A model to classify the disease pattern based on similarity degree method

In order to help human expert resolve the problem of diagnosing disease, we analyze the comparability and relativity between pattern recognition and disease diagnosis in terms of the solution means, and propose the theoretical model of disease-similarity-degree pattern recognition on the basis of certainty factors vectors and fuzzy membership factors vectors, and its corresponding data structure mode. In addition, the software hierarchy of model and recognition algorithm, and its practice method are designed. Field experiment statistics demonstrate that: compared with the individual human expert, the proposed model be able to obtain a favorable accuracy rate of diagnosis over 85%, and reduce a rate of misdiagnosis effectively, which provided with a preferential comprehensive diagnosis performance.