Automated Waste Classification Using Machine Learning for Sustainable Waste Management in Dhaka: A Comparative Evaluation of SVM, Random Forest, KNN, and Neural Network Approaches

Md Alauddin Mazumder

doi:10.25163/data.2110805

Data Modeling

Mathematical and Computational Data Modeling

Citations

4.5k

Views

Articles

Submit

Volume 2 Number 1 2021

RESEARCH ARTICLE (Open Access)

Previous Next Contents Vol 2 (1)

Automated Waste Classification Using Machine Learning for Sustainable Waste Management in Dhaka: A Comparative Evaluation of SVM, Random Forest, KNN, and Neural Network Approaches

Abstract 1. Introduction 2. Methodology 3. Results 4. Discussion 5. Conclusion Author Contribution Acknowledgement Competing Financial Interests References

Md Alauddin Mazumder1*

+ Author Affiliations

Data Modeling 2 (1) 1-8 https://doi.org/10.25163/data.2110805

Submitted: 10 May 2021 Revised: 02 July 2021 Accepted: 12 July 2021 Published: 14 July 2021

Abstract

Urban waste mismanagement remains one of the most persistent environmental challenges facing rapidly growing cities in the developing world, and Dhaka is no exception. With conventional sorting methods proving increasingly inadequate against rising waste volumes, there is a pressing — if still underexplored — need for scalable, automated classification solutions. This study evaluates the effectiveness of four machine learning algorithms — Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbors (KNN), and a shallow Neural Network (NN) — for automated image-based waste classification using the TrashNet benchmark dataset, comprising six waste categories: glass, paper, cardboard, plastic, metal, and trash. Images underwent standardized preprocessing including resizing, contrast enhancement, histogram equalization, and augmentation. Features were extracted using color, texture, and shape descriptors, and models were evaluated on accuracy, precision, recall, and F1-score. Random Forest achieved the highest classification accuracy at 66.60%, followed by SVM at 60.28%, KNN at 47.04%, and NN at 43.87%. These results reveal the practical utility of ensemble methods for structured feature-based classification, while also surfacing meaningful limitations of shallow neural architectures under constrained data conditions. The findings suggest that with expanded, locally curated datasets and deeper architectures, automated waste classification could meaningfully support waste sorting infrastructure in cities like Dhaka, contributing to renewable energy pipelines and broader sustainability goals.

Keywords: waste classification, machine learning, image processing, computer vision, municipal solid waste, Random Forest, Dhaka

References

Abadi, M., Agarwal, A., Barham, P., Brevdo, E., Chen, Z., Citro, C., Corrado, G. S., Davis, A., Dean, J., Devin, M., Ghemawat, S., Goodfellow, I., Harp, A., Irving, G., Isard, M., Jia, Y., Jozefowicz, R., Kaiser, L., Kudlur, M., … Zheng, X. (2015). TensorFlow: Large-scale machine learning on heterogeneous systems (Version 2.11.0) [Software]. Google Brain.

Aral, R. A., Keskin, S. R., Kaya, M., & Haciömeroglu, M. (2019). Classification of TrashNet dataset based on deep learning models. In Proceedings of the 2019 IEEE International Conference on Big Data (pp. 2058–2062). IEEE.

Bellman, R. (1961). Adaptive control processes: A guided tour. Princeton University Press.

Bhuiyan, M. R. A., Mamur, H., & Begum, J. (2021). A brief review on renewable and sustainable energy resources in Bangladesh: Current landscape and energy transition. Cleaner Engineering and Technology, 4, 100208.

Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5–32.

Canny, J. (1986). A computational approach to edge detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 8(6), 679–698.

Cortes, C., & Vapnik, V. (1995). Support-vector networks. Machine Learning, 20(3), 273–297.

Cover, T., & Hart, P. (1967). Nearest neighbor pattern classification. IEEE Transactions on Information Theory, 13(1), 21–27.

Hu, M.-K. (1962). Visual pattern recognition by moment invariants. IRE Transactions on Information Theory, 8(2), 179–187.

Huda, A. S. N., & Mekhilef, S. (2014). Biomass energy in Bangladesh: Current status and prospects. Renewable and Sustainable Energy Reviews, 30, 504–517.

Islam, M. R., & Rahman, M. M. (2019). A prospective assessment of biomass energy resources, potential, technologies and challenges in Bangladesh. Renewable and Sustainable Energy Reviews, 107, 217–225.

LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436–444.

Masud, M., Ananno, A. A., Arefin, A., & Ahamed, R. (2019). Perspective of biomass energy conversion in Bangladesh. Clean Technologies and Environmental Policy, 21(5), 1009–1023.

Ojala, T., Pietikäinen, M., & Mäenpää, T. (2002). Multiresolution gray-scale and rotation invariant texture classification with local binary patterns. IEEE Transactions on Pattern Analysis and Machine Intelligence, 24(7), 971–987.

Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Prettenhofer, P., Weiss, R., Dubourg, V., Vanderplas, J., Passos, A., Cournapeau, D., Brucher, M., Perrot, M., & Duchesnay, É. (2011). Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 12, 2825–2830.

Thung, G., & Yang, M. (2016). Classification of trash for recyclability status [Dataset]. Hugging Face.

Data Modeling

Article Contents

Automated Waste Classification Using Machine Learning for Sustainable Waste Management in Dhaka: A Comparative Evaluation of SVM, Random Forest, KNN, and Neural Network Approaches

Abstract

References

Recommended articles

Stay connected