Hypertensive patient intelligent question-answering system based on knowledge graph and establishment method thereof

Abstract

The invention relates to a hypertensive patient self-comprehensive management intelligent question-answering system based on a knowledge graph and an establishment method thereof. The system comprisesa user system interaction layer, a voice text conversion layer, a matching reasoning layer, an information search layer and a resource layer, wherein the resource layer is used for realizing data acquisition, knowledge extraction, knowledge fusion and processing and knowledge updating to form a framework of the knowledge graph; the user system interaction layer is used for carrying out language interaction with a user and answering questions raised by the user; the voice text conversion layer is used for completing conversion between voice and text; the knowledge graph can analyze the questions presented by the natural language of the user, understand the questions of the user and complete knowledge reasoning; the information search layer is used for searching or completing reasoning in the resource layer based on the text converted by the voice text conversion layer; and the matching reasoning layer is used for matching the information searched by the information search layer with the questions raised by the user, outputting matched answers after matching succeeds, and outputting specific answers if matching fails.

Description

technical field [0001] The invention relates to the field of medical and health technologies, in particular to an intelligent question-and-answer system for self-comprehensive management of hypertensive patients based on a knowledge graph and a method for establishing the same. Background technique [0002] In China, the prevalence of hypertension among residents aged 18 and over is 25.2%, the treatment rate is 41.1%, and the control rate is 13.8%. A low control rate greatly increases the risk of developing hypertensive complications. Foreign clinical trials have shown that good blood pressure control can reduce the risk of stroke by 30%-43%, and reduce the risk of myocardial infarction by 15%, which can save a lot of social medical resources and greatly reduce the pressure on patients. [0003] Active community intervention and self-management are effective ways to control hypertension. Among them, grass-roots medical and health institutions (community health service cent...

AI Technical Summary

Problems solved by technology

[0009] ① The existing technology does not have the active supervision function of the intelligent system

This process relies on users to actively ask questions, but there is no system to actively intervene in the process of patient self-management. It cannot actively remind and supervise patients to complete healthy behaviors. It plays a small role in patients' lives, and the effect is limited.

[0010] ② Existing technology cannot guide the self-comprehensive management of hypertensive patients

The lack of these health behavior management makes the function of the existing technology very limited, which greatly affects the effect of helping hypertensive patients to manage themselves

[0011] ③Limited scenarios that can be solved by existing technologies

The scope of the hypertension knowledge base and the question-and-answer knowledge base mentioned in the prior art only includes blood pressure values, symptoms, special populations, and diet types, and the breadth of knowledge is small, resulting in limited answerable questions for hypertensive patients

[0012] ④ The existing technology is based on deep learning, and the user experience is limited

The deep learning model needs to rely on large-scale labeled data, and can only do simple pattern recognition. When it comes to any content that requires reasoning, the deep learning model cannot achieve

Therefore, the existing technology cannot make reasonable reasoning based on the user's questions and the equipped knowledge base, which is likely to bring an unpleasant communication experience to the user.

[0013] ⑤ The existing technology does not consider the problem of individualization

The existing technology does not consider the individual needs of users, and does not collect and store user information. The system can only answer universal or even fixed-mode questions, and cannot understand the characteristics of users.

This leads to the system not being able to store the user's changes, and unable to provide information for the user's long-term disease management

[0014] ⑥Existing technology does not consider the language of popularization

Existing technologies often use professional terminology to build knowledge bases, which inevitably contain obscure content. In addition, the effect of machine learning is limited, and most of the hypertensive patients are middle-aged and elderly people, which may bring difficulties for users to understand

Images