Construction Method Of Fuzzy Decision Trees For Identification The Computer System State
Svitlana Gavrylenko, Viktor Chelak, Oleksii Hornostal · 2022
The efficiency of using fuzzy decision trees to identify the state of a computer system has been studied. The method of fuzzy classification is offered. Based on it, a model was created, trained and tested with real data According to the results of the research, a method of identifying the state of a computer system using the classifier of fuzzy decision trees with a special procedure for the formation of fuzzy sets and membership functions is proposed. The software in which the offered technique of the decision of a problem of identification of a condition of computer system is realized and investigated is developed. The evaluation of the efficiency of the developed classifier in comparison with standard machine learning algorithms is performed. Prospects for further research are the development of an ensemble of decision-making and fuzzy decision-making trees based on boosting and bagging using meta-algorithm improvements and a method for calculating classifier weights, as well as a special final decision-making procedure based on common and fuzzy classifier decisions.