Research On Disease Information Extraction Technology In Online Intelligent Consultation System

With the rapid development of Internet technology and its products,online inquiry has become a new way of inquiry gradually integrated into people's lives.Through the network health consulting platform,people through the online doctor to submit their own disease information,they can easily get free helps,such as diagnosis,medical knowledge,the doctor's advice and so on.However,online health consulting platform is also faced with the lack of online physician resources,platform consulting procedures bothered,online response results quality and other issues.According to the mental health problems,this paper focuses on how to extract the key disease information from the text of the psychological problems in the network effectively based on text information extraction technology.It will realize the online intelligent diagnosis system by constructing a text information extraction system.In this paper,the research focuses on the disease information extraction of the patient's readme text,mainly in the following areas:Firstly,the obvious disease entity extracted.We can discover the online text not only has the simple disease information but also includes other entity information,such as the patient's gender,age,duration of disease and so on,if you analyze the patient's readme.In this paper,the information is classified into obvious entity information.So we present a method for extracting explicit disease entities from a patient's readme text based on conditional random field model.This method can solve the problem of long distance entity feature in the readme text preferably.In the experiment,the method achieves good results of 80.54%.Secondly,the hidden physical entity extracted.In the patient readme text,it contains the hidden information.These information need to derive.The main factor is that the patient description is not professional.For this kind of disease information,this paper proposes a method based on sentence vector similarity calculation.The method can achieve the classification of statements in the text and obtain the hidden information contained in the statement according to the classification results.It focuses on vector similarity calculation.In the open experiment,the method has achieved satisfactory results of 71.27%.Finally,based on the above theoretical research,from the point of view of the project,this paper completes the design and implementation of the disease information extraction subsystem in the online diagnosis system.And there is a simple and intuitive display of early forms of system.