An Analysis of Support Vector Machines and Decision Trees in Handwritten Digits Recognition

• Author(s): Kabiru Labaran Bala ; Dr Musa Argungu ; Dr Dan Lami Gabi ; Aminu Labaran ; Abubakar Abduraruaf
• Paper ID: 1710047
• Page: 377-385
• Published Date: 18-08-2025
• Published In: Iconic Research And Engineering Journals
• Publisher: IRE Journals
• e-ISSN: 2456-8880
• Volume/Issue: Volume 9 Issue 2 August-2025

Abstract

A Support Vector Machines (SVMs) and Decision Trees in handwritten digits recognition. The primary research aim is to find performance and suitability of these algorithms. Most of the time, the physically constructed digits are not inside the domains of proportionate size, due to the similarities between the numbers, such as 1 and 7, 5 and 6, 3 and 8, 2 and 5, 2 and 7. The performance and accuracy of machine learning algorithms heavily depend on the quality and diversity of the dataset used for training and testing. This research limited by the availability of suitable handwritten digit datasets (MNIST), which impact the generalizability of the results. Using MNIST dataset consists of a large number of 28x28 pixel grayscale images of handwritten digits (0 through 9) and 60,000 Training Images and 10,000 Testing. In this study, different machine learning methods, which are SVM and Decision Trees architectures are used to achieve high performance on the digit string recognition problem. In these methods, images of digit strings are trained with the SVM and Decision Trees model methods structure by sliding a fixed size window through the images labeling each sub-image as a part of a digit or not. The research ultimately revealed that Support Vector Machines was the classifier, with 98% accuracy. The lowest score of accuracy goes to Decision Tree with 96% of accuracy and we Recommend the Use of SVM as the Best Algorithms for the handwritten digits recognition but also we suggest that combination of difference Datasets for more accuracy of the recognition MNIST dataset a most of U.S written style..

Keywords

Machine Learning, Support Vector Machine, Decision Tree, MNIST, digit recognition and Data Set.

Citations

IRE Journals:
Kabiru Labaran Bala , Dr Musa Argungu , Dr Dan Lami Gabi , Aminu Labaran , Abubakar Abduraruaf "An Analysis of Support Vector Machines and Decision Trees in Handwritten Digits Recognition" Iconic Research And Engineering Journals Volume 9 Issue 2 2025 Page 377-385

IEEE:
Kabiru Labaran Bala , Dr Musa Argungu , Dr Dan Lami Gabi , Aminu Labaran , Abubakar Abduraruaf "An Analysis of Support Vector Machines and Decision Trees in Handwritten Digits Recognition" Iconic Research And Engineering Journals, 9(2)