CN113934858A - Health consultation realization method and device based on medical knowledge map retrieval technology - Google Patents

Health consultation realization method and device based on medical knowledge map retrieval technology Download PDF

Info

Publication number
CN113934858A
CN113934858A CN202111105853.8A CN202111105853A CN113934858A CN 113934858 A CN113934858 A CN 113934858A CN 202111105853 A CN202111105853 A CN 202111105853A CN 113934858 A CN113934858 A CN 113934858A
Authority
CN
China
Prior art keywords
entity
medical knowledge
library
health consultation
consultation
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202111105853.8A
Other languages
Chinese (zh)
Inventor
章瑶
庄国强
童良宇
詹进林
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ylz Information Technology Co ltd
Yilianzhong Zhiding Xiamen Technology Co ltd
Original Assignee
Ylz Information Technology Co ltd
Yilianzhong Zhiding Xiamen Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Ylz Information Technology Co ltd, Yilianzhong Zhiding Xiamen Technology Co ltd filed Critical Ylz Information Technology Co ltd
Priority to CN202111105853.8A priority Critical patent/CN113934858A/en
Publication of CN113934858A publication Critical patent/CN113934858A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
    • G06F16/367—Ontology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/3331—Query processing
    • G06F16/334—Query execution
    • G06F16/3344—Query execution using natural language analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
    • G06F16/374—Thesaurus
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H70/00—ICT specially adapted for the handling or processing of medical references
    • G16H70/20—ICT specially adapted for the handling or processing of medical references relating to practices or guidelines
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H70/00—ICT specially adapted for the handling or processing of medical references
    • G16H70/40—ICT specially adapted for the handling or processing of medical references relating to drugs, e.g. their side effects or intended usage
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H70/00—ICT specially adapted for the handling or processing of medical references
    • G16H70/60—ICT specially adapted for the handling or processing of medical references relating to pathologies

Abstract

The invention relates to the field of natural language processing, in particular to a method and a device for realizing health consultation based on a medical knowledge map retrieval technology, wherein the method comprises the following steps: establishing an entity list according to the medical knowledge graph, processing various entities of the entity list to realize multi-mode entity matching, and establishing a health consultation intention library and a pretreatment library according to entity relations; preprocessing the consultation problem, filtering out key entities and entity types, and combining a health consultation intention library with the entity types to obtain an entity relation chain; and splicing according to the entity relationship chain to generate a knowledge graph retrieval statement, retrieving an answer in the medical knowledge graph and displaying the answer. The problems are responded by the medical knowledge map, the health consultation intention library and the preprocessing library, so that the defects of low response speed, low accuracy, support of only one turn, large workload and long time consumption of deep learning labeling are overcome, multi-turn conversation is realized, answers are quickly retrieved from the medical knowledge map, and millisecond-level response is realized.

