← Back to Publications List

Predictive Modeling Of Biodiversity Loss In Bangladesh Wetlands Using Machine Learning

Students & Supervisors

Student Authors
Abrar Sharif
Bachelor of Science in Computer Science & Engineering, FST
Ekramul Hasib
Bachelor of Science in Computer Science & Engineering, FST
Supervisors
Md. Mortuza Ahmmed
Associate Professor, Faculty, FST

Abstract

Wetlands play an indispensable part in the ecosystem, biodiversity & environment of Bangladesh. These wetlands are facing compelling threats to biodiversity loss and environmental challenges. This study reports on the issue of biodiversity loss, and by analyzing environmental issues in the wetlands of Bangladesh via statistical and machine learning tools, using a general dataset from the year 1994 to 2023. This dataset provides Impact Score, Adoption Rate, Policy Effectiveness, Innovation Index and Sustainability Rating. Five predictive models: Linear Regression, Decision Tree, Random Forest, XGBoost, and Support Vector Regression (SVR), were used to predict two primary outputs: Impact Score (biodiversity loss) and Sustainability Rating. Random Forest and XGBoost were found to make the best predictions and detect subtle, nonlinear relationships among ecological, policy, and adoption-related variables. According to SHAP analysis, the research reveals Policy Effectiveness as the largest contributor to Sustainability Rating, followed by Innovation Index. On the contrary, the Adoption Rate has an inverse relationship with the Impact Score in the context that greater adoption of environmentally friendly habits essentially camouflages the loss of biodiversity. A forecast from 2024 to 2030 displays a consistent decline in biodiversity loss coupled with an increase in sustainability; therefore, policies are enforced, and innovative efforts continue. We found out that what is still driving biodiversity stress is variable policy action and low policy success despite the improvement in innovation and sustainability practices. This research puts forth the value of machine learning in ecological prediction and brings to light the requirement for better governance and policy action to address biodiversity loss in Bangladesh’s wetlands.

Keywords

Biodiversity Loss Machine Learning SHAP Policy Effectiveness Sustainability Bangladesh.

Publication Details

  • Type of Publication:
  • Conference Name: 7th International Conference on Integrated Sciences (ICIS) 2025
  • Date of Conference: 25/10/2025 - 25/10/2025
  • Venue: Eastern University Campus, Ashulia, Dhaka, Bangladesh
  • Organizer: Eastern University, Bangladesh
View on Publisher