Dynamic Sequential Pattern Recognition Applied in Medical Diagnosis

摘要

In medical diagnosis the cost of the feature measurements is important. Therefore sequential classifiers which take in account the costs of the features have to be used. In the first part of the report methods are investigated to reduce the computations as much as possible. In the cases where still to much computation time is required, suboptimal procedures must be used. Data are available on patients with pancreas cancer, patients with liver cancer and healthy people. In the last chapter some comments are given about interaction of the physician with the algorithm and topics for further research are listed. (Author)