SIGMOIDAL THEORY

被引：10

作者：

DRAKOPOULOS, JA

机构：

[1] Stanford University, Department of Computer Science, Knowledge Systems Laboratory, Palo Alto, CA 94304-0106

来源：

FUZZY SETS AND SYSTEMS | 1995年 / 76卷 / 03期

关键词：

FUZZY SETS; MEMBERSHIP FUNCTIONS; ANALYSIS; NEURAL NETWORKS; PATTERN RECOGNITION;

D O I：

10.1016/0165-0114(95)00028-5

中图分类号：

TP301 [理论、方法];

学科分类号：

081202 ;

摘要：

Sigmoidal theory is a new theory that defines and analyzes a new family of functions called sigmoidal functions. A number of theorems proved show that there is an infinite number of classes of sigmoidal functions and provide us with a method of constructing those classes. Subsequently, multi-sigmoidal functions can be constructed using classes of sigmoidal functions. The applications of sigmoidal functions range from fuzzy sets to neural networks and pattern recognition. In particular, sigmoidal functions are shown to be very appropriate functions to model fuzzy set membership. The formalization of sigmoidal functions here is expected to increase the interest in them and result in a number of new applications.

引用

页码：349 / 363

页数：15

共 16 条

[1]

[Anonymous], 1987, LEARNING INTERNAL RE

[2]

DASGUPTA B, 1993, ADV NEURAL INFORMATI, V5, P615

[3]

Dawson M. R. W., 1992, Connection Science, V4, P19, DOI 10.1080/09540099208946601

[4]

DRAKOPOULOS JA, 1994, KSL9442 STANF U KNOW

[5]

DRAKOPOULOS JA, 1995, 4TH INT C ART NEUR N

[6]

DRAKOPOULOS JA, 1991, KSL9175 STANF U KNOW

[7]

GUPTA MM, 1975, FUZZY LOGIC KNOWLEDG, P69

[8] CONNECTIONIST LEARNING PROCEDURES [J].

HINTON, GE .

ARTIFICIAL INTELLIGENCE, 1989, 40 (1-3) :185-234

[9]

Kaufmann A, 1975, INTRO THEORY FUZZY S

[10]

Lapedes A., 1987, LA UR 87 2662 C 8706

← 1 2 →