Enhancing COVID-19 detection accuracy with decision tree and support vector machine learning models

The purpose of this study was to determine how effectively the innovative decision tree algorithm could detect COVID-19 in comparison to the support vector machine technique. The Components and Procedures: A Kaggle data source library dataset contains 8,400 data values, which are utilised in the inquiry that has been recommended. Two sets are created from the data that was obtained: the training set, which has 6720 records and accounts for 80 percent of the total, and the testing set (1680, or 20 percent ). The accuracy, sensitivity, and specificity of a decision tree are each taken into consideration while determining its score. According to the findings, the decision tree, with a 97.00 percent accuracy rate, is able to correctly identify and detect a greater number of targets compared to the support vector machine, which has a 73.4760 percent accuracy rate. The significance value for both accuracy and loss is 0.001 (p < 0.05). The independent sample T-test was used to acquire this result, and it indicates that there is statistical significance between the two methods. This result was obtained by examining independent samples. In addition, the programmeClinCalc that was utilised for the purpose of determining the sample size maintained the G power at 0.80 and alpha at 0.05, taking into consideration both accuracy and loss. When compared to the support vector machine approach, the performance of decision trees is superior in the COVID-19 identification dataset that was used in this work.