Diagnosis of sleep disorders in traditional Chinese medicine based on adaptive neuro-fuzzy inference system

Traditional Chinese medicine (TCM) relies on a combination of the four diagnostic methods of inspection, listening and smelling, inquiry, and palpation to diagnose sleep disorders. This method relies on the doctor's practice experience and his mastery of TCM theory and is subjective in nature, it is necessary to make the TCM diagnosis objective. We propose using an Adaptive neuro-fuzzy inference system (ANFIS) to diagnose sleep disorders in TCM, a method with adaptive reasoning capabilities and objectivity. Sleep disorder symptoms are first compressed by a genetic algorithm based on mutual information. According to the diagnostic process of TCM, the model has four inputs: inspection(A), listening and smelling(O), inquiry(F), and palpation(P). We quantified the four linguistic variables A, O, F, P and fed them into the model, which was trained to generate the membership of each input variable, and the model used the derivation rules to form the output to complete the diagnosis. After testing, the accuracy of the proposed model was 97.6%. In a test comparison with six TCM doctors with more than 10 years of experience, our proposed model tested correctly in all 100 medical cases. The diagnoses of the six doctors were consistent and correct for 92 cases, however, there were 8 cases where their diagnoses diverged, indicating that our proposed model has the ability of objective reasoning. Our work provides a reference for the objectification of sleep disorders in TCM diagnosis.

Automated summary

The study introduces the use of an Adaptive neuro-fuzzy inference system (ANFIS) to diagnose sleep disorders in Traditional Chinese Medicine, achieving an accuracy of 97.6% and 100% correct testing results in 100 medical cases. Comparison tests with six TCM doctors show that the proposed model has the ability of objective reasoning in diagnosing sleep disorders.