Online department triage model training method and device

Abstract

The embodiment of the invention provides an online department triage model training method and device, and the method comprises the steps: firstly carrying out the pre-training of a prediction model through a corpus text in a general field, obtaining a trained first prediction model, carrying out the further training through a corpus text in a medical field, and obtaining a second prediction model; and finally, in order to complete the tasks of department triage and hospital guide, training the second prediction model again through the prediction objective function and the triage objective function, ending the training when the prediction objective function and the triage objective function both reach convergence, and removing the prediction objective function to obtain the department triage model. Through three times of training in the above steps, a large number of different corpus texts are respectively processed, and the accuracy of the department triage model can be effectively improved. Meanwhile, due to the improvement of the accuracy, a registered or doctor-seeing department can be recommended to the user after information interaction with the patient for a small number oftimes, and the registration and doctor-seeing efficiency is improved.

Description

technical field [0001] This application relates to the field of information interaction, and specifically relates to an online department triage model training method and device. Background technique [0002] Due to the popularization of Internet technology, more and more hospitals have launched online registration and consultation services. However, when conducting online registration or consultation, most patients may not know which department they should choose for registration or consultation, which wastes a lot of time and leads to low efficiency of online registration and consultation. [0003] In order to solve the above problems, the current triage and guidance system usually needs to perform multiple information interactions (up to 10 times) with the patient, and some even require the user to select body parts on the interactive interface. The operation is more complicated. Nevertheless, the current triage and guidance system is still unable to recommend accurate ...

AI Technical Summary

Problems solved by technology

However, when conducting online registration or consultation, most patients may not know which department they should choose for registration or consultation, which wastes a lot of time and leads to low efficiency of online registration and consultation

[0003] In order to solve the above problems, the current triage and guidance system usually needs to perform multiple information interactions (up to 10 times) with the patient, and some even require the user to select body parts on the interactive interface. The operation is more complicated

Nevertheless, the current triage and guidance system is still unable to recommend accurate departments for patients to register or consult

At the same time, the departments of the current triage and guidance system are not detailed enough, and patients under the age of 6 are directly assigned to the Department of Pediatrics, but the actual Department of Pediatrics also includes many sub-departments, which need to be further divided

Images