| |
Profile |
|
Bai Qifeng
PHD student
N320 CSIT
Department of
Computer Science
ANU, Australia
Bai.qifeng@anu.edu.au |
|
| View Profile Details |
| |
My Topic |
|
|
Fuzzy Information Retrieval
Interdiscipline topic includes: Fuzzy logic and Artifical intelligence, Natural language processing, and part of Search Technology on WWW.
|
|
|
| |
Blog |
|
|
2 Mar 2006
Books I am reviewing
Speech and Language Processing: An Introduction to Natural Language Processing, Computational Linguistics, and Speech Recognition
Search Engine and Information Retrieval
Artificial Intelligence- A modern approach
Artificial Intelligence and Fuzzy Control
Fuzzy Control
J2EE WebService
JavaServer Faces
Hiberante
Working Schedule
Tues 16:00 - 24:00
Thur 16:00- 24:00
Sat 18:00 - 24:00
|
Add a comment | Read comments
|
11 Nov 2005
Abstract of my master thesis
The hierarchical fuzzy signature, like a tree structure, is vector valued fuzzy sets, where each vector component can be a further vector valued fuzzy set. In order to construct fuzzy signature, for a specific disease, the first thing should be to classify the group members of which present similar symptoms. However, the result from the commonly used K-Means Clusters Analysis is not very satisfactory. It is therefore that this project uses Fuzzy Cluster Method Clustering (FCMC) to achieve better results rather than that of K-means. Through an adaptive searching approach, FCMC should be able to generate appropriate clusters and the membership degrees of data points, which can convey more meaning to support feasible interpretation. After getting the specific disease suspect group, we need to explore the common features in this group, which will help us to construct fuzzy signatures. Factor Analysis (FA) explores or establishes correlational structure among the observed clusters. In that way, measured variables (symptoms) may be reconstructed by a small set of parameters, which could form sub-trees. Also, the results can represent the underlying structure in a concise and interpretable form. Then, possible fuzzy signatures which represent the underlying structure of the disease can build hierarchical rule bases and be used for further research. We construct hierarchical fuzzy rule bases based on the fuzzy signature and possible weighted aggregations among the fuzzy signature sets are implemented to test the applicability of the fuzzy signatures. |
Add a comment | Read comments
|
|
|
|
