ABSTRACT
Postharvest quality assessment of fruits and vegetable is pivotal in ensuring ideal freshness, reduction in waste, and enhancing ne economic value of agricultural produce. The increase in application of machine learning (ML) in this domain has enabled important strides in automated detection and classification systems. This paper recommends a novel model, ForestSHAP, which integrates SHapley Additive exPlanations (SHAP) and Random Forest to provide both accurate prediction and interpretability for the classification of postharvest fruits and vegetables. ForestSHAP shows robust performance across numerous ML classifiers, including Logistic Regression, Support Vector Machines (SVM), and ensemble techniques like LightGBM and Stacking, evaluated on metrics such as sensitivity, accuracy, F1-score, specificity, precision, and ROC-AUC. The SHAP framework is applied to improve model transparency, identifying the most influential features affecting quality of produce. This work demonstrates that ML, combined with SHAP explainability, can significantly contribute to reducing postharvest losses and improving supply chain efficiency by guiding informed decisions based on model outputs.
References
- [1] Yuen, A., Li, J., Zhang, T., & Chen, M. (2024). ‘Freshness Recognition of Fruits and Vegetables Using CNN and BiLSTM.’ Journal of Agriculture Technology and Food Safety, 12(3), 245-260.
- [2] Tapia-Mendez, C., Hernandez, P., & Torres, M. (2023). ‘Ripeness Classification and Maturity Assessment Using CNN Models.’ International Journal of Food Science and Technology, 18(2), 101-115.
- [3] Mukherjee, A., Biswas, S., & Roy, D. (2022). ‘Mushroom Freshness Detection with SVM and ANN Models.’ Agricultural Informatics Journal, 15(2), 156-168.
- [4] Zhu, H., Liang, T., & Chen, J. (2023). ‘Evaluating Machine Learning Techniques for Ripeness and Freshness Classification in Mangoes.’ Food Technology Review, 15(3), 222-239.
- [5] Lundberg, S. M., Lee, S.-I. (2017). ‘A Unified Approach to Interpreting Model Predictions.’ Advances in Neural Information Processing Systems, 30, 4765-4774.
- [6] Rodriguez, C., Martinez, A., & Chavez, E. (2023). ‘Ensemble Learning for Postharvest Quality Monitoring in Peppers.’ AI in Food Science, 11(4), 102-115.
- [7] Amani, H., & Aghamohammadi, N. (2023). ‘Grading Fruit Varieties Using CNN and Particle Swarm Optimization.’ Computational Agriculture Journal, 22(4), 310-325.
- [8] Mukhiddinov, D., Anarbekov, T., & Mukhtorov, A. (2022). ‘Application of CNN-Based Approaches for Postharvest Classification of Fruits.’ Journal of Food Technology and Quality Assurance, 10(1), 85-97.
- [9] Mohi-Alden, K., & Soltani, B. (2023). ‘Computer Vision Systems for Pepper Ripeness Classification Using MLP.’ Postharvest Biology and Technology, 20(5), 478-492.
- Azarmdel, H., Rezaei, M., & Rahimi, Z. (2020). ‘Grading Dragon Pearl Fruits Using CVS and Machine Learning Models.’ Journal of Fruit Science and Technology, 14(3), 205-220.
- Kheiralipour, K., & Pormah, R. (2017). ‘Cucumber Quality Classification Using CVS and ANN.’ Journal of Cucumber Quality Analysis, 9(2), 100-110.
- Wang, X., Wu, X., & Qian, H. (2024). ‘Deep Learning Approaches for Postharvest Freshness Detection of Fruits and Vegetables.’ Journal of Agricultural Informatics, 18(2), 134-147.
- Gupta, R., Singh, V., & Sharma, P. (2023). ‘AI-Driven Quality Assessment of Postharvest Produce Using Hyperspectral Imaging.’ Computational Agriculture & Food Processing, 10(1), 23-36.
- Elmasry, G., & Barbin, D. (2023). ‘Machine Vision for Non-Destructive Quality Assessment in Postharvest Fruits.’ Postharvest Biochemistry Journal, 9(1), 56-72.
- Nguyen, T., Tran, Q., & Pham, D. (2023). ‘AI-Enhanced Detection of Bruises in Apples Using Image Analysis.’ Journal of Food AI Technologies, 5(2), 145-160.
- Kim, J., Park, S., & Lee, J. (2023). ‘Implementation of Deep CNNs for Maturity Classification in Tomatoes.’ International Journal of Food Science AI, 12(3), 175-190.
- Sanchez, G., & Patel, M. (2023). ‘SVM-Based Approach for Quality Grading in Citrus Fruits.’ Computational Food Science, 8(2), 89-102.
- Ogunleye, A., & Olaniyi, A. (2023). ‘Data Augmentation Techniques for Improving Postharvest Fruit Classification Models.’ Journal of Agriculture Data Science, 16(2), 190-203.
- Zhang, Y., & Tang, H. (2023). ‘LSTM Networks for Time-Series Analysis of Postharvest Fruit Shelf Life.’ Agricultural Informatics & Technology, 7(4), 205-218.
- Wang, Y., Liu, S., & Zhu, F. (2023). ‘Hybrid AI Systems for Ripeness Detection of Multiple Fruit Types.’ AI and Food Technology, 14(3), 134-152.
- Singh, A., & Singh, B. (2023). ‘Role of AI in Predicting Postharvest Quality of Vegetables Using Spectral Data.’ Journal of Food Science & AI Research, 9(1), 112-127.
- Deng, X., & Wei, Z. (2023). ‘Evaluation of AI-Based Methods for Non-Invasive Quality Testing of Postharvest Vegetables.’ Journal of Agricultural Informatics, 18(4), 78-92.
- Rahman, M., & Kabir, M. (2023). ‘ML-Enabled Systems for Postharvest Shelf-Life Prediction in Tropical Fruits.’ AI in Food and Agriculture, 9(2), 142-155.
- Ali, R., & Mustafa, Z. (2023). ‘Advanced CNN Architectures for Classification and Grading of Postharvest Fruits.’ Food Science AI Journal, 7(3), 210-225.
- Gao, L., & Sun, Y. (2023). ‘Application of ML Models in Postharvest Storage and Packaging of Vegetables.’ International Journal of Postharvest Technology, 10(3), 130-144.
- Xu, J., & Huang, G. (2023). ‘Comparative Analysis of AI Techniques for Postharvest Quality Detection.’ Journal of Agricultural Informatics & AI Applications, 17(1), 121-138.
- Patel, A., & Mehta, K. (2023). ‘Development of an Integrated AI System for Postharvest Quality Monitoring.’ Computational Agriculture & Food Science Journal, 6(3), 193-207.
Download all article in PDF
