Author/Editor		Gamberger, Dragan; Lavrač, Nada; Grošelj, Ciril
Title		Experiments with noise filtering in a medical domain
Type		članek
Source		In: Bratko I, Džeroski S, editors. Machine learning. Proceedings of the 16th international conference (ICML'99); 1999 Jun 27-30; Bled. San Francisco: Morgan Kaufmann publishers,
Publication year		1999
Volume		str. 143-51
Language		eng
Abstract		The paper presents a series of noise detection experiments in a medical problem of coronary artery disease diagnosis. The following algorithms for noise detection and elimination are tested: a saturation filter, a classification filter, a combined classification filter, a combined classification-saturation filter, and a consensus saturation filter. thedistinguishing feature of the novel consensus saturation filter is its high reliability which is due t the multiple detection of potentially noisy examples. Reliable detection of noisy examples is important for the analysis of patient records in medical databases, as wel as for the induction of rules from filtered data, representing genuine characteristics of the diagnostic domain. Medical evaluation in the problem of coronaty artery disease diagnosis shows that the dected noisy examples are indeed noisy or non-typical class representatives.
Descriptors		CORONARY DISEASE ARTIFICIAL INTELLIGENCE ALGORITHMS ELECTROCARDIOGRAPHY CORONARY ANGIOGRAPHY