AgroPackAI: An AI-Powered Dual-Model Machine Learning System for Agro-Waste Eco-Packaging Recommendation with Simultaneous CO2 and Cost Prediction for Indian SMEs

AgroPackAI is a dual-model AI-powered eco-packaging recommendation system achieving R2 = 0.9793 for cost prediction (Random Forest) and R2 = 0.9853 for CO2 emission prediction (XGBoost) — new state-of-the-art benchmarks for both packaging ML tasks — with sub-200 ms real-time inference on consumer hardware. The system directly addresses all five identified research gaps: an integrated recommendation pipeline (G1), quantified CO2 prediction (G2), dual R2 > 0.97 (G3), 7 Indian agro-waste materials (G4), and a deployed BI dashboard (G5). The quantitative finding that agro-waste materials are simultaneously the most cost-competitive and most environmentally optimal packaging options in the Indian market establishes the economic case for EPR Rules 2022 compliance, not merely the regulatory one. Future directions for research: (1) real-time integration of SME inventory and purchasing systems; (2) user acceptance testing by packaging experts; (3) deep learning enhancements (TabNet, attention-based models); (4) expansion of eco-material database to include bio-composites;

[1] R. Geyer, J. R. Jambeck, and K. L. Law, "Production, use, and fate of all plastics ever made," Science Advances, vol. 3, no. 7, p. e1700782, 2017. [2] Central Pollution Control Board, "Annual report on plastic waste management in India," Ministry of Environment, Forest and Climate Change, Govt. of India, New Delhi, Tech. Rep., 2022. [3] Dept. for Promotion of Industry and Internal Trade, "Annual e-commerce report 2023," Ministry of Commerce and Industry, Govt. of India, New Delhi, Tech. Rep., 2023. [4] Punjab Remote Sensing Centre, "Assessment of paddy straw burning and its environmental impact in Punjab and Haryana," PRSC, Ludhiana, India, Tech. Rep., 2021. [5] R. A. Ilyas et al., "Characterization of sugar palm nanocrystalline cellulose reinforced sugar palm starch bionanocomposites," J. Biobased Mater. Bioenergy, vol. 12, no. 4, pp. 1-12, 2022. [6] EcoInvent Centre, "EcoInvent database version 3.9," Swiss Centre for Life Cycle Inventories, Zurich, 2023. [Online]. Available: https://www.ecoinvent.org [7] K. Danach, H. E. Dirani, and H. Rkein, "Revolutionizing supply chain management with AI: A path to efficiency and sustainability," IEEE Access, vol. 12, pp. 1-18, 2024. [8] L. Breiman, "Random forests," Machine Learning, vol. 45, no. 1, pp. 5-32, 2001. [9] T. Chen and C. Guestrin, "XGBoost: A scalable tree boosting system," in Proc. 22nd ACM SIGKDD Int. Conf. KDD, 2016, pp. 785-794. [10] ISO, "ISO 18602:2013 - Packaging and the environment: Optimization of the packaging system," ISO, Geneva, 2013.