Review and Comparison of Evaluation Indicators Based on Support Vector Machine Classification

Support Vector Machine (SVM) is a widely-used supervised learning algorithm with extensive applications in classification problems. It is crucial to use appropriate evaluation metrics to assess the performance of SVM classification models. This article reviews and compares commonly used SVM classification evaluation indicators, including accuracy, precision, recall, Fl score, ROC curve, and AUC value. The comparison of classification results using these indicators shows variation across different datasets. Thoughtful selection of appropriate evaluation indicators is essential for accurate model evaluation. This research can assist researchers and practitioners in choosing an optimal number of datasets for a more precise evaluation of SVM classification model performance.