CN114155962A - Data cleaning method and method for constructing disease diagnosis by using knowledge graph - Google Patents
Data cleaning method and method for constructing disease diagnosis by using knowledge graph Download PDFInfo
- Publication number
- CN114155962A CN114155962A CN202210123200.0A CN202210123200A CN114155962A CN 114155962 A CN114155962 A CN 114155962A CN 202210123200 A CN202210123200 A CN 202210123200A CN 114155962 A CN114155962 A CN 114155962A
- Authority
- CN
- China
- Prior art keywords
- data
- user
- disease
- participant
- information given
- 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
Links
Images
Classifications
-
- 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
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- 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
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- 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
- G06F16/35—Clustering; Classification
-
- 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
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
- G06F40/295—Named entity recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Databases & Information Systems (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Public Health (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Pathology (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Animal Behavior & Ethology (AREA)
- Human Computer Interaction (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
A data cleaning method comprises the following steps: step 1, acquiring and storing information given by each participant of diseases related to a user through manual customer service; step 2, converting the information given by each participant of the disease related to the user into word vector data which can be identified by a computer and manually identifying the meaning of the information given by each participant of the disease related to the user; step 3, inputting the word vector data into a recurrent neural network to understand the meaning expressed by each participant; and 4, performing two-dimensional training on the word vector data and the artificial identification obtained in the step 2 so as to perform emotion classification on subsequently input information given by each participant of diseases related to the user. Also relates to a method for constructing disease diagnosis by using the knowledge map.
Description
Technical Field
The invention relates to the field of application of a knowledge graph in disease diagnosis, in particular to a data cleaning method and a disease diagnosis method constructed by using the knowledge graph and using the data cleaning method.
Background
In disease diagnosis, a doctor has a large amount of false information and emotional information in the communication process because a non-professional person communicates with the doctor, and the essence of the doctor is to analyze the cause of the disease according to symptoms, determine various optional treatment schemes based on the symptoms and the cause, and select the optimal treatment scheme from the various optional treatment schemes or the treatment scheme most related to the symptoms and the cause. With the development of artificial intelligence and the improvement of knowledge graph technology, network resources can be utilized to the maximum extent, the accumulated experience of each diagnosis and treatment can be utilized, and the future trend is fixed, but how to best remove the fake and truthful, and accurately grasp symptoms, etiology and treatment scheme, important information such as answers of patients, inquiry of experts, feedback of experts and the like needs to be held in each manual question and answer. This is also a critical need for patients and users who are concerned about health.
In addition, with irregular work and rest of most teenagers, the risk of diseases is greatly improved. The knowledge of medical and health nutrition management in this part of adolescents is relatively weak. The search for the etiology and treatment scheme after the disease is the most important ring of the whole system, the patient can input corresponding symptom phenomena to inquire the possible etiology associated with the symptom phenomena, and the well obtains the final etiology by inquiring the system to find the corresponding treatment scheme.
Disclosure of Invention
In view of the above problems and the disadvantages of the prior art, the present patent proposes a data cleaning method, which includes the following steps:
step 1, acquiring and storing information given by each participant of diseases related to a user through manual customer service;
step 2, converting the information given by each participant of the disease related to the user into word vector data which can be identified by a computer and manually identifying the meaning of the information given by each participant of the disease related to the user;
step 3, inputting the word vector data into a recurrent neural network to understand the meaning expressed by each participant;
and 4, performing two-dimensional training on the word vector data and the artificial identification obtained in the step 2 so as to perform emotion classification on subsequently input information given by each participant of diseases related to the user.
Preferably, the step 4 further includes a step 41 of performing two-dimensional training on the word vector data and the artificial identification obtained in the step 2, so as to perform keyword classification on subsequently input information given to each user-related disease participant.
Preferably, the step 2 further comprises a step 21 of performing keyword identification on the information given by each participant of the user-related disease and converting the keyword into a number.
Preferably, the information given to each participant in the user-related disease includes one or both of: expert inquiry information, user response information related to diseases given by the user and expert judgment results.
Preferably, the method further comprises a step 5 of constructing and storing the triple data of the keywords of the expert query information, the keywords of the user answer information and the keywords of the expert judgment result.
Preferably, the information given to each participant in the user-related disease includes one or all of: symptoms of the disease, etiology of the disease, and treatment regimen for the disease.
Preferably, the keyword classification is realized by a Bi-LSTM + CRF model.
The invention also provides a disease diagnosis method constructed by using the knowledge graph, and the data cleaning method is used.
Preferably, step 6 is included, the triple data are matched with the existing medical atlas entity data so as to update the background knowledge atlas data of the intelligent question answering module.
Preferably, step 7 is included to determine the final diagnosis and provide a treatment plan in the form of query features from the updated knowledge map data.
