Solution of Fuzzy Pattern Recognition Inverse Problem
The fuzzy pattern recognition is a kind of popular pattern recognition presently. This paper studies the fuzzy pattern recognition two basic model of inverse problem. (1) The standard model of fuzzy vector set race, and to identify the sample for the sing
- PDF / 455,226 Bytes
- 7 Pages / 439.37 x 666.142 pts Page_size
- 88 Downloads / 201 Views
Solution of Fuzzy Pattern Recognition Inverse Problem Cuilan Mi, Xinchun Wang and Jianming Liu Abstract The fuzzy pattern recognition is a kind of popular pattern recognition presently. This paper studies the fuzzy pattern recognition two basic model of inverse problem. (1) The standard model of fuzzy vector set race, and to identify the sample for the single point vector, (2) The standard model and to identify samples are fuzzy vector set clan. Put forward the weight calculation method and an example are presented. Keywords The fuzzy pattern recognition • Fuzzy vector set family • The inverse problem • Weight
21.1 Introduction Pattern recognition is compared and matching the identification of the object to be characteristic information and mode library characteristic information, and give target belong to the judgment of the model. In the usual pattern recognition, the model is clear, clear, for sure. But there are a lot of problems, the model itself is not very clear, has some fuzzy. And the fuzzy pattern recognition is refers to in pattern recognition, model is fuzzy, namely standard mode library provided in the fuzzy model test. The fuzzy pattern recognition in the engineering problems to a wide range of applications. These applications to more positive, negative meaning application is rare. In this paper, the fuzzy pattern recognition two basic models of the inverse problem is presented, and the solving methods. C. Mi (*) · X. Wang College of Science, Hebei United University, Hebei, e-mail: [email protected] X. Wang e-mail: [email protected] J. Liu Mathematics and Information Science Department, Tangshan Normal College, Tangshan, Hebei, China e-mail: [email protected]
W. Du (ed.), Informatics and Management Science I, Lecture Notes in Electrical Engineering 204, DOI: 10.1007/978-1-4471-4802-9_21, © Springer-Verlag London 2013
155
C. Mi et al.
156
21.2 The Fuzzy Pattern Recognition Basic Model From the Recognition model of recognition process, fuzzy pattern recognition method can be divided into two types: direct method and indirect method, the direct method according to maximum subjection principle of classification, and indirect method according to choose nearly principle classified [1]. Maximum subjection principle: A domain is on the U a fuzzy subsets of m (m model), and compose a standard model base is U x0 fixed element, if m
A˜ i 0 (x 0 ) = ∨
k=1
A˜ k (x 0 )
(21.1)
then think the relative membership x 0 in Ai 0 fuzzy subsets nearly principle: Set A˜ 1 , A˜ 2 , · · · , A˜ m is domain of m a U fuzzy subset (m model), and compose a standard model base, Stay recognition model is fuzzy subsets, if m
˜ = ∨ σ ( A˜ k , B) ˜ σ0 ( A˜ i 0 , B)
(21.2)
k=1
Argue that B˜ is closest to A˜ i 0 , σ is the close degree. About the definition of close degree, a lot of literature are discussed, Among them the literature [2] discusses more in detail, the commonly used close degree have case close degree, distance of press close At present, the common problem of the fuzzy pattern recognition acc
Data Loading...
