Michie, D., Problem decomposition and the learning of skills, in Proceedings of the European Conference on Machine Learning, pp. 17-31, Springer-Verlag, 1995.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....of examples to construct rules for intermediate concepts in the hierarchy. In comparison with standard decision tree induction techniques, structured induction exhibits about the same classification accuracy with the increased transparency and lower complexity of the developed models. Michie [26] emphasized the important role of structured induction in the future and listed several real problems that had been solved in this way. Mozetic [27, 28, 7] employed another scheme for structuring the learning problem. That approach was particularly aimed at automated construction of system models ....

D. Michie. Problem decomposition and the learning of skills. In N. Lavrac and S. Wrobel, editors, Machine Learning: ECML-95, Notes in Artificial Intelligence 912, pages 17--31. Springer-Verlag, 1995.

....approach. In the third part we examine the algorithm on several artificial data and real applications. 2 Feature Decomposition Approach The purpose of decomposition methodology is to break down a complex problem into several manageable problems. In artificial intelligence, according to Michie [30], finding a good decomposition is a major tactic both for ensuring the transparent end product and for avoiding the combinatorial explosion. Decomposition methodology can be considered as effective strategy for changing the representation of a learning problem. In fact Kusiak [22] consider ....

....of neural networks and decision trees on several multi class problems from the UCI repository. Function decomposition was originally developed in 50 s for improving the design of switching circuits. Recently this approach has been adopted by the machine learning community. Shapiro [38] and Michie [30] used a manual decomposition of the problem and an expert assisted selection of examples to construct rules for the concepts in the hierarchy. In comparison with standard decision tree induction techniques, structured induction exhibits about the same classification accuracy with the increased ....

[Article contains additional citation context not shown here]

Michie, D., "Problem decomposition and the learning of skills," in Proceedings of the European Conference on Machine Learning, Springer-Verlag, PP. 17-31, 1995.

....examples is used or an expert is consulted to build a corresponding decision tree. In comparison with standard decision tree induction techniques, Shapiro s approach exhibits about the same classification accuracy with the increased transparency and lower complexity of the developed models. Michie [12] emphasizes the important role the structured induction will have in the future development of machine learning and lists several real problems that were solved in this way. The work presented here is based on our own decomposition algorithm [23] in which we took the approach of Curtis [7] and ....

D. Michie. Problem decomposition and the learning of skills. In N. Lavrac and S. Wrobel, editors, Machine Learning: ECML-95, Notes in Artificial Intelligence 912, pages 17--31. Springer-Verlag, 1995.

....of examples to construct rules for intermediate concepts in the hierarchy. In comparison with standard decision tree induction techniques, structured induction exhibits about the same classification accuracy with the increased transparency and lower complexity of the developed models. Michie [25] emphasized the important role of structured induction in the future and listed several real problems that had been solved in this way. Mozetic [26, 27, 7] employed another scheme for structuring the learning problem. That approach was particularly aimed at automated construction of system models ....

D. Michie. Problem decomposition and the learning of skills. In N. Lavrac and S. Wrobel, editors, Machine Learning: ECML-95, Notes in Artificial Intelligence 912, pages 17--31. Springer-Verlag, 1995.

....of constructive induction [24] The original goal of constructive induction was to adding task specific features to the example representations to improve the accuracy and understandability of the learned concept. The induction of a hierarchy of intermediate concepts termed structured induction [25,26] is also similar. Here the domain expert is involved in attribute selection for decomposing a problem into sub problems recursively and then induction is done bottom up. Our work differs from these previous ones in the following way. Earlier, while an expert provided rules for adding these ....

D. Michie. Problem decomposition and the learning of skills. In Machine Learning: ECML-95, Lecture Notes in Artificial Intelligence, 912, (eds. N. Lavrac and S. Wrobel), Berlin, Heidelberg, New York: Springer Verlag, pp. 17-31, 1995.

.... Science, the acquisition of skills without the availability of explicit knowledge has been reported in (Stanley et al. 1989; Lane, 1988) In Robotics and Machine Learning, the bottom up approach is the most widely followed one (Liu and Asada, 1993; Lee et al. 1994; Urbancic and Bratko, 1994; Michie, 1995; Reignier et al. 1995; Kaiser and Dillmann, 1996) However, only few researchers working in the latter fields have considered the step of abstraction, i.e. the generation of explicit, possibly declarative knowledge from skills that are learned in a bottom up manner. Recent work in this area ....

Michie, D. (1995). Problem decomposition and the learning of skills. In Proceedings of the European Conference on Machine Learning, Heraklion, Crete, Greece.

No context found.

Michie, D., Problem decomposition and the learning of skills, in Proceedings of the European Conference on Machine Learning, pp. 17-31, Springer-Verlag, 1995.

No context found.

Michie, D.: Problem decomposition and the learning of skills. In Lavrac, N., Wrobel, S., eds.: Machine Learning: ECML-95. Notes in Artificial Intelligence 912. Springer-Verlag (1995) 17--31

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