The patent scheme has the advantages that: aiming at the patients, the reasons of the symptoms can be checked as soon as possible, and an effective treatment scheme can be obtained at the first time to relieve the continuous aggravation of the illness state; for the management of diseases (such as urine disease), corresponding suggestions can be provided to adjust the life of the patient (such as' whether the diabetic patient has an influence on the offspring.); aiming at the aspect of doctors (hospitals), the pressure of the hospitals (doctors) can be reduced, and more resources can be provided for some very urgent patients to treat; aiming at the broad masses: the patent system can provide corresponding knowledge in the aspect of health preservation, help is provided for the healthy life of the masses, and diseases are avoided.
Description of the terms: jieba: realizing word segmentation of natural language and proposing a few words of tone with nonsense stop words) find my head somewhat hidden doing pain "→ [" i, found recently ", i have some hidden doing pain ].
Word2vec, natural language is converted into digital identification, and because the algorithm model does not accept natural language input and needs to be converted into digital, unlike one-hot (0001000), the Word2vec can retain natural language information after conversion. word2vec principle: word vector conversion preferentially carries out one-hot coding on words, Word2vec randomly initializes a W weight matrix, inner product operation is carried out on the W weight matrix and the converted one-hot coding, weighted average calculation is carried out on the result after the inner product operation to obtain theta, finally, the Word2vec adds a fully-connected network W, and the Word vector = theta W.
Bi-LSTM is a bidirectional cyclic neural network, and the training of word vectors can consider the situation of the current moment and the situation after the current moment, so that the Bi-LSTM has a remarkable effect on processing natural languages. Such as: "I feel I have a fan plug, should be because I fan inflammation and cause, not cold, because I do not feel other uncomfortable places" consider the whole talk to carry out the linguistic analysis through Bi-LSTM, reduce the error with great probability.
And (3) a CRF undirected graph model, which analyzes input data by using a hidden state. Principle of CRF: CRF was introduced to find the probability of labeling the most likely sequence under this one, as measured by Bi-LSTM, given the probability distribution of each wordAnd (4) transmitting the calculated hidden layer (N-dimensional vector) into a CRF model, and processing the local normalization problem existing in the Bi-LSTM by using the CRF model. The specific calculation formula is as follows
The LSTM is a recurrent neural network, and the model understands the meaning of the input sentence through training the word vector so as to judge the language is positive language, negative language and other languages. LSTM principle: the LSTM model will sort the fed vector data into numbers for each position adding weight W, where the LSTM will add the full link layer at the end to add b bias to the formula. The model self-adjusts the W weight and b bias to achieve the best fit by continuously feeding data into the model.
Match, similar to NOSQL language, the operation of adding, deleting, and modifying the graph database.
Drawings
FIG. 1 is a schematic flow chart of a data cleaning method according to the present invention;
FIG. 2 is a schematic flow chart of a method for constructing a disease diagnosis using a knowledge graph according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The diagnosis scheme based on the knowledge graph comprises a one-stype diagnosis and a combination scheme of intelligent question-answer inquiry and manual question-answer inquiry. The invention provides a data cleaning method, and the data cleaning method is shown in figure 1 and comprises the following steps:
step 1, denoted by reference numeral 1001, acquires and stores information given by each participant of a disease related to a user through manual customer service;
step 2, denoted by reference numeral 1002, converting information given for each participant of a user-related disease into computer-recognizable word vector data and manually identifying the meaning of the information given for each participant of the user-related disease;
step 3, denoted by reference numerals, inputting the word vector data into a recurrent neural network to understand the meaning expressed by each participant;
and 4, indicated by reference numerals, performing two-dimensional training on the word vector data and the artificial identification obtained in the step 2 so as to perform emotional classification on subsequently input information given by each participant of the diseases related to the user.
The method of the present invention for diagnosing a disease based on a knowledge base is fully shown in fig. 2, wherein a user first finds disease symptoms, also referred to as a user finding phenomenon 9001, and a primary model algorithm one-type is used for judging 9002, and a definite disease cause and treatment plan are given as far as the judgment is concerned. If not, then the intelligent question-answering model 9004 is entered, where there is a conventional knowledge map, under this model algorithm if a decision can be made for the disease, then the synonym library 9007 is entered, then the analysis tools jieba and word2vec processing 9008 are entered, then the sentiment classification 9009 is done and a confirmation 9010 is done.
When the intelligent question 9004 cannot judge the disease, the process proceeds to an artificial inquiry link 9005, in this link, as shown in fig. 2, the process may directly proceed to a confirmed diagnosis process 9006, and further proceed to a synonym library 9007, and then proceed to a process tool jieba and word2vec process 9008, and further proceed to an emotion classification link 9009, at this time, data cleaning is performed, and a knowledge map of the intelligent question is updated and stored 9012, so that a machine inquiry is initiated again in a more sophisticated knowledge map database 9013, and a confirmed diagnosis is performed according to the cause and scheme given by the new knowledge map database.
