A COMPARATIVE STUDY OF MACHINE LEARNING TECHNIQUES FOR IMAGE CLASSIFICATION

Machine learning is a promising tool for developing systems for various applications, such as medical diagnosis and computer vision. In this study, we analyze the performance of five different algorithms on the MNIST database for image classification. The five are CNN, SVM, KNN, RF, and DBN. The algorithms' performance is evaluated using various metrics. According to our experimental results, CNNs performed the best on the MNIST database when it comes to image classification, followed by SVM, KNN, RF, and DBN, with 98.2%, 99.1%, and 98.2% accuracy, respectively. Different algorithms performed well on different metrics. The findings of our study suggest that CNNs is an ideal technique for image classification on MNIST, as other algorithms such as RF and SVM performed well. The decision-making process regarding which algorithm to use depends on the application and the dataset's characteristics. Our study, which analyzed the performance of five machine learning algorithms, serves as a valuable guide for practitioners and researchers in the field.