Michalski, R.S., Imam, I.F.: Learning problem-oriented decision structures from decision rules: the AQDT-2 system. Proceedings of 8th International Symposium Methodologies for Intelligent Systems. Lecture Notes in Artificial Intelligence 869. Springer Verlag, Heidelberg (1994) 416--426  Home/Search   Document Details and Download   Summary   Related Articles   Check

....are obtained. 1 Introduction Decision trees and decision rule systems are widely used in different applications as algorithms and as a form of knowledge representation. Problems of comparative analysis of decision trees and decision rule systems are interesting both for theory and for practice [2, 6]. In this paper we consider relationships between parameters of a decision rule system and the minimal depth of a decision tree which solves the problem of the search of all realizable rules from the system. The necessity to find all realizable rules arises, for example, if we consider problems ....

Michalski, R.S., Imam, I.F.: Learning problem-oriented decision structures from decision rules: the AQDT-2 system. Proceedings of 8th International Symposium Methodologies for Intelligent Systems. Lecture Notes in Artificial Intelligence 869. Springer Verlag, Heidelberg (1994) 416--426

....say, four leaves, the partitioning algorithm tries to partition the instances directly into the four leaves, thus creating a composite node. CHAPTER 6. OBLIVIOUS READ ONCE DECISION GRAPHS 186 Y Y Y Y N N N N N Y Figure 6.16: A Pylon. Question marks denote tests at the branching nodes. Michalski Imam (1994) and Imam (1995) suggested the use of decision structures, which are decision trees with complex tests at nodes, disjunctive values on the edges, and probabilistic classifications at the leaves. Polynomially size decision structures are more powerful than univariate trees because of the wider ....

....because the accuracy was very low (m of n 3 7 10 just predicts the majority class) A GLD provides a way of displaying up to about ten dimensions in a graphical representation that can be easily understood. GLDs were described in Michalski (1978) and later used in Thrun et al. 1991) Wnek Michalski (1994), and Kohavi (1994d) Each possible instance in the projected space defined by the decision table s schema has exactly one box that is shaded according to the prediction made there (all GLDs shown below have only two shades of grey for the classes, except for DNA which has three classes and thus ....

Michalski, R. S. & Imam, I. F. (1994), Learning problem-oriented decision structures from decision rules: The aqdt-2 system, in Z. W. Ras & M. Zemankova, eds, "Proceedings of the 8th International Symposium on Methodology for Intelligent Systems (ISMIS-94)", Lecture Notes in Artificial Intelligence 914, Springer Verlag, pp. 416--426.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC