Benchmarking Classical and Deep Learning Models for Handwritten Digit Classification

  • Unique Paper ID: 181773
  • Volume: 12
  • Issue: 1
  • PageNo: 5579-5584
  • Abstract:
  • Handwritten digit recognition is an important task in image classification, with real-world uses in areas like banking, education, and postal systems. Accurately recognizing digits can help automate systems and reduce human errors. In this study, we used the MNIST dataset, which contains 60,000 images of handwritten digits (0 to 9), and is commonly used to test and compare image classification models. We compared the performance of three popular machine learning models—K-Nearest Neighbors (KNN), Support Vector Machine (SVM), and Multi-Layer Perceptron (MLP)—with a deep learning model called Convolutional Neural Network (CNN). Before training, we applied normalization to the data and used Principal Component Analysis (PCA) to reduce dimensions for better model performance. Among all the models tested, CNN performed the best with an accuracy of 99.16%, followed by KNN (97.69%), MLP (97.36%), and SVM (91.70%). We also looked at other metrics like precision, recall, and F1-score, and used visual tools like confusion matrices and bar plots to compare results. This study shows that deep learning (CNN) is more effective than traditional machine learning methods for recognizing handwritten digits and can be a strong foundation for future work in this field.

Copyright & License

Copyright © 2025 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

BibTeX

@article{181773,
        author = {Vaishnavi S Nikalje and Prof. Vishnupant Potdar and Dr. Nagnath Biradar},
        title = {Benchmarking Classical and Deep Learning Models for Handwritten Digit Classification},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {1},
        pages = {5579-5584},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=181773},
        abstract = {Handwritten digit recognition is an important task in image classification, with real-world uses in areas like banking, education, and postal systems. Accurately recognizing digits can help automate systems and reduce human errors. In this study, we used the MNIST dataset, which contains 60,000 images of handwritten digits (0 to 9), and is commonly used to test and compare image classification models. We compared the performance of three popular machine learning models—K-Nearest Neighbors (KNN), Support Vector Machine (SVM), and Multi-Layer Perceptron (MLP)—with a deep learning model called Convolutional Neural Network (CNN). Before training, we applied normalization to the data and used Principal Component Analysis (PCA) to reduce dimensions for better model performance. Among all the models tested, CNN performed the best with an accuracy of 99.16%, followed by KNN (97.69%), MLP (97.36%), and SVM (91.70%). We also looked at other metrics like precision, recall, and F1-score, and used visual tools like confusion matrices and bar plots to compare results. This study shows that deep learning (CNN) is more effective than traditional machine learning methods for recognizing handwritten digits and can be a strong foundation for future work in this field.},
        keywords = {Handwritten Digit Recognition, MNIST, Convolutional Neural Network (CNN), K-Nearest Neighbors (KNN), Multi-Layer Perceptron (MLP), Support Vector Machine (SVM), Principal Component Analysis (PCA), Image Classification},
        month = {June},
        }

Cite This Article

  • ISSN: 2349-6002
  • Volume: 12
  • Issue: 1
  • PageNo: 5579-5584

Benchmarking Classical and Deep Learning Models for Handwritten Digit Classification

Related Articles