Description

Health consultation realization method and device based on medical knowledge map retrieval technology
Technical Field
The invention relates to the field of natural language processing, in particular to a method and a device for realizing health consultation based on a medical knowledge map retrieval technology.
Background
With the famous Turing test proposed by Turing in 1950, the intelligent question-answering system opened up the historical prelude, and a large number of scholars studied the intelligent question-answering system in various ways. The question-answering system on the market at present mainly adopts a method based on question similarity calculation; there are also a few question-answering systems that employ deep learning algorithms.
For example, CN202011064355.9 patent "an intelligent question and answer method and system based on long and short text matching", published as 2020.12.08, discloses an intelligent question and answer method and system based on long and short text matching, which can accurately locate a document paragraph where a question input by a user is located, and extract an answer. The method comprises the following steps: fusing a text similarity BM25 algorithm with a long and short text similarity calculation method based on a sendLA topic model, and matching user-input problems and corresponding paragraphs in a database from a word level and a sentence level respectively; extracting answers corresponding to the questions from the corresponding paragraphs based on the machine reading understanding model;
CN201710334888.6 patent "question-answering system and method based on deep learning", published as 20171107, which comprises a question-answering subsystem for receiving input questions and preprocessing the input questions; the system comprises a deep learning subsystem and a knowledge base subsystem, wherein the deep learning subsystem is used for extracting feature information in preprocessed input problems, generating corresponding first word vector information and obtaining a plurality of recommendation problems according to a problem classification model, a problem matching model and the first word vector information, the knowledge base subsystem is used for judging whether standard problems corresponding to unidentified problems exist in a knowledge base or not, marking the standard problems corresponding to the unidentified problems when the standard problems corresponding to the unidentified problems exist in the knowledge base, creating new standard problems according to the unidentified problems when the standard problems corresponding to the unidentified problems do not exist in the knowledge base, and marking the new standard problems.
However, the similarity algorithm usually adopts a question-and-answer form to carry out a dialogue, the response time is usually in direct proportion to the size of a knowledge base, while the health consultation relates to a plurality of contents such as diseases, examination items, medicines, foods and the like, the response speed corresponding to the similarity algorithm cannot meet the requirement, and the accuracy rate is low; the deep learning algorithm has high requirements on the material data, needs a large amount of training data for model training, has more medical knowledge content, has large labeling workload, and cannot support high-frequency knowledge updating in a model training process which consumes a long time.
Disclosure of Invention
In order to solve the defects of low response speed, low accuracy, only single-round support, large workload of deep learning and labeling and long time consumption when the health consultation uses a similarity algorithm in the prior art, the invention provides a health consultation realization method based on a medical knowledge map retrieval technology, which comprises the following steps:
establishing an entity list according to the medical knowledge graph, processing various entities of the entity list to realize multi-mode entity matching, and establishing a health consultation intention library and a pretreatment library according to entity relations;
preprocessing a consultation problem, filtering out key entities and entity types, and combining the health consultation intention library with the entity types to obtain an entity relation chain;
and splicing to generate a knowledge graph retrieval statement according to the entity relation chain, retrieving an answer in the medical knowledge graph and displaying the answer.
In one embodiment, the method for matching multi-mode entities by establishing an entity list based on a medical knowledge graph and processing various entities by using an AC automaton comprises the following steps:
initializing a Trie;
adding a medical entity Keyword into the Trie and constructing a success table according to a success function;
after all the medical entity keywords are added, checking and creating a failure table;
outputting the hit pattern string according to the failure table when the health consultation question is input.
In one embodiment, the health consultation intention library is established according to the entity relation condition of the medical knowledge map, wherein the entity relation condition comprises disease symptoms, causes, complications, drug indications and a disease diet recommendation preprocessing library.
In one embodiment, the preprocessing library comprises a similar word replacement library and a stop word library; the similar word replacing library is used for replacing similar words for the input health consultation problems so as to realize consultation problem pretreatment.
In one embodiment, the key entities and entity types are filtered out through an AC automaton, and when the entity names are not filtered out from the questions, the entity names corresponding to the entity types are obtained from the historical question records.
In one embodiment, the AC automaton filters out key entity names from the preprocessed consultation problems, and obtains entity names corresponding to entity types according to a preprocessed entity dictionary;
and storing the historical entity name in a card slot form, and replacing and updating when the card slot obtains the same type of entity name again.
In one embodiment, when the entity name is not captured in the question, the entity type is inferred by the health consulting the intent library and based on the intent, and the entity name corresponding to the entity type is obtained from the historical question record.
In one embodiment, the answers are obtained according to the consultation question retrieval, and are displayed in the question-answering system after being spliced.
The invention also provides a health consultation device based on the medical knowledge map retrieval technology, which is characterized in that: the system comprises a generation module, a query module and a query module, wherein the generation module is used for establishing an entity list according to a medical knowledge graph, processing various entities in the entity list to realize multi-mode entity matching, and establishing a health consultation intention library and a pretreatment library according to an entity relationship;
the processing module is used for filtering out key entities and entity types after preprocessing the consultation problem, and acquiring an entity relation chain by combining the entity types through the health consultation intention library;
and the retrieval display module is used for generating a knowledge graph retrieval statement according to the entity relation chain in a splicing manner, retrieving answers in the medical knowledge graph and displaying the answers.
The present invention also provides an apparatus comprising a processor and a memory for storing computer program instructions, which are executed by the processor to implement any one of the above-mentioned health consultation implementation methods based on medical knowledge-map retrieval technology.
