Wageningen Agricultural University, Wageningen, The Netherlands. ISBN 90-5485-662-9.

2

Aarnink A.J.A., A. Hol, and N.W.M. Ogink 2016, Ammonia emission from organic pig houses determined with local parameters. In: Proc. CIGR-AgEng Conf., Aarhus, Denmark, 26-29 June 2016, pp 1-9.

3

Ahn S.B., S.H. Lee, and R.W. Kim 2023, Design of a small-scale smart swine engineering test and demonstration facility for greenhouse gas emissions and indoor environment research of livestock facilities. Rural Resource 65:42-49.

4

Andretta I., F.M.W. Hickmann, A. Remus, C.H. Franceschi, A.B. Mariani, C. Orso, M. Kipper, M.P. Létourneau-Montminy, and C. Pomar 2021, Environmental impacts of pig and poultry production: Insights from a systematic review. Front Vet Sci 8:750733. doi:10.3389/fvets.2021.750733

10.3389/fvets.2021.750733
5

American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) 2017, ASHRAE Handbook－Fundamentals (SI edition). ASHRAE, Atlanta, GA.

6

Atakora J.K.A., K. Basri, T. Matsuo, and Y. Koike 2011, Effect of diet and manure management on ammonia emissions from pig facilities. Agric Eng Int: CIGR J 13:1-10.

7

Basak J.K., B. Paudel, N.C. Deb, D.Y. Kang, and H.T. Kim 2024, Modeling ammonia concentration in swine buildings using biophysical data and machine learning algorithms. Comput. Electron. Agric. 225:109269. doi:10.1016/j.compag.2024.109269

10.1016/j.compag.2024.109269
8

Bhatt N.P., and S. Varma 2023, An enhanced LightGBM model with data analytical approach for crop recommendation. Proc. Int. Conf. Electron. Renew. Syst. (ICEARS), Tuticorin, India, pp 1538-1544. doi:10.1109/ICEARS56392.2023.10085596

10.1109/ICEARS56392.2023.10085596
9

Blanes-Vidal V., M.N. Hansen, S. Pedersen, and H.B. Rom 2008, Reduction of odour and ammonia emission from pig slurry by acidification: Effects of pH, temperature and mixing. Agric Ecosyst Environ 124:237-244. doi:10.1016/j.agee.2007.09.009

10.1016/j.agee.2007.09.009
10

Bogireddy S.R., and H. Murari 2024, Enhancing crop yield prediction through random forest classifier: A comprehensive approach. Proc Int Conf Smart Electron Commun (ICOSEC), pp 1663-1668. doi:10.1109/ICOSEC61587.2024.10722249

10.1109/ICOSEC61587.2024.10722249
11

Castrillón N., M. Cardona, L. López, and J. Jiménez 2020, Assessment of methane emissions for different typologies of fattening swine facilities in the department of Antioquia, Colombia. Agron Res 18:55-67.

12

Deng Y., X. Chen, M. Yin, C. Wang, P. Dong, Z. Xie, J. Sun, and J. Wen 2024, Research on predicting microclimate in pig house based on machine learning algorithms. Preprint, Research Square. doi:10.21203/rs.3.rs-4734553/v1

10.21203/rs.3.rs-4734553/v1
13

Eggleston H.S., L. Buendia, K. Miwa, T. Ngara, K. Tanabe, eds 2006, 2006 IPCC guidelines for national greenhouse gas inventories. Institute for Global Environmental Strategies, Hayama, Japan.

14

Feng K., Y. Wang, R. Hu, and R. Xiang 2022, Continuous measurement of ammonia at an intensive pig farm in Wuhan, China. Atmosphere 13:442. doi:10.3390/atmos13030442

10.3390/atmos13030442
15

Greenhouse Gas Inventory and Research Center of Korea (GIR) 2023, National greenhouse gas inventory report of Korea 2023. GIR, Seoul, Korea.

