부산대학교 도서관

Machine learning classification algorithms and their applications are becoming increasingly popular in today's era of information science. Choosing an algorithm that is appropriate for the problem and application is a challenging task. The quality of the prediction model is significantly influenced by selecting the right algorithm. The purpose of the study and comparative analysis highlights the benefits and drawbacks of various classification algorithms from all perspectives. The Machine Learning (ML) based classification algorithms such as Logistic Regression (LR), K-Nearest Neighbour(KNN), DT (Decision Tree), SVM (Support Vector Machine), and Naïve Bayes (NB) are assessed for comparative analysis. These algorithms were tested and analysed using various datasets acquired and used from the UCIML repository. Algorithms are evaluated using well-established effective measures for accuracy, recall, and precision. A variety of datasets are used to monitor, measure, and compare each algorithm's performance in terms of metrics. Each algorithm's effectiveness and efficacy are analysed and compared from all aspects. Based on overall performance, the algorithms are ranked in the following order: logistic regression, decision tree, SVM, KNN, and Nave Bayes.