Based on the above, compared with the prior art, the health consultation realization method and device based on the medical knowledge map retrieval technology, provided by the invention, have the advantages that the health problems of consultation are responded through the medical knowledge map, the health consultation intention library and the preprocessing library, and the defects of low response speed, low accuracy, large workload of annotation and long time consumption when only supporting single-round conversation and using deep learning in the prior art are overcome, so that multiple rounds of conversation are realized, answers are quickly retrieved from the medical knowledge map, and millisecond-level response is realized.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts; in the following description, the drawings are illustrated in a schematic view, and the drawings are not intended to limit the present invention.
FIG. 1 is a diagram of the steps of a health consultation implementation method based on medical knowledge map retrieval technology provided by the invention;
fig. 2 is a flow chart of the AC automaton provided by the present invention for processing various entities to implement multi-pattern entity matching.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some embodiments of the present invention, but not all embodiments; the technical features designed in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other; all other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In the description of the present invention, it is to be noted that all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs, and are not to be construed as limiting the present invention; it will be further understood that terms used herein should be interpreted as having a meaning that is consistent with their meaning in the context of this specification and the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
Specific examples are given below:
referring to fig. 1, a health consultation implementation method based on a medical knowledge map retrieval technology includes the following steps: establishing an entity list according to the medical knowledge graph, processing various entities of the entity list to realize multi-mode entity matching, and establishing a health consultation intention library and a pretreatment library according to entity relations;
preprocessing the consultation problem, filtering out key entities and entity types, and acquiring an entity relation chain by combining a health consultation intention library with the entity types;
and splicing to generate a knowledge graph retrieval statement according to the entity relation chain, retrieving an answer in the medical knowledge graph and displaying the answer.
Compared with the prior art, the health consultation implementation method based on the medical knowledge map retrieval technology solves the defects that in the prior art, when the health consultation uses a similarity algorithm, the response speed is low, the accuracy is low, only a single-round conversation is supported, and when deep learning is used, the workload is large and the time consumption is long by replying through the medical knowledge map, the health consultation intention library and the preprocessing library, so that multiple rounds of conversations are realized, answers are quickly retrieved from the medical knowledge map, and millisecond-level response is realized.
Specifically, an entity list is established according to the medical knowledge graph, various entities of the entity list are processed to realize multi-mode entity matching, and a health consultation intention library and a preprocessing library are established according to the entity relationship;
establishing an entity list based on the medical knowledge graph, wherein the entity list comprises 18 medically related entity lists including hospitals, departments, doctors, diseases, disease aliases, symptoms, diagnoses, treatment modes, prevention, examination items, medicines, medicine aliases, medicine manufacturers, indications, contraindications, cautions, foods and sports, and processing various entities by adopting an AC automatic machine to realize multi-mode entity matching.
Specifically, referring to fig. 2, the step of processing various entities by using an AC automaton to realize multi-mode entity matching specifically includes the following steps: initializing a Trie; adding a medical entity Keyword into the Trie and constructing a success table according to a success function; after all the medical entity keywords are added, checking and creating a failure table; and outputting a hit pattern string according to the failure table when the health consultation question is input, wherein the pattern string refers to the corresponding entity name.
Then, a health consultation intention library is established according to the entity relation condition of the medical knowledge map, wherein the entity relation condition comprises disease symptoms, causes, complications, drug indications and disease diet recommendation, for example, the disease symptom consultation generally inquires: what symptoms are the XX disease? What is the XX disease present? What is often the case with XX disease? It should be understood that XX herein refers to the pronoun of a disease.
Then, the preprocessing library comprises a similar word replacing library and a stop word library; the similar word replacing library is used for replacing similar words for the input health consultation problems so as to realize the pretreatment of the consultation problems; preferably, the similar words are stored in a dictionary form, such as: { "headache": headache "], headache": headache "[" headache "] } to accelerate the replacement rate. The stop word removing library is used for removing virtual words which have no practical significance per se, such as 'yes' and 'yes', reducing data redundancy and improving precision and accuracy of question processing.
Preferably, the data processing is completed in the model initialization process, and the information is stored in a memory variable or a shared database, so that the time-consuming problem caused by reloading and processing data in the interface calling process is avoided.
Specifically, after the consultation problem is preprocessed in the step, key entities and entity types are filtered, and the health consultation intention library is combined with the entity types to obtain an entity relation chain:
firstly, preprocessing the consultation problem, wherein the preprocessing comprises but is not limited to stop words and similar word replacement; then, the key entities and entity types are filtered out by the AC automaton.
Specifically, the AC automaton filters out key entity names from the preprocessed consultation problems, and obtains entity types of corresponding entities according to a preprocessed entity dictionary, preferably, when the entity names are not captured in the problems, the entity types are pushed back through a health consultation intention library according to the intention; and storing the historical entity name in a card slot mode to acquire the entity name of the corresponding entity type from the historical problem record, and replacing and updating to ensure accurate response when the card slot acquires the entity name of the same type again.
During implementation, the preprocessed user problem is put into an AC automaton, key entity names are filtered out, and the user problem is solved according to a preprocessed entity dictionary, such as { entity name: entity type, directly obtaining entity type of corresponding entity; the health consultation adopts a card slot form to store historical entity names, and sets max _ history values and storage _ time values of conversation strategies, if the max _ history values are set to be 8, namely only the entity names of the latest 8 conversations are stored, the storage _ time values are set to be 12, the name storage time is 12 hours, card slot information is automatically emptied after 12 hours, and when the card slot obtains the same type of entity names again, the card slot is replaced and updated.
