Research Group Project and Description (cont'd)
GAFSW: Genetic Algorithms for Feature Selection and Weighting in Character Recognition Systems	Nawwaf Kharma, Faten Hussein, Rabab Ward Feature weighting is the general case of feature selection, and hence should perform better than feature selection, at least in some situations. The initial purpose of this study was to test the validity of this hypothesis within the context of character recognition systems. However, we ended up carrying out two sets of studies, which in turn produced some unexpected but justified results. The first set compares the performance of Genetic Algorithm (GA)-based feature selection to GA-based feature weighting, under various conditions. The second set of studies evaluates the performance of the better method (which turned out to be feature selection) in terms of optimal performance and time. The results of these studies show that (a) feature set selection prior to classification is important for k-nearest neighbour classifiers, in the presence of redundant or irrelevant features; and (b) that GAs are effective methods for feature selection. However, their scalability to highly-dimensional problems, in practice, is still an open problem.
Novel Set-Theoretic Definitions of Common Fuzzy Hedges, Theory and Application	Hongjian Shi, Nawwaf Kharma, Rabab Ward The subject of this study is fuzzy linguistic hedging, used to modify membership functions. Our investigation will exceed the traditional definition given by Zadeh(and others), upon which our research is based. We will present new and more general definitions for four of the most commonly used hedges. These hedges are very, more or less, positively and negatively. The effect of applying each hedge to a membership function will be described both qualitatively and quantitatively in this paper. We will also describe conditions under which a specific hedge is valid, chart the examples, and explore the properties of the hedge operations. A software program called FuzzyMarkx has been developed for understanding and applying the above four hedges to list of data. It is specially designed for altering school grade distributions. It allows users to view the new grades, the grade distribution as well as the graphical representation of the hedge being applied.
PalmPrints: A Novel Co-Evolutionary AlgorithmFor Clustering Finger Images	Nawwaf Kharma, Ching Y. Suen, and Pei F. Guo  The purpose of this study is to explore an alternative means of hand image classification, one that requires minimal human intervention. The main tool for accomplishing this is a Genetic Algorithm (GA). This study is more than just another GA application; it introduces (a) a novel cooperative co-evolutionary clustering algorithm with dynamic clustering and feature selection; (b) an extended fitness function, which is particularly suited to an integrated dynamic clustering space. Despite its complexity, the results of this study are clear: the GA evolved an average clustering of 4 clusters, with minimal overlap between them.
	Page 1, Page 2, Page 3, Page 4
DEC Lab Home Last Updated April 6, 2004