Method and system for constructing intelligent question answering system based on deep learning and knowledge graph

Abstract

The invention discloses a method and system for constructing an intelligent question-and-answer system based on deep learning and knowledge graphs. A crawler is used to obtain medical inquiry data sets from the Internet, and data preprocessing is performed to obtain labeled data sets; A word segmentation dictionary in the medical field, and merged with a medical dictionary as a systematic word segmentation dictionary; construct a knowledge map related to diseases and symptoms, and perform disease entity alignment and symptom entity alignment; obtain labeled data sets based on disease entity alignment; build depth-based Learned language model; build a knowledge map-based query optimization algorithm that combines user context information; build a training data set for language model and knowledge map fusion and perform model fusion training to obtain a pre-diagnosis fusion model based on language model and knowledge map. Based on deep learning and knowledge graphs, the present invention realizes active consultation interaction combined with user complaint information, and disease pre-diagnosis based on user complaint and consultation information.

Description

technical field [0001] The present invention relates to the field of deep learning, visualization and question answering systems, in particular to a method and system for constructing an intelligent question answering system based on deep learning and knowledge graphs. Background technique [0002] At present, China is facing a series of health risks: the aging population, the huge number of patients with chronic diseases, the "younger" senile diseases, the normalization of sub-health, the rise of medically sensitive groups, and serious psychological problems. The medical field has always been an important direction of informatization development, and the scale of informatization investment in the medical industry has been increasing year by year. After a period of development, the online medical system has gradually changed from illusion to reality, which has become an inevitable trend. Through the online question-and-answer system, users can consult the medical system wit...

AI Technical Summary

Problems solved by technology

[0003] However, as far as the traditional medical question answering system is concerned, it usually faces the following important problems: (1) It is impossible to accurately locate the disease based on the shared symptoms; (2) The unprofessional description of the user causes ambiguity; (3) It does not have the active interaction function combined with the context

There are two main strategies for the construction of intelligent diagnosis and treatment question answering system: (1) data-based method, mainly through the search mechanism of the Internet, to search for similar information from the massive data on the Internet to assist in the diagnosis, and the amount of data collected corresponding to this strategy It is very large, and the number of disease categories that can be covered is also large; however, the knowledge density of network-based text data is low, and it is impossible to form an effective reasoning mechanism. 2) Knowledge-based reasoning mechanism, which mainly introduces fuzzy decision-making technology into disease classification. The advantage of this strategy is that it is easy to use existing deterministic disease knowledge, and can form understandable and displayable decision-making basis. Provide direct auxiliary support; but the disadvantage is that it relies too much on high-quality data, and the coverage of the total number of diseases and categories will be very limited

Images