In the embodiment, when a user inquires about 'what symptom is caused by cold' for the first time, the disease entity 'cold' is obtained through an AC automaton and stored in a disease card slot; if the user asks for 'what medicine can be taken by lung cancer' for the second time, the captured disease entity 'lung cancer' can replace the latest disease card slot; if the user asks for "what medicine can be taken for the second time? If the entity name is not captured in the question, the model can combine with the health consultation intention library, and according to the intention (what medicine to eat) the entity type is pushed down as the disease, the latest disease entity is obtained from the historical disease card slot, thereby accurately positioning the intention of the user as to what medicine to eat when the user catches a cold; if the user inquires for the third time about "what side effect is in amoxicillin? ", the newly acquired drug entity" amoxicillin "will be stored in the drug card slot drug".
The health consultation intention library is combined with the entity type to obtain an entity relation chain, and when the health consultation intention library is implemented, whether intention characteristic words are in the question or not is judged based on the health consultation intention library, the user intention is determined by combining the entity type, and the entity relation chain is obtained. If "what symptoms the lung cancer has" is "the lung cancer" is disease entity (disease), and the symptoms correspond to disease symptoms in the health counseling intention library (has _ symptom), the entity relationship chain of the medical knowledge map can be defined: lung cancer-has _ symptom- > symptom.
Specifically, in the step, a knowledge graph retrieval statement is generated according to the entity relation chain in a splicing manner, an answer is retrieved in the medical knowledge graph, and the answer is displayed:
according to the obtained entity relation chain, generating a knowledge graph retrieval statement by splicing, retrieving a question answer in a medical knowledge graph by a knowledge graph retrieval technology, taking the lung cancer Symptom relation chain as an example, converting the entity relation chain 'lung cancer-has _ Symptom- > Symptom' into a medical knowledge graph retrieval statement 'MATCH p ═ (a: Disease { name:' lung cancer '}) - [ r: has _ Symptom ] - > (b: Symptom) RETURN b.name', connecting a graph database, retrieving corresponding Symptom information, splicing results, retrieving according to a consultation question to obtain an answer, and displaying the answer in a question-answer system after splicing.
When the method is implemented, after a user inputs an advisory question in a question-answering system, the question-answering system acquires the advisory question and then preprocesses the advisory question through a preprocessing library, the preprocessed advisory question filters key entities and entity types through an AC automaton, wherein health advisory adopts a card slot form to store historical entity names, when the entity names are not captured in the question, the health advisory intention library is used for pushing the entity types backwards according to intentions, the entity names corresponding to the entity types are acquired from the historical card slot, so that an entity relation chain is acquired, the entity relation chain is spliced to generate a knowledge map retrieval statement, answers to the question are retrieved from a medical knowledge map, and the answers are spliced in the question-answering system to reply.
The invention also provides a health consultation device based on the medical knowledge map retrieval technology, which comprises a generation module, a pre-processing module and a health consultation meaning library, wherein the generation module is used for establishing an entity list according to the medical knowledge map, processing various entities in the entity list to realize multi-mode entity matching, and establishing the health consultation meaning library and the pre-processing library according to the entity relation; the processing module is used for filtering out key entities and entity types after preprocessing the consultation problem, and acquiring an entity relation chain by combining the entity types through the health consultation intention library; and the retrieval display module is used for generating a knowledge graph retrieval statement according to the entity relation chain in a splicing manner, retrieving answers in the medical knowledge graph and displaying the answers.
During implementation, the generation module establishes an entity list according to the medical knowledge graph, processes various entities of the entity list to realize multi-mode entity matching, establishes a health consultation intention library and a preprocessing library according to entity relations to realize generation of a consultation frame, preferably completes data processing in the model initialization process, stores information in a memory variable or a shared database, and avoids the problem of time consumption caused by reloading and processing data in the interface calling process.
The processing module is used for filtering out key entities and entity types after preprocessing the consultation problems, and acquiring an entity relation chain by combining the entity types through the health consultation intention library;
when the query question is implemented, after a user inputs a query question in a question answering system, the question answering system acquires the query question and then preprocesses the query question through a preprocessing library, the preprocessed query question filters out key entities and entity types through an AC (alternating current) automaton, when the entity name is not captured in the question, the health query idea library is used for pushing back the entity type according to intention, the health query adopts a clamping groove form to store historical entity names so as to acquire the entity name corresponding to the entity type from a historical question record when the entity name is not filtered out in the question, and then the health query idea library is used for acquiring an entity relationship chain by combining the entity types;
and the retrieval display module splices the entity relationship chains to generate a knowledge map retrieval statement, retrieves answers of the questions in the medical knowledge map, and splices the answers in the question-answering system for reply.
In summary, compared with the prior art, the health consultation implementation method and device based on the medical knowledge map retrieval technology, provided by the invention, have the advantages that the health questions consulted are responded through the medical knowledge map, the health consultation intention library and the preprocessing library, and the defects that in the prior art, when the health consultation is carried out by using a similarity algorithm, the response speed is low, the accuracy is low, only a single-round conversation is supported, and when deep learning is used, the workload is large and the time consumption is long are overcome, so that multiple rounds of conversations are realized, answers are rapidly retrieved from the medical knowledge map, and millisecond-level response is realized.
In addition, it will be appreciated by those skilled in the art that, although there may be many problems with the prior art, each embodiment or aspect of the present invention may be improved only in one or several respects, without necessarily simultaneously solving all the technical problems listed in the prior art or in the background. It will be understood by those skilled in the art that nothing in a claim should be taken as a limitation on that claim.
Although terms such as medical knowledge-graph, entity list, health consultation intent library, pre-processing library, entity relationship chain, knowledge-graph search statement … … are used more often herein, the possibility of using other terms is not excluded. These terms are used merely to more conveniently describe and explain the nature of the present invention; they are to be construed as being without limitation to any additional limitations that may be imposed by the spirit of the present invention; the terms "first," "second," and the like in the description and in the claims, and in the drawings, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.

