Peter Johansson Classification of ECGs and Microarray Data Using Support Vector Machines Master Thesis Advisor: Mattias Olsson Abstract: Support Vector Machines are a special form of artificial neural networks, where the penalty function is chosen such that the classification problem is transformed into a convex quadratic problem. In this thesis Support Vector Machines are used on Medical data. First on two problems using ECGs as input to make diagnoses and it is also applicated on a multi-class problem, using microarray data to classify four different cancer types. December 2001 LU TP 01-36