Markus Ringnér, Patrik Edén and Peter Johansson Classification of expression patterns using artificial neural networks In A Practical Approach to Microarray Data Analysis (eds. D.P. Berrar, W. Dubitzky and M. Granzow, Kluwer Academic Publishers), 201-215 (2002) Abstract: We present an artificial neural network based method for classification of gene expression data. This method is successfully applied to the example of classifying small round blue-cell tumors into distinct diagnostic categories. The key components of this classification procedure are principal component analysis for dimensional reduction and cross-validation to optimize the training of the classifiers. In addition, we describe a way to rank the genes according to their importance for the classification. Random permutation tests are introduced to asses the significance of the classification results. LU TP 02-25