| As the development of information technology, database technique is capable of processing and storing more and more large amount of data. On the other hand, the current techniques of discovering and extracting the knowledge hiding in the data and using the extracted knowledge to support decision-making are far from satisfactory. In this thesis, a set of pattern recognition approaches and methodologies are proposed for effective knowledge discovery and uncertainty reasoning from large and complex datasets.;Recently advances in biotechnology allow researchers to measure expression levels of thousands of genes simultaneously. Analysis of data produced by such experiments offers potential insights into gene functions and regularity mechanisms. There are thousands of methods are proposed to the gene expression data analysis, but the results are various. Uncertainties contained in the dataset may be one of the causes. The methodologies proposed in the thesis attempt to attenuate the effects of such uncertainties by proper association of multi-faceted measurements. Experimental results show that the research is useful on handling the large amount of highly intertwining datasets in knowledge discovery.;Keywords. Pattern Recognition, Microarray Gene Analysis, Uncertainty Reasoning, Classification, Gene Set Analysis. |
