THE USEFULNESS OF A MACHINE LEARNING APPROACH TO KNOWLEDGE ACQUISITION

This paper presents results of experiments showing how machine learnin g methods are useful for rule induction in the process of knowledge ac quisition for expert systems. Four machine learning methods were used: ID3, ID3 with dropping conditions, and two options of the system LERS (Learning from Examples based on Rough Sets): LEM1 and LEM2. Two know ledge acquisition options of LERS were used as well. All six methods w ere used for rule induction from six real-life data sets. The main obj ective was to test how an expert system, supplied with these rule sets , performs without information on a few attributes. Thus an expert sys tem attempts to classify examples with all missing values of some attr ibutes. As a result of experiments, it is clear that all machine learn ing methods performed much worse than knowledge acquisition options of LERS. Thus, machine learning methods used for knowledge acquisition s hould be replaced by other methods of rule induction that will generat e complete sets of rules. Knowledge acquisition options of LERS are ex amples of such appropriate ways of inducing rules for building knowled ge bases.