MIME-Version: 1.0 Content-Location: file:///C:/B13362B3/6SS.htm Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset="us-ascii" SPECIAL SESSION

IFSA 2007 = World Congress

Cancun, Me= xico, June 18-21, 2007<= /p>

&n= bsp;

&n= bsp;

SPECIAL= SESSION 6SS

 

Fuzzy Tree= s

 

Organizers: Zhiheng Huang and Masoud = Nikravesh

{zhiheng, nikravesh}@cs.berkeley.edu

Electrical Engineering and Computer

Science University of California= at Berkeley

CA 94720, USA

Phone: 1-510-643-4523 (Zhiheng) and 1-510-643-4522 (Masoud)

Fax: 1-510-643-4519

 

SCOPE: The main aim of th= is special session is to bring together researchers and scientists who have the same interest in fuzzy trees, which include fuzzy decision trees, fuzzy signatures, fuzzy multiple aggregator trees, pattern trees and other form of trees such as fuzzy hierarchical structures.

 

Fuzzy decision tree induc= tion is the fuzzy extension of classical decision tree induction. Fuzzy decision tr= ees have been widely used in varied applications and consistently show good accuracy in classification applications. Fuzzy signatures have been propose= d to model complex structures of the data points using different aggregation operators such as MIN, MAX and AVERAGE. Fuzzy multiple aggregator trees whi= ch make use of more choices of aggregation operators have been introduced and = the tree induction method has been proposed by means of genetic programming. Pattern trees have been suggested to represent the structures of classes via different aggregations over fuzzy terms. The experimental work shows they c= an outperform fuzzy decision trees.

 

Fuzzy trees are an effect= ive and efficient tool for fuzzy classification and decision analysis. Conventional fuzzy rules can be extracted from fuzzy trees, and therefore transparency a= nd comprehensibility are preserved. In addition, tree representations are able to significantly reduce the computation expense at the test stage, compared to conventional fuzzy rule bases.

 

The objective of the spec= ial session is to provide an opportunity to exchange ideas on the theoretical research and/or applications of fuzzy trees. It would be interesting to com= pare different fuzzy trees to identify the advantages and drawbacks of each fuzzy tree. Hopefully the discussion in this special session will lead to new res= earch directions for fuzzy trees.

 

The organizers welcome contributions on the fields of = (but not limited to):

 

- fuzzy decision trees

- fuzzy signatures

- fuzzy multiple aggregator= s trees

- pattern trees

- fuzzy hierarchical struct= ures

- other forms of trees

 

We encourage submissions of papers that report on  theoretical<= /span> advances in fuzzy trees as well as papers that concern applications of such fields.<= /p>

 

Proposed contribution sho= uld be sent by e-mail to the organizers of the session by October 16, 2006. The same deadlines as for the regular papers to be submitted to the IFSA 2007 World Congress will be observed. Submitted papers must comply to the IFSA 2007 World Congress guidelines for submission. All submitted papers will be refereed by experts in the field.