Sequential classification for microarray and clinical data

Sequential classification uses in a stepwise process only part of the data (evidence) for partial classification, i.e., classifying only objects with sufficient evidence and leaving the rest unclassified. In the following steps the procedure is repeated using additional data until all objects are classified. This is especially useful when data become available only at certain points in time, as in surgical decision making, i.e., clinical patient data, lab data, or cDNA microarray expression data from tissue samples become available before, during and after the operation. Surgeons are interested in classifying patients into low or high risk groups, which might need special measures, e.g., prolonged intensive care.