Comparative Analysis of Machine Learning and Deep Learning Algorithms for Brain Stroke Classification: A Literature Survey

This literature review examines and compares the Deep Learning (DL) and Machine Learning (ML) algorithms for brain stroke classification. Each method’s performance was evaluated using accuracy, precision, recall and F1-score metrics. Deep Learning models have great accuracy if applied to large datasets from a variety of angles, while Machine Learning performs well on small datasets. The main aim of this survey is the futuristic development of AI-based solutions for brain stroke image classification related to future healthcare research. The study, therefore, compares different ML and DL algorithms performance based on the accuracy of brain stroke classification. In this comparative analysis, the two well-known algorithms, SVM (ML algorithm) and CNN (DL model), are selected for the practical evaluation of classification accuracy. The benchmarking of the CNN and the SVM model was performed using the 'Brain_Data_Organised CT Image Dataset' dataset, which is a publicly available dataset from Kaggle. The CNN model achieves a training accuracy of 99% and testing accuracy of 94% after 20 epochs, while the SVM model has attained a classification accuracy of 86.4%. The results demonstrate that the deep learning models perform better than traditional machine learning algorithms in terms of accuracy and predictive power.