E. Bloedorn, R. Michalski, and J. Wnek. Matching methods with problems: A comparative analysis of constructive induction approaches. Technical report, Machine Learning and Inference Laboratory, MLI 94-2, George Mason University, Fairfax, VA., 1994.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....Kernels A useful distinction between di#erent classes of kernels is based on driving force . We distinguish between semantics, syntax, model, and data as the driving force of the kernel definition. A similar terminology has been used previously in the context of constructive induction algorithms [2] Semantics is the ideal driving force for the definition of proximities. It is related to so called knowledge driven approaches of incorporating expert knowledge into the domain representation. Syntax is often used in typed systems to The search strategy determines how the hypothesis space is ....

.... 1 if z is is a member of the concept and 1 otherwise) that performs ideally. We distinguish the following three issues crucial to good kernels: completeness, correctness, and appropriateness. A similar terminology has been used previously in the context of constructive induction algorithms [2]. Completeness refers to the extent to which the knowledge incorporated in the proximity is su#cient for solving the problem at hand. A proximity is said to be complete if it takes into account all the information necessary to represent the concept that underlies the problem domain. Correctness ....

E. Bloedorn, R. Michalski, and J. Wnek. Matching methods with problems: A comparative analysis of constructive induction approaches. Technical report, Machine Learning and Inference Laboratory, MLI 94-2, George Mason University, Fairfax, VA., 1994.

....(KDD) it is very often the pre processing stage, namely transformation and reduction of data, that play an important role. A good transformation reduction technique may result in new attributes that are more appropriate for a data mining algorithm. A wide range of constructive induction techniques [6, 7, 9, 15] has been explored aiming at enriching the language of propositional learners [4, 5] The usual way is to add new attributes that are computed from existing ones using a priori given functions. One important class of functions includes linear combinations of numerical attributes. To prevent a ....

Bloedorn, E., Michalski, R.S. and Wnek, J., "Matching Methods with Problems: A Comparative Analysis of Constructive Induction Approaches," Reports of the Machine Learning and Inference Laboratory, MLI 94-2, School of Information Technology and Engineering, George Mason University, Fairfax, VA, May 1994.

....space expansion is very useful when the attribute construction operators are well matched with the problem at hand. Representation space contraction, however, must be performed with great care, as it may lead to a removal of information that is crucial for learning a correct hypothesis [5]. For this reason, default thresholds on the space contraction operators are set conservatively. The following sections describe these operators in more detail. 4.2.1 Space Expansion: Attribute Construction By GENERATE The GENERATE method for constructing new attributes employs both mathematical ....

Bloedorn, E., Michalski, R.S., and Wnek, J., "Matching Methods with Problems: A Comparative Analysis of Constructive Induction Approaches", Reports of the Machine Learning and Inference Laboratory, MLI 94-2, George Mason University, Fairfax, VA, 1994.

....one operator, or without any automated representation space modification. 4. An Empirical Comparison 4.1 Descriptions of methods evaluated In order to determine the effectiveness of different CI methods, a set of experiments was performed. These experiments are described in greater detail in (Bloedorn, et. al, 1994). This set of experiments samples a wide variety of possible learning problems including: misclassification noise, attribute value noise, overprecision, inappropriate attributes and irrelevant attributes. In all of these experiments the AQ15c program was used as the learning algorithm (Wnek et ....

Bloedorn, E., Michalski, R.S. and Wnek, J., "Matching Methods with Problems: A Comparative Analysis of Constructive Induction Approaches," Reports of the Machine Learning and Inference Laboratory, MLI 94-2, Center for AI, George Mason University, Fairfax, VA, 1994.

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