In a variation on the preferred embodiment of this patent, the data cleansing method includes the steps of: s3011, training an emotion classification model, S30111, converting data of a patient obtained by using an artificial customer service into word vectors and manually identifying meanings answered by a user (0 [ yes ], 1 [ no ], 2 [ other ]), wherein the word vectors converted by S30112 are computer-recognizable data, and feeding the word vectors into an LSTM recurrent neural network to enable the LSTM model to better understand the meaning of expression of the patient. (X: sentence word vector, Y: 0/1/2); s30113, training the model through the provided X and Y, so that the model can carry out emotional classification on subsequently input dialogs; s3012, a keyword classification model, S30121, converting the data of the expert inquiry obtained by the artificial customer service into a word vector and manually identifying the keyword of the expert inquiry (such as whether dizziness is felt or not, the keyword: dizziness); s30122, carrying out word vector processing on the whole natural language (key information (multiple)) Y (queried by the X expert) to convert the Chinese characters into numbers; s30123, feeding XY to a Bi-LSTM + CRF model at the same time, wherein the Bi-LSTM + CRF model scans the word vector input by X and performs recognition training by combining the extracted information data dimension given by Y; s302, after emotion analysis is carried out on the patient answer information, keyword extraction operation is carried out on the patient answer information through the process the same as S3012, and S3021, by combining the emotion classification model and the entity recognition model according to rules, the expert inquiry keywords, the patient answer keywords and the final expert judgment diagnosis result keywords are constructed, and triple data are stored.
A preferred embodiment of the method for constructing a disease diagnosis using a knowledge graph of the present invention is: one-stype diagnosis, namely partial simply and directly diagnosed diseases (such as the fact that a cold takes some medicines, whether the scratch surface needs to be coated with medicines little, etc.), intelligent question-answer inquiry, namely, the question which cannot be judged through one-stype enters the module to be continuously judged, and manual question-answer inquiry, wherein the last step that the former two machines cannot be judged is transferred to manual processing, comprises the following steps of S1 extracting disease-related information (which can be taken from most medical websites and the data of the question-answer inquiry is accumulated), S2 constructing a medical knowledge graph by using the relationship between the data utilization of disease basic information, treatment scheme information, complication information, etc., S201 disassembling some related attributes of the disease information, well-developed symptoms and treatment measures, establishing nodes and node attributes by adopting CREATE statements in Neo4j (PS: when the data is more, adopting Import commands to guide S202 to establish the relationship between CREATE statements, generating from disease name nodes backwards, constituting a complication of the disease; the disease is generated backwards from the disease name node and constitutes the disease cause: generating backward from disease cause nodes to form a disease treatment scheme; and generating backward from the disease name node to form a disease treatment scheme.
The S3 data washing comprises the following steps: s3011, training an emotion classification model, S30111, converting data of a patient obtained by using an artificial customer service into word vectors and manually identifying meanings answered by a user (0 [ yes ], 1 [ no ], 2 [ other ]), wherein the word vectors converted by S30112 are computer-recognizable data, and feeding the word vectors into an LSTM recurrent neural network to enable the LSTM model to better understand the meaning of expression of the patient. (X: sentence word vector, Y: 0/1/2); s30113, training the model through the provided X and Y, so that the model can carry out emotional classification on subsequently input dialogs; s3012, a keyword classification model, S30121, converting the data of the expert inquiry obtained by the artificial customer service into a word vector and manually identifying the keyword of the expert inquiry (such as whether dizziness is felt or not, the keyword: dizziness); s30122, carrying out word vector processing on the whole natural language (key information (multiple)) Y (queried by the X expert) to convert the Chinese characters into numbers; s30123, feeding XY to a Bi-LSTM + CRF model at the same time, wherein the Bi-LSTM + CRF model scans the word vector input by X and performs recognition training by combining the extracted information data dimension given by Y; s302, after emotion analysis is carried out on the patient answer information, keyword extraction operation is carried out on the patient answer information through the process the same as S3012, and S3021, by combining the emotion classification model and the entity recognition model according to rules, the expert inquiry keywords, the patient answer keywords and the final expert judgment diagnosis result keywords are constructed, and triple data are stored.
S4, the extracted expert inquiry key information is matched and associated with the existing medical atlas entity data S401, a symptom node is established by using a CREATE establishing statement, after the disease is spliced by combining the correlation between the patient answers and the patient answers, a treatment scheme corresponding to the symptom is established by using the CREATE establishing statement; s5, updating the background knowledge map database of the intelligent question answering module by using the sorted triple data; s501, as long as the user enters the manual module, indicates that the questions asked by the current user are not recorded in the database, and the questions are directly sorted and added into the database in the mode of S3. S6 determining the final diagnosis and providing a treatment plan by interrogating the features: s601 inquires of the user whether the problem exists by using the symptoms already stored in the database to determine the disease and provide a corresponding treatment scheme. The inquiry process of S602 is similar to the manual module, and the expert inquiry is changed into the machine inquiry, and finally the diagnosis scheme is confirmed. The S7 answers and reasoning include: s701, forward derivation, wherein S7011 adopts entity recognition trained in the S3 process to perform entity recognition on the input natural language, extracts entities needing to be queried and relational phenomena, and does not need to enter if the input is directly the required entities; s7012 asks the patient for symptoms and attributes corresponding to the disease using a Match query statement. S702, reverse reasoning: the disease determined by using the shape is used as a unique index Key value to start from a phenomