Abstract

Abstract. Although successful in medical diagnostic problems, inductive learning systems were not widely accepted in medical practice. In this paper two di erent approaches to machine learning in medical appli-cations are compared: the system for inductive learning of decision trees Assistant, and the naive Bayesian classi er. Both methodologies were tested in four medical diagnostic problems: localization of primary tumor, prognostics of recurrence of breast cancer, diagnosis of thyroid diseases, and rheumatology. The accuracy of automatically acquired diagnostic knowledge from stored data records is compared and the interpretation of the knowledge and the explanation ability of the classi cation process of each system is discussed. Surprisingly, thenaiveBayesian classi er is superior to Assistant in classi cation accuracy and explanation ability, while the interpretation of the acquired knowledge seems to be equally valuable. In ad-dition, two extensions to naive Bayesian classi er are brie y described: dealing with continuous attributes, and discovering the dependencies among attributes.

Citations

4403	A mathematical theory of communication - Shannon - 1948
2888	Induction of Decision Trees - Quinlan - 1986
682	The CN2 induction algorithm - Clark, Niblett - 1989
296	The multi-purpose incremental learning system aq15 and its testing application to three medical domains - Michalski, Mozetic, et al. - 1986
291	Rule induction with cn2: Some recent improvements - Clark, Boswell - 1991
175	Adaptive Pattern Recognition and Neural Networks - Pao - 1988
137	On changing continuous attributes into ordered discrete attributes - Catlett - 1991
104	Learning by being told and learning from examples: an experimental comparison of the two methods of knowledge acquisition in the context of developing an expert system for soybean disease diagnosis - Michalski, Chilausky - 1980
96	Unknown attribute values in induction - Quinlan - 1989
88	Discovering rules by induction from large collections of examples - Quinlan - 1979
78	Probability and The Weighing of Evidence - Good - 1950
72	Rule Induction Using Information Theory - Smyth, Goodman - 1991
50	The Estimation of Probabilities - Good - 1965
48	Inductive knowledge acquisition: A case study - Quinlan, Compton, et al. - 1987
40	Estimating Probabilities: A Crucial Task - Cestnik - 1990
29	New measurements highlight the importance of redundant knowledge - Gams - 1989
28	Experiments in automatic learning of medical diagnostic rules - Konenko, Bratko - 1994
24	A method for attribute selection in inductive learning systems - Baim - 1988
16	Information based evaluation criterion for classifier's performance - Kononenko, Bratko - 1991
14	The deepening of expert systems - Steels - 1988
12	An expert system for the interpretation of thyroid assays in a clnical laboratory - Horn, Compton, et al. - 1985
7	Bayesian neural networks - Kononenko - 1989
5	Learning If Then Rules in Noisy Domains - Clark, Niblett - 1986
5	Inductive Acquisition of Expert Knowledge. Turing - Muggleton - 1990
3	An Experiment in Automatic Learning of Diagnostic Rules - Bratko, Mulec - 1980
3	Informativity-Based Splitting of Numerical Attributes into Intervals - Cestnik - 1989
3	Wong A.K.C., Automatic Construction of Expert Systems from Data: A Statistical Approach - Chan - 1989
3	An Experimental Comparison of Inducing Decision Trees and Decision - Elomaa, Holsti
3	Torasso P., Learning of Fuzzy Production Rules for Medical Diagnoses - Lesmo, Saitta - 1982
3	Decision Tree Induction using Domain Knowledge - Nunez - 1990
2	Kononenko I., Learning Rules from Incomplete and Noisy - Bratko - 1987
2	Dezelic G., Tezak-Bencic M., Durrigl T., Medical Decision Making Using InductiveLearning Program - Kern
2	ID3, sequential Bayes, naive Bayes and Bayesian neural networks - Kononenko - 1989
2	Semi-naiveBayesian classi er - Kononenko - 1991
2	An Experiment in Machine learning of Redundant Knowledge - Kononenko - 1991
2	Janc T., Induction of Reliable Decision - Kononenko, Jauk - 1988
2	Automatic construction of diagnostic rules - Mozetic, Lavrac, et al. - 1986
2	Verordnung zur digitalen Signatur (Signaturverordnung -- SigV). Entered into force as from November 1st - unknown authors - 1997
2	Bundesamt fr Sicherheits - unknown authors - 1983
1	Mozetic I.,Lavrac N.,KARDIO: A study in deep and qualitative knowledge for expert systems, Cambridge,MA - Bratko - 1989
1	Bratko I.,Learning redundant rules in noisy domains - Cestnik - 1988
1	Kononenko I.& Bratko I., ASSISTANT 86 : Aknowledge elicitation tool for sophisticated users - Cestnik - 1987
1	ITRULE: An Information Theoretic Rule-Induction Algorithm - Goodman, Smyth
1	Pirnat V., Signi cance Level Based Classi cation with Multiple Trees - Karalic - 1991
1	Feedforward Bayesian Neural Networks and Continuous Attributes - Kononenko - 1991
1	Machine learning and knowledge acquisition. In: International Handbook of Information Technology and Automated O ce Systems - Michie - 1985
1	Al Attar, Use of Sequential Bayes with Class Probability Trees - Michie, A - 1991
1	Common Criteria for Information Technology Security Evaluation (CC), Version 2.0 - unknown authors - 1998
1	DINV 66391-1: Spezifikation der Schnittstelle zu Chipkarten mit Digitaler Signatur-Anwendung / Funktion nach SigG und SigV - unknown authors - 1998