Research 

Home ► Research ►

Current Research

I am currently pursuing a PhD in extending rough set theory and augmenting the Rough Set Attribute Reduction (RSAR) algorithm to operate incrementally on uncertain and mutable data, using probabilistic methodologies. The application testbed is proposed to be email categorisation. I am studying under the expert supervision of Dr Qiang Shen and Dr Miles Osborne and the auspices of the Approximate and Qualitative Reasoning Group of the Centre for Intelligent Systems and their Applications (CISA) of the Division of Informatics of the University of Edinburgh.

PhD Proposal

This is my PhD proposal. It is effectively a report of work done during the first year of the degree, and an outline of work to be done in the remaining time.

The document is available in compressed Postscript Postscript (229 kbytes) for printing on black and white printers, or PDF (735 kbytes) for viewing or printing on colour printers.

More information will appear here as it is made available.

Research Interests

My research interests generally orbit around rough set theory; fuzzy logic and fuzzy systems; information filtering and retrieval; machine learning and knowledge elicication; and genetic algorithms and evolutionary computing in general.

Publications 

Q. Shen and A. Chouchoulas. A fuzzy-rough approach for generating classification rules. Pattern Recognition, 2002 (to appear).

A. Chouchoulas and Q. Shen. Rough Set-Aided Keyword Reduction for Text Categorisation. Journal of Applied Artificial Intelligence, 15(9):843-873, 2001.

J. Márin-Blazquez, A. Chouchoulas and Q. Shen. An overview of recent approaches to fuzzy modelling. Proceedings of the 2001 UK Workshop on Computational Intelligence, 49-56, 2001.

Q. Shen and A. Chouchoulas. Rough Set-Based Dimensionality Reduction for Supervised and Unsupervised Learning. Applied Mathematics and Computer Science, Special Issue on Rough Sets and their Applications, 11(3):101-119, 2001.

Q. Shen and A. Chouchoulas. FuREAP: A Fuzzy-Rough Estimator of Algae Populations. Artificial Intelligence in Engineering, 15(1):13-24, 2001.

Q. Shen and A. Chouchoulas. Selection of Features in Transparent Fuzzy Modelling. To appear in Proceedings of FUZZ-IEEE 2001.

A. Chouchoulas and Q. Shen. Rough Set-Based Dimensionality Reduction for Multivariate Adaptive Regression Splines. Proceedings of the 2nd International Conference on Rough Sets, pages 112-119, 2000.

Q. Shen and A. Chouchoulas. A Modular Approach to Generating Fuzzy Rules with Reduced Attributes for the Monitoring of Complex Systems. Engineering Applications of Artificial Intelligence, 13(3):263-278, 2000.

Q. Shen and A. Chouchoulas. Knowledge-Based Fault Detection in Industrial Plants Supported by Rough-Fuzzy Learning. Proceedings of the IFAC Symposium on Fault Detection, Supervision and Safety for Technical Processes, pages 669-674, 2000.

A. Chouchoulas and Q. Shen. A Rough Set-Based Approach to Text Classification. Proceedings of the 7th International Workshop on Rough Sets (Lecture Notes in Artificial Intelligence, No. 1711), pages 118-127, 1999.

Q. Shen and A. Chouchoulas. Combining Rough Sets and Data-Driven Fuzzy Learning for Generation of Classification Rules. Pattern Recognition, 32(12), pages 2073-2076, 1999.

Q. Shen and A. Chouchoulas. Data-Driven Fuzzy Rule Induction and Its Application to Systems Monitoring. Proceedings of the 8th International Conference on Fuzzy Systems, pages 928-933, 1999.

A. Chouchoulas and Q. Shen. Rough Set-Assisted Rule Induction for Plant Monitoring, Proceedings of the 4th International Joint Conference on Information Science, pages 316-319, 1998.

Past Research 

A Rough Set Approach to Text Classification

My MSc (by Research) dissertation dealt with a means of simplifying text classification domains through the use of rough set theory, and the Rough Set Attribute Reduction (RSAR) algorithm in particular. This was tested experimentally using supervised email categorisation as an example, with several different IF/IR classification methodologies.

The MSc thesis is available in PDF (2323 kbytes) and compressed Postscript (1351 kbytes).

Generating Fuzzy Classification Rules from Crisp Examples

Between January 1998 and December 2000, I worked as a Research Associate for the Approximate and Qualitative Reasoning Group.

The research project revolved around Fuzzy System, Rule Induction and Data Mining. Dr. Qiang Shen was the grant holder; the project itself was funded by NCR. The research looked into Generating Fuzzy Classification Rules from Crisp Examples.

A Genetic Algorithm-Based Information Filter for Usenet

My BSc dissertation was based on a hybrid information filter for Usenet. The system used reinforcement-based learning in conjunction with a genetic algorithm to evolve interface agents that could estimate what topics interest the user. The agents also explored new domains for information that could possibly interest the user, thus applying both exploitation and exploration concepts to the search space.

The dissertation is available in compressed Postscript (349 kbytes).