16

Ivanova-Peneva S.G., J. Arogo, R.H. Zhang, and G.L. Riskowski 2008, Ammonia emissions from swine farms: A review of measurement techniques and modeling approaches. Agric Ecosyst Environ 128:1-17. doi:10.1016/j.agee.2008.05.014

10.1016/j.agee.2008.05.014
17

Kim K.Y., H.J. Ko, H.T. Kim, Y.S. Kim, Y.M. Roh, C.M. Lee, and C.N. Kim 2008, Quantification of ammonia and hydrogen sulfide emitted from pig buildings in Korea. J Environ Manage 88:195-202. (in Korean). doi:10.1016/j.jenvman.2007.02.003

10.1016/j.jenvman.2007.02.003
18

KREI (2023) Agricultural outlook 2023 Korea: Beef, pork, and dairy supply and demand trends. Korea Rural Economic Institute, Seoul, Korea.

19

Kumar B.A., S. Bhavani, S.K. Babu, Y. Ramesh, and S. Devi 2024, A Pareto distribution-based gradient boosting for sustainable agriculture. Proc Int Conf Commun Electron Syst, pp 1307-1311. doi:10.1109/I-SMAC61858.2024.10714692

10.1109/I-SMAC61858.2024.10714692
20

Kythreotou N., G. Florides, and S.A. Tassou 2012, A proposed methodology for the calculation of direct consumption of fossil fuels and electricity for livestock breeding, and its application to Cyprus. Energy 40:226-235. doi:10.1016/j.energy.2012.01.077

10.1016/j.energy.2012.01.077
21

Lee S.H., H.J. Ko, K.Y. Kim, H.T. Kim, and Y.M. Roh 2006, Study on ammonia emission characteristics of pig slurry. J Environ Sci Int 15:23-31. (in Korean). doi:10.5322/JES.2006.15.1.023

10.5322/JES.2006.15.1.023
22

Ma T.M., H.S. Chen, X. Wang, Q. Xie, and Y. Wang 2022, Study on ammonia concentration prediction model of pigsty based on LSTM neural network. Scholars J Agric Vet Sci 9:80-84. doi:10.36347/sjavs.2022.v09i07.001

10.36347/sjavs.2022.v09i07.001
23

Maazallahi A., S. Thota, N.P. Kondaboina, V. Muktineni, D. Annem, A.S. Rokkam, M. Amini, M. Salari, P. Norouzzadeh, E.M. Snir, and B. Rahmani 2024, Naive Bayes and random forest for crop yield prediction. Preprint, Research Square. doi:10.21203/rs.3.rs-4345189/v1

10.21203/rs.3.rs-4345189/v1
24

Ministry of Agriculture, Food and Rural Affairs, Livestock Management Division; NH Agribusiness, Livestock Consulting Division 2021, Standard livestock housing design 2021: Swine. NH Livestock Information Center, Seoul, Korea. (in Korean)

25

Mosquera J., N. Edouard, F. Guiziou, R.W. Melse, A.L. Riis, S.G. Sommer, and E. Brusselman 2011, Decision document on the revision of the VERA protocol on air cleaning technologies: Measuring techniques for the determination of the removal efficiency for ammonia. Livest. Res., Wageningen UR, Report 767.

26

Ogejo, J.A., R.S. Senger, and R.H. Zhang 2010, Global sensitivity analysis of a process-based model for ammonia emissions from manure storage and treatment structures. Atmospheric Environment, 44: 3621-3629. doi:10.1016/j.atmosenv.2010.06.053

10.1016/j.atmosenv.2010.06.053
27

Pedregosa F., G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, and E. Duchesnay 2011, Scikit-learn: Machine learning in Python. J Mach Learn Res 12:2825-2830.

28

Peng K., Y. Wang, R. Hu, and R. Xiang 2022a, Continuous measurement of ammonia at an intensive pig farm in Wuhan, China. Atmosphere 13:442. doi:10.3390/atmos13030442

10.3390/atmos13030442
29

Peng S., J. Zhu, Z. Liu, B. Hu, M. Wang, and S.K. Pu 2022b, Prediction of ammonia concentration in a pig house based on machine learning models and environmental parameters. Animals 13:165. doi:10.3390/ani13010165

10.3390/ani13010165
30

Philippe, F.X., and B. Nicks 2015, Review on greenhouse gas emissions from pig houses: CO₂, CH₄ and N₂O by animals and manure. Agric For Meteorol 202:69-81. doi:10.1016/j.agrformet.2014.12.007

