FIT Božetěchova 2, seminární místnost UPGM L220 24.10.2008

Many publications are addressed to microarray-based prediction and demonstrate that microarray data are informative. An ignored matter is how good microarray data in comparison with clinical parameters for outcome prediction is. Related topic is whether a combination of these both sources can give better prediction results than just classifiers based on either clinical parameters or microarray data. This work considers two methods that combine clinical parameters and microarray data. The first method uses Random forest that has suitable features for prediction with microarray data. The second method joins up logistic regression and boosting.