Claims (10)

1. A health consultation realization method based on a medical knowledge map retrieval technology is characterized by comprising the following steps: the method comprises the following steps:
establishing an entity list according to the medical knowledge graph, processing various entities of the entity list to realize multi-mode entity matching, and establishing a health consultation intention library and a pretreatment library according to entity relations;
preprocessing a consultation problem, filtering out key entities and entity types, and combining the health consultation intention library with the entity types to obtain an entity relation chain;
and splicing to generate a knowledge graph retrieval statement according to the entity relation chain, retrieving an answer in the medical knowledge graph and displaying the answer.
2. The method for realizing health consultation based on medical knowledge-graph search technology according to claim 1, wherein: establishing an entity list based on the medical knowledge graph, and processing various entities by adopting an AC automatic machine to realize multi-mode entity matching, which specifically comprises the following steps:
initializing a Trie;
adding a medical entity Keyword into the Trie and constructing a success table according to a success function;
after all the medical entity keywords are added, checking and creating a failure table;
outputting the hit pattern string according to the failure table when the health consultation question is input.
3. The method for realizing health consultation based on medical knowledge-graph search technology according to claim 1, wherein: and establishing a health consultation intention library according to the entity relation condition of the medical knowledge map, wherein the entity relation condition comprises disease symptoms, causes, complications, drug indications and disease diet recommendation.
4. The method for realizing health consultation based on medical knowledge-graph search technology according to claim 1, wherein: the preprocessing library comprises a similar word replacing library and a stop word library; the similar word replacing library is used for replacing similar words for the input health consultation problems so as to realize consultation problem pretreatment.
5. The method for realizing health consultation based on medical knowledge-graph search technology according to claim 1, wherein: and filtering out key entities and entity types through an AC automaton, and acquiring entity names of corresponding entity types from the historical problem records when the entity names are not filtered out from the problems.
6. The method for realizing health consultation based on medical knowledge-graph search technology according to claim 5, wherein: filtering key entity names from the preprocessed consultation problems by the AC automaton, and acquiring entity names corresponding to entity types according to a preprocessed entity dictionary;
and storing the historical entity name in a card slot form, and replacing and updating when the card slot obtains the same type of entity name again.
7. The method for realizing health consultation based on medical knowledge-graph search technology according to claim 5, wherein: when the entity name is not captured in the question, the entity type is deduced backwards according to the intention by consulting the idea library through health, and the entity name corresponding to the entity type is obtained from the historical question record.
8. The method for realizing health consultation based on medical knowledge-graph search technology according to claim 1, wherein: and searching according to the consultation questions to obtain answers, splicing the answers, and displaying in a question-answering system.
9. Health consultation device based on medical knowledge map retrieval technology, which is characterized in that: the system comprises a generation module, a query module and a query module, wherein the generation module is used for establishing an entity list according to a medical knowledge graph, processing various entities in the entity list to realize multi-mode entity matching, and establishing a health consultation intention library and a pretreatment library according to an entity relationship;
the processing module is used for filtering out key entities and entity types after preprocessing the consultation problem, and acquiring an entity relation chain by combining the entity types through the health consultation intention library;
and the retrieval display module is used for generating a knowledge graph retrieval statement according to the entity relation chain in a splicing manner, retrieving answers in the medical knowledge graph and displaying the answers.
10. An apparatus of a processor and a memory, characterized in that: the memory is used for storing computer program instructions which are executed by the processor to complete the health consultation realization method based on the medical knowledge-map retrieval technology according to any one of claims 1 to 8.
CN202111105853.8A 2021-09-22 2021-09-22 Health consultation realization method and device based on medical knowledge map retrieval technology Pending CN113934858A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111105853.8A CN113934858A (en) 2021-09-22 2021-09-22 Health consultation realization method and device based on medical knowledge map retrieval technology

