Sergei I. Gorbachev

this paper evaluates the performance of three data analysis techniques-decision trees, neural networks, and self-organizing Kohonen maps-in identifying cardiovascular disease patterns. Based on a dataset of 1,328 patient records, the models were compared regarding classification accuracy for target variables and demographic attributes. The neural network achieved the lowest error rate for disease prediction (3.61%), while the decision tree showed robust performance (6.17%) with higher interpretability. The findings underscore the necessity of method selection based on data specificity and the...

this article explores the application of rule-based machine learning, specifically decision trees, for predicting corporate bankruptcy within a hierarchical framework. Using a dataset from the Taiwan Economic Journal (1999–2009), the study evaluates the economic performance indicators that critically influence financial distress. The decision tree model achieved an error rate of 1.6%, outperforming neural networks (2.67%) and Kohonen maps (2.98%). The most significant factors identified include Persistent EPS (22.4%), Working Capital to Equity ratio (20.1%), and Pretax Income to Capital (15.5%...