Author/Editor		Kavalar, MS; Povalej, P; Homšak, M; Kokol, P; Štiglic, MM
Title		Recognizing the most important risk factors in diagnosing pediatric asthma with decision trees
Type		članek
Source		In: Sepiashvili R, editor. Asthma: from genes to clinical management. Proceedings of the 17th world asthma congress; 2003 Jul 5-8; Saint-Petersburg. Bologna: Monduzzi editore, International proceedings division,
Publication year		2003
Volume		str. 235-9
Language		eng
Abstract		Despite the frequency with which it occurs, pediatric asthma is often not diagnosed correctly or soon enough. In this paper we present the results of an intelligent data analysis based on decision trees used for determining the most important risk factors for diagnosis of asthma in very young children because of inability to perform pulmonary function tests, which are the golden standard in older children. The study included 106 Slovenian children aged 2-8 years. All decision trees were highly aecurate in a sense of classifying unseen test cases. The most interesting decision tree classified unseen test cases with 97.3% accuracy. Many well-known decision pathways were shown and more importantly some new interesting patterns, which deserve to be more carefully examined, were exposed.
Descriptors		ASTHMA DECISION MAKING, COMPUTER-ASSISTED DECISION TREES CHILD