10.1016/j.agrformet.2014.12.007
31

Pushpalatha A.M., and P.K. Rani 2023, Effective crop yield prediction using gradient boosting to improve agricultural outcomes. Proc Int Conf Commun Electron Syst, pp 1-6. doi:10.1109/ICNWC57852.2023.10127269

10.1109/ICNWC57852.2023.10127269
32

Sefeedpari P., S. Khoshnevisan, F. Ghahderijani, S. Rafiee, and R. Rezaei 2024, Model adaptation and validation for estimating methane and ammonia emissions from fattening pig houses: Effect of manure management system. Animals 14:964. doi:10.3390/ani14060964

10.3390/ani14060964
33

Stinn J.P., D.S. Andersen, and D.R. Schmidt 2014, Ammonia and greenhouse gas emissions from a modern U.S. swine breeding–gestation–farrowing system. Atmos Environ 99: 315-323. doi:10.1016/j.atmosenv.2014.09.037

10.1016/j.atmosenv.2014.09.037
34

Tabase R.K., G. Næss, and Y. Larring 2023, Ammonia and methane emissions from small herd cattle buildings in a cold climate. Sci Total Environ 868:166046. doi:10.1016/j.scitotenv.2023.166046. 2023.166046

10.1016/j.scitotenv.2023.166046
35

Verde M.T., M. Toral, A. Sanz, L. del Prado, and A. de Vega 2023, A measurement system for enteric CH₄ emissions monitoring from ruminants in livestock farming. Acta IMEKO 12:1618. doi:10.21014/actaimeko.v12i4.1618

10.21014/actaimeko.v12i4.1618
36

Vyawahare N., M. Bhanse, A. Bhagat, and B. Bhonde 2025, Smart crop earning prediction model using optimized classification algorithm. Social Sci Res Netw (Preprint).

10.2139/ssrn.5095632
37

Wang P. 2023, Predictive machine learning models of methane emissions using farm environmental data. Proc Int Conf Comput Intell Data Sci, pp 881-887. doi:10.1145/3653081.3653229

10.1145/3653081.3653229
38

Wicklin R. 2021, The Hampel filter for robust outlier detection. SAS Blogs, SAS Institute Inc. Available at: https://blogs.sas.com/content/iml/2021/06/01/hampel-filter-robust-outliers.html (Accessed October 23, 2025).

39

Wolter M., S. Prayitno, and F. Schuchardt 2002, Comparison of greenhouse gas emissions from solid pig manure during storage versus during composting with respect to different dry matter contents. Landbauforsch. Völkenrode 52:167-174.

40

Xie Q., J.Q. Ni, and Z. Su 2017, A prediction model of ammonia emission from a fattening pig room using adaptive neuro-fuzzy inference system. J Hazard Mater 325:301-309. doi:10.1016/j.jhazmat.2016.12.010

10.1016/j.jhazmat.2016.12.010
41

Yamparla R., H.S. Shaik, N.S.P. Guntaka, P. Marri, and S. Nallamothu 2022, Crop yield prediction using random forest algorithm. Proc Int Conf Commun Electron Syst, pp 1538-1543. doi:10.1109/ICCES54183.2022.9835756

10.1109/ICCES54183.2022.9835756
42

Zong C., L. Chen, and Y. Yang 2015, Emission characteristics and control strategies of ammonia from livestock and poultry farms. Environ Sci Pollut Res 22:12634-12645. doi:10.1007/s11356-015-4563-8

10.1007/s11356-015-4563-8
Information
  • Publisher :The Korean Society for Bio-Environment Control
  • Publisher(Ko) :(사)한국생물환경조절학회
  • Journal Title :Journal of Bio-Environment Control
  • Journal Title(Ko) :생물환경조절학회지
  • Volume : 34
  • No :4
  • Pages :604-618
  • Received Date : 2025-09-29
  • Revised Date : 2025-10-23
  • Accepted Date : 2025-10-30
  • DOI :https://doi.org/10.12791/KSBEC.2025.34.4.604
Journal Informaiton Journal of Bio-Environment Control Journal of Bio-Environment Control
Online Submission
  • NRF
  • KOFST
  • crossref crossmark
  • crossref cited-by
  • crossref funder-registry
  • crosscheck
  • orcid
  • open access
  • ccl
Journal Informaiton Journal Informaiton - close