Using Support Vector Machines instead of Naive Bayes, a New Model for Predicting Complications from Black Fungus Infection

Abstract

Aim: To enhance the accuracy in early prediction and detection of black fungus infection using Novel Support Vector Machine algorithm compared with Naive bayes. Materials and Methods: This study contains two groups i.e, Novel Support Vector Machine algorithm and Naive bayes. Each group consists of a sample size of 20 and the study parameters include alpha value 0.05, beta value 0.2, and the G power value 0.8. Their accuracies are compared with each other using different sample sizes also. Results: The Statistical SVM 87.7% more accurate than the Naive Bayes algorithm of 78.3% in identifying the black fungus in humans with significance value 0.001 (p=0.05). Conclusion: The SVM model is significantly better than the NB algorithm in identifying black fungus infection in humans. It can be also considered as a better option for the identification of black fungus infection.