Logo

CIRG - Research  -  Data Mining 



[Bullet] Home
About
News
[Bullet]

- NN
- DM
- SI
- EC
- MAS
- AIS
- IA
- Bioinf
- Games
- Opt
- FA
- Industry

Publications
People
Resources
Links
Contact Us

OVERVIEW

The data and text mining focus area has the objective of developing new techniques for knowledge discovery and to improve existing techniques. The focus area is also active in applying data mining techniques to solve real-world problems in consultation to South African industries.

Some of the questions being addressed are how to mine knowledge from data with continuous classes, how to cope with extremely large databases, more efficient data clustering methods and how to extract knowledge in environments where data changes over time. Tools are currently under development which address these questions.

ACTIVE MEMBERS

List the current members actively doing research in this focus area. [ Show ]

ALUMNI MEMBERS

P Lutu

PhD Completed in 2010

A Louis

M.Sc Started in 2006

E Dean

M.Sc Started in 2002

G Nel

M.Sc Completed in 2005

E Papacostantis

M.Sc Started in 2004
Hons-B.Sc Completed in 2003

G Potgieter

M.Sc Completed in 2003
Hons-B.Sc Completed in 2001

D Rodic

PhD Completed in 2005
M.Sc Completed in 1999

GROUP PUBLICATIONS

List publications of this research focus area. [ Show ]

MEMBER PROFILE



 Name:

 Evangelos Papacostantis

Portrait photo

 E-mail:

 Evangelos.Papaconstantis@rmb.co.za

 Group(s):

 Swarm Intelligence
 Evolutionary Computation
 Finance

 

 Degree specific information: M.Sc

 Title:

 Co-Evolutionary Approach to Probabilistic Game Learning using PSO

 Abstract:

The domain of complex board games has been under the microscope for many years in the field of AI. These games provide an ideal testing ground to explore a number of AI techniques, concepts and approaches. Games that are more representative of real world problems are probabilistic games or otherwise called non-deterministic games. These games have imperfect information, meaning that a players' actions within the game is determined by random/probabilistic elements. Co-evolution techniques have successfully been applied to these games, allowing competitive and reasonably intelligent agents to immerge. Co-evolution techniques allow unintelligent/random agents to compete against each other and incrementally learn from better performing agents. This research will generally investigate how co-evolution and PSO techniques can be used to find competitive probabilistic game playing agents. Different PSO topologies are going to be examined together with different co-evolution schemes. Backgammon and Poker are the two probabilistic games which are going to be used for evaluation purposes.

 Supervisor / Co-Supervisor:

 AP Engelbrecht
 N Franken

 Thesis:

 Not available for download yet.

 

 Degree specific information: Hons-B.Sc

 Title:

 Development of advanced data analysis/data mining tool

 Abstract:

The cerebral cortex is arguably the most fascinating structure in all of human physiology. The ability to associate items according to their similarities/differences was first pointed out by Aristotle. This project includes the creation of an advanced data analysis/data mining tool, by attempting to simulate the cerebral cortex. It will employ self-organizing feature maps, that compress high dimensional data into two dimensional maps, accomplished by different unsupervised training algorithms. It will allow features such as association, classification, path finding and forecasting. A number of visual techniques will be implemented on the two dimensional maps to reveal all these features. The tool will be further more improved by adding knowledge exploration features with the use of decision trees, rule inductions and genetic algorithms.

 Supervisor / Co-Supervisor:

 AP Engelbrecht

 Thesis:

 Not available for download yet.




You are visitor #153243
Contact webmaster
Back to top

QualNet Network Simulator University Program Valid XHTML 1.0! Valid CSS!


Computational Intelligence Research Group
University of Pretoria
Copyright © 2017