Article,
HEART DISEASE DIAGNOSIS WITH TREE STRUCTURAL NAÏVE BAYES
CENTRAL ASIAN JOURNAL OF MATHEMATICAL THEORY AND COMPUTER SCIENCES, 4 (5): 134-143 (May 2023)
Abstract
Among the world population, the Disease of Heart is one of the biggest mortality and morbidity causes. This disease's precise prediction and early detection might decline rate of mortality rate certainly. Learning machines are utilized to consider several problems in the science of information. In Fortune, one of efficient methods for classification is naïve bayes (NB) is because of the ability of it for learning inherent features of data. Although, generally such method groups data with just one single that makes this less efficient relatively in several classes for big classification issue. In the article, we present the tree structural naïve bayes (Tree-NB) that classifies big classification in small classifications with utilizing structure of tree. The particular classifier is adjusted after division for every small classification. By several classifiers that are employed, Tree-NB is able to complement each other in performance of classification as well as one classifier issue is solved. As all several classifiers are end-to-end frameworks, automatically Tree-NB is able to learn nonlinear relationship among output and input data with no extraction of feature. For verifying our model validity, we compare modern methods with Tree-NB by utilizing dataset of UCI. Experimental results illustrate that Tree- NB is able to obtain the higher performance in less time of training. Average Tree- NB accuracy is 1.19 % higher than the other modern methods also it possesses higher average recall and precision.
