Machine learning-based forecasting of food waste generation by category : a short-term predictive analysis

AuthorSeldon, Meto
Call NumberAIT Thesis no.EV-26-09
Subject(s)Food waste--Forecasting--Thailand
Machine learning
NoteA thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Environmental Engineering and Management
PublisherAsian Institute of Technology
AbstractInstitutional food waste management often lacks source-specific data for effective mitigation. This study addresses this gap by quantifying, categorizing, analyzing, and forecasting food waste at the Asian Institute of Technology (AIT). Over an 81-day period (November 2025 February 2026), a mixed-methods approach was employed, combining direct field measurements from 46 zones with resident surveys to identify behavioral drivers. The results indicated a total measured waste of 12,007.27 kg, which was extrapolated to an estimated campus-wide total of 21,847.73 kg, with an average measured waste generation of 148.24 ± 3.98 kg/day. The primary contributors were shops (36.04%) and kitchen dormitories (32.16%). Waste generation consistently peaked on Mondays and Tuesdays and increased during examination period.Three machine learning models were evaluated: Random Forest, Gradient Boosting and support Vector Regression (SVR). SVR emerged as the most reliable model (R2 = 0.6538), offering superior stability and generalization on unseen data compared to Random Forest, which exhibited overfitting. Feature importance analysis identified “Lag 1” (previous day’s waste) as the primary predictors, indicating persistent campus waste behavior.This research highlights the shift from routine disposal to proactive, category specific interventions such as portion control and campus food bank. These strategies align with Thailand’s Bio-Circular-Green (BCG) Economy Model, providing a data-driven framework for institutional sustainability.
Year2026
TypeThesis
SchoolFaculty of Civil and Environmental Engineering (2026)
DepartmentOther Field of Studies (No Department)
Academic Program/FoSEnvironmental Engineering and Management (EEM)
Chairperson(s)Thammarat Koottatep
Examination Committee(s)Ghimire, Anish;Tatchai Pussayanavin
Scholarship Donor(s)Global Water and Sanitation (GWSC);AIT Scholarship
DegreeThesis (M. Sc.) - Asian Institute of Technology, 2026


Usage Metrics
View Detail5
Read PDF1
Download PDF0