Detail Nonparametric fuzzy regression and its applications Bibliografi Author: Cheng, Chi-Bin ; Lee, E. Stanley (Advisor) Topik: ENGINEERING; INDUSTRIAL|OPERATIONS RESEARCH|ENGINEERING; SYSTEM SCIENCE Bahasa: (EN )    ISBN: 0-591-86971-3     Penerbit: KANSAS STATE UNIVERSITY     Tahun Terbit: 1998     Jenis: Theses - Dissertation Fulltext: 9833770.pdf (0.0B; 0 download) Abstract Various approaches for performing fuzzy regression analysis have been developed in the past decade. Most of them assume the functional forms used to define the regression relation are possibilistic linear equations with fuzzy coefficients. Thus, the problem is reduced to the estimation of fuzzy parameters (coefficient) with a given functional form. However, in practice, the true functional form which defines the relationship between dependent and independent variables is frequently unknown and highly nonlinear and complicated. Therefore, there is a need to develop approaches, known as nonparametric approaches, which do not require the assumption of the regression functional relationship. Three new nonparametric approaches for fuzzy regression analysis are proposed in this dissertation. The first one borrows the classical techniques from statistics, in which two of the most commonly used methods, k-nearest neighbor and kernel smoothing, are fuzzified to carry out the fuzzy regression analysis. The second approach uses the fuzzy radial basis function network, which is a fuzzification of an ordinary radial basis function neural network. The third approach is a modification of a Sugeno fuzzy inference system and it is further converted into a fuzzy adaptive network. The latter two approaches conduct the nonlinear regression in more implicit manners and they both have learning abilities. These two approaches also have better performances in high dimensions and their structures enable us to incorporate constraints in the fuzzy regression analysis. Numerical examples are used to demonstrate these three approaches. A survey of the applications of fuzzy regression analysis in various fields, including economics, engineering, sales forecasting, ergonomics and quality improvement, are carried out. Two new applications to a welding quality design problem and an environment design problem are developed. These two problems are modeled by fuzzy regression through the approach of fuzzy adaptive network; and then the optimal settings of their input parameters are found by the fuzzy response surface methods based on the regression models built. Opini Anda Klik untuk menuliskan opini Anda tentang koleksi ini! Lihat Sejarah Pengadaan  Konversi Metadata   Kembali

Copyright © 2006, 2007 Unika Atma Jaya, all rights reserved