Krantz, D., Luce, R., Suppes, P, & Tversky, A. (1971). Foundations of measurement Vol. 1. Academic Press New York  Home/Search   Document Not in Database   Summary   Related Articles   Check

....the claim that this method lessens the burden of knowledge acquisition and opens new perspectives for knowledge management. 1. Introduction Knowledge has become a crucial asset that no organization can afford to disregard. In an industrial setting, management of knowledge comes in many forms [10]. We take the view that the best way when starting to analyze and to model processing of knowledge is to take a look at humans who do this job: They learn, collect experience, solve problems, invent, produce, control, monitor, plan, negotiate and decide. In modern industry, however, the central ....

....significantly explains the data. The random structure is set up by taking the relation y randomly follows x and by instantiating it with all concepts. 1 In mathematics and psychometrics a structure made up only of a universe and various relations is usually called a relational structure [10]. Thus, the concept structures used in knowledge tracking may also be called relational structures. 2 While the first example of a concept structure is usually labeled a semantic network, the second one is generally called a frame. 3 2 3 5 4 5 3 .3 .4 .3 .1 .2 .1 .3 t m . a) b) c) ....

Krantz, D., Luce, R., Suppes, P, & Tversky, A. (1971). Foundations of measurement Vol. 1. Academic Press New York

.... 1995) it is known, that the function h with minimal risk is the Bayes s optimal function: h (d; d 0 ) ae 1 if P ( ffl j(d; d 0 ) 1 2 Gamma1 otherwise : 3) However, the Bayesian approach for preference learning is inconsistent, because stochastic transitivity may not hold (Suppes et al. 1989). We will demonstrate this fact by the following example. Let us consider a document space with 27 documents. The documents are described by three distinct feature vectors D = fd; d 0 ; d 00 g, which separate the document space into three sets of nine documents each, one set for each feature ....

Suppes, P.; Krantz, D. H.; Luce, R. D.; and Tversky, A. 1989. Foundations of Measurement Vol. II. San Diego: Academic Press Inc.

....learning of preference relations reduces to a standard classification problem if pairs of objects are considered. This, however, is not true in general because the properties of transitivity and asymmetry may be violated by traditional Bayesian approaches which may violate stochastic transitivity (Suppes et al. 1989). Considering pairs of objects, the task of learning reduces to finding a utility function that best reflects the preferences induced by the unknown distribution PXY . Our learning procedure on pairs of objects is an application of the large margin idea known from data dependent Structural Risk ....

....Therefore we call this problem also the problem of preference learning. It was shown that the Bayes optimal decision function on pairs of objects can result in a function p which is no longer transitive on X (Herbrich et al. 1998) This is also known as the problem of stochastic transitivity (Suppes et al. 1989). Note also that the demand of transitivity and asymmetry effectively reduces the space of admissible classification functions p acting on pairs of objects. Distribution independent bounds on R s pref The main advantage of the distribution independent model for ordinal regression is given by the ....

Suppes, P., D. H. Krantz, R. D. Luce, and A. Tversky (1989). Foundations of Measurement Vol. II. San Diego: Academic Press Inc.

....of learning a preference relation reduces to a classification problem: The classifier has to assign pairs (x (1) x (2) of objects from the space X to classes and OE. The Bayes s optimal h , however, may not fulfill conditions (2) and (3) because stochastic transitivity may not hold [9]. Let us consider a space X of 27 objects with three different representations x 1 ; x 2 and x 3 . Each object is assigned one out of no. of objects per feature vector rank x 1 x 2 x 3 r 5 0 0 4 r 4 5 0 0 r 3 0 9 0 r 2 0 0 5 r 1 4 0 0 Table 1 Example for the violation of stochastic ....

P. Suppes, D. H. Krantz, R. D. Luce, and A. Tversky. Foundations of Measurement Vol. II. Academic Press Inc., San Diego, 1989.

No context found.

Patrick Suppes, David H. Krantz, R. Duncan Luce, and Amos Tversky. Foundations of Measurement Vol. II. Academic Press Inc., San Diego, 1989.

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