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111105853.8A CN113934858A (en) 2021-09-22 2021-09-22 Health consultation realization method and device based on medical knowledge map retrieval technology

Publications (1)

Publication Number Publication Date
CN113934858A true CN113934858A (en) 2022-01-14

Family

ID=79276175

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111105853.8A Pending CN113934858A (en) 2021-09-22 2021-09-22 Health consultation realization method and device based on medical knowledge map retrieval technology

Country Status (1)

Country Link
CN (1) CN113934858A (en)

Similar Documents

Publication Publication Date Title
CN110929016A (en) Intelligent question and answer method and device based on knowledge graph
US8108381B2 (en) System and method for analyzing electronic data records
CN112667794A (en) Intelligent question-answer matching method and system based on twin network BERT model
CA2852101A1 (en) Method and system for validation of claims against policy with contextualized semantic interoperability
US20200020423A1 (en) A method and system for matching subjects to clinical trials
Pang et al. BiobankConnect: software to rapidly connect data elements for pooled analysis across biobanks using ontological and lexical indexing
US11145390B2 (en) Methods and systems for recommending filters to apply to clinical trial search results using machine learning techniques
CN111899829A (en) Full-text retrieval matching engine based on ICD9/10 participle lexicon
CN112035757A (en) Medical waterfall flow pushing method, device, equipment and storage medium
Roberts et al. Annotating logical forms for EHR questions
Shi et al. Information extraction from FDA drug Labeling to enhance product-specific guidance assessment using natural language processing
CN109299238B (en) Data query method and device
Liao et al. Medical data inquiry using a question answering model
EP3901875A1 (en) Topic modelling of short medical inquiries
CN115620886B (en) Data auditing method and device
US11481557B2 (en) Clinical terminology mapping with natural language processing
Halland et al. Investigations into the use of SNOMED CT to enhance an OpenMRS health information system
AU2020265819A1 (en) System and method for phrase comparison consolidation and reconciliation
Chandrashekar et al. Ontology mapping framework with feature extraction and semantic embeddings
US20220293253A1 (en) Systems and methods using natural language processing to improve computer-assisted coding
Geissbuhler et al. Discussion of “Biomedical Infor-matics: We Are What We Publish”
EP3964980A1 (en) Automatically recommending an existing machine learning project as adaptable for use in a new machine learning project
CN113934858A (en) Health consultation realization method and device based on medical knowledge map retrieval technology
Montenegro et al. The hope model architecture: a novel approach to pregnancy information retrieval based on conversational agents
CN112561714B (en) Nuclear protection risk prediction method and device based on NLP technology and related equipment

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination