| Verdenius, F., and Engels, R. 1997. A Process Model for Developing Inductive Applications. Proceedings of the Seventh Belgian-Dutch Conference on Machine Learning Tilburg, NL, 119-128 |
.... although Brazdil argues that it is important to do so [Brazdil, 1998] The only implemented IDA like system we are aware of was presented by Engels et al. who describe a user guidance module for DM processes called CITRUS ( Engels, 1996] Engels et al. 1997] Wirth et al. 1997] and [Verdenius and Engels, 1997]) In particular, the user guidance module uses a task method decomposition [Chandrasekaran et al. 1992] to guide the user through a stepwise refinement of a high level DM process, in order to help the user to construct the best plan using a limited model of operations. Finished plans are ....
Verdenius, F., and Engels, R. 1997. A Process Model for Developing Inductive Applications. Proceedings of the Seventh Belgian-Dutch Conference on Machine Learning Tilburg, NL, 119-128
.... predict which algorithms will be better; the rules looked at measurable characteristics of the data (e.g. number of cases, number of attributes, kurtosis) Engels et al. propose to implement a user guidance module for KD processes ( Engels, 1996] Engels et al. 1997] Wirth et al. 1997] and [Verdenius and Engels, 1997]) In particular, the user guidance module uses a task method decomposition [Chandrasekaran et al. 1992] to guide the user through a stepwise refinement of a highlevel KD process, in order to help the user to construct the best plan using a limited model of operations. Finished plans are compiled ....
Verdenius, F., and Engels, R. A Process Model for Developing Inductive Applications. Proceedings of the Seventh Belgian-Dutch Conference on Machine Learning Tilburg, NL, 119-128
....shortcomings, and at the same time deliver some support, the definition of a methodology for application of learning techniques in general is therefore an obligatory part of our work. Work that is performed towards definition of such a methodology can be found in [EEHV97] Eng96] ELS97] and [VE97] User Guidance and Support aims at a decreased development time as well as better results of executed KDD processes. Next to a description of the context in which a KDD process is to be executed (characteristics of the data to be used, requirements on the solution, etc. our approach also uses ....
F. Verdenius and R. Engels. A Process Model for Developing Inductive Applications. In W. Daelemans and P. Flach, editors, Proceedings of the 7th Belgian-Dutch Conference on Machine Learning (Benelearn), Tilburg, Netherlands, 1997. 11
....various settings, and helps to plan projects where such techniques are involved. 1 Introduction Machine Learning techniques become popular tools for solving real world problems. Recently several surveys in real world applications that exploit these techniques have been reported (Rudstrom, 1995) (Verdenius, 1997a) With the transition of these techniques from a research environment to the industry, the research focus shifts from technical issues towards the process of designing and implementing applications. In literature this development is reflected in the growing number of reports on process models ....
....lacking of a task analysis phase, and the presence of task specific analysis tools. An example of a task specific approach is that of (Brodley and Smyth, 1995) being completely configured around the classification task. Examples of task independent approaches are described in (Engels, 1996) and (Verdenius, 1997b) In these approaches derivation of a task decomposition is a central notion in the process. It often seems to be the case that technique specificity implies task specificity. Application orientation Process models can be oriented towards a specific application mode, or they can be mode ....
[Article contains additional citation context not shown here]
Verdenius, F. and Engels, R. (1997). A process model for developing inductive applications. In Daelemans, W. and Flach, P., editors, Proceedings of the 7th Belgian-Dutch Conference on Machine Learning (Benelearn), Tilburg, Netherlands.
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