Logo image
An Integrated Machine Learning and Agent-Based Modelling Pipeline for Socio-Environmental Forecasting
   

An Integrated Machine Learning and Agent-Based Modelling Pipeline for Socio-Environmental Forecasting

Ogechi Smart Ekejiuba, Olayinka Anthony Ojo, Ogunlola John Olalere, Blessing Alice Alao-Olatunji Negin Aboutorabi
2026 International Conference on Cybersecurity, Digital Forensics, and AI Applications (ICCSDFAI), pp.303-309
2026 International Conference on Cybersecurity, Digital Forensics, and AI Applications (ICCSDFAI) (Istanbul, Turkiye, 25/06/2026–27/06/2026)
17/08/2026
: https://doi.org/10.1109/ICCSDFAI70505.2026.11647879
Agent-based modeling agent-based modelling Climate climate migration Forecasting Human factors Hybrid modelling LightGBM Machining Modeling Printing scenario simulation SHAP Simulation socio-environmental systems Machine Learning Meteorology
Socio-environmental systems have nonlinear dynamics, feedback, and behavioural responses of heterogeneity, which undermines the traditional forecasting techniques. This paper proposes a combined modelling pipeline that involves deploying ensemble machine learning and agent-based simulation to aid in predicting and scenario analysis within complex socio environmental scenarios. The suggested pipeline connects regression-based prediction models and behavioural simulation layer, which allows aggregate forecasts to drive the micro-level adaptive dynamics and policy feedback mechanisms. The review of the framework is done on the case of the climate-driven migration using the multi-source panel data collected in 20 countries from 2009 to 2023 and included a set of climate, socio-economic and governance indicators. Comparative tests with Random Forest, XGBoost, CatBoost, and LightGBM prove the statement that gradient-boosting models have the highest predictive power, as LightGBM can explain about 61% of the variance of the actual migration flows that it predicts. SHAP as a model of explainability on extreme temperatures and institutional capability singles out the events as the defining predictors. The agent-based element also explains the effect of behavioural adaptation and policy actions in system-level outcomes in other situations. The findings show that a combination of predictive and behavioural modelling enhances interpretability, strength and the depth of analysis when used to do socio-environmental forecasting.

(2)

pdf
An Integrated Machine Learning and Agent-Based Modelling Pipeline for Socio-Environmental Forecasting592.52 kB
Accepted Open Access  — This license enables reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator.
url
Link to published version
Published (Version of record) Publisher may require payment for access  — This item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation (such as any Fair Dealing allowances allowed by copyright law). For other uses you need to obtain permission from the rights-holder(s).
3
Logo image

Usage details for all content viewed and downloaded in this site are shared with IRUS-UK (Institutional Repository Usage Statistics UK). Cookies are used to remember your decision. Click Accept to accept usage details sharing and the cookies.