A brain tumor diagnostic system with automatic learning abilities

Abstract

A medical diagnostic system, BTDS (the brain tumors diagnostic system), is presented. BTDS consists of an expert system and a learning system. The former can help diagnosticians in judging the causes of brain tumors according to computed tomography pictures, and the latter, based on a revised inductive learning method, is especially suitable in the case of brain tumors. By applying the proposed learning mechanism, the initial knowledge base can be constructed and new rules can be learned from noise data in the process of inference. The consideration and process of system implementation is introduced, and the learning mechanism is presented. An example is given to illustrate BTDS, and the test results are shown to evaluate the performance of the system.<>