Inducing a Membership Function of a Category in a Fuzzy-Set-Based Attribute Space
Lists
Only you see this
Abstract
Learning from examples has been studied widely. A new approach is presented by Narazaki and Ralescu, where by assigning suitable numerical values to any of the old symbolic attribute values, they give a method to make examples belonging to the same category (concept) concentrated. This paper is an improvement to their study. Instead of a real value from an interval, a vector of fuzzy membership values is used for the assignment in this paper. The vector is obtained based on building a fuzzy set base for each attribute. The new approach presented here involves Narazaki and Ralescu's method as a special case. Its effectiveness is simply explained and verified with three experiments. Results of the experiments show that the new approach is superior to the previous studies.
BibTeX
@article{liu1997inducing,
author = {Weijie Liu and Michio Sugeno},
title = {{Inducing a Membership Function of a Category in a Fuzzy-Set-Based Attribute Space}},
journal = {International Journal of Uncertainty Fuzziness and Knowledge-Based Systems},
year = {1997},
doi = {10.1142/s0218488597000130},
}Loading related content...