Doctor recommendation method

The invention discloses a doctor recommendation method. Comprising the following steps: collecting text data containing previous patient chief complaint symptoms and corresponding treatment departments; preprocessing the text data to obtain text vectorized data; constructing a text classification model by adopting a convolutional neural network and a Transform model; selecting and optimizing parameters of the text classification model to construct a department recommendation model; based on a department recommendation model, inputting a chief complaint symptom of the target patient to obtain a recommended department; performing multi-dimensional matching according to the matching degree of the target patient and the doctor; and positioning all doctors in the recommended department, calculating the matching degree of the target patient and the doctors by adopting a doctor-patient matching degree model, and recommending the doctors to the target patient according to the matching degree. According to the method,