197 0

Interval Type-2 퍼지접근방법을 이용한 Competitive Learning

Title
Interval Type-2 퍼지접근방법을 이용한 Competitive Learning
Other Titles
Interval Type-2 Fuzzy Approach to Competitive Learning
Author
이상수
Alternative Author(s)
Lee, Sang Soo
Advisor(s)
이정훈
Issue Date
2009-08
Publisher
한양대학교
Degree
Master
Abstract
Fuzzy Competitive Learing(FCL)은 주변 패턴의 특성을 고려하지 않고 항상 일정하게 학습률 파라미터를 감소 시키면서 학습을 하기 때문에 학습 수렴 속도가 느리며 최적의 학습 결과를 얻기 어렵다는 단점이 있다. 이와 같은 단점을 보완하기 위해 본 논문에서는 학습률 파라미터에 Interval type-2 방법을 적용하여 학습을 진행하는 동안 주변 패턴에 특성에 따른 적절한 학습률을 계산하는 방법을 제안한다. 제안한 Interval type-2 FCL 알고리즘은 기존의 FCL알고리즘과 비교했을 때 학습 수렴 속도가 빠르고 학습 결과도 좋았다.; Fuzzy Competitive Learning (FCL) has some drawbacks such as slow learning due to the use of a constant learning rate. In this thesis, we propose a method to estimate the learning rate in accordance with the pattern characteristics by applying interval type-2 fuzzy sets. By doing so, the proposed algorithm shows the ability to learn faster and obtain improved clustering results compared to FCL. Several experimental results are shown.
URI
https://repository.hanyang.ac.kr/handle/20.500.11754/143526http://hanyang.dcollection.net/common/orgView/200000412092
Appears in Collections:
GRADUATE SCHOOL[S](대학원) > ELECTRONIC,ELECTRICAL,CONTROL & INSTRUMENTATION ENGINEERING(전자전기제어계측공학과) > Theses (Master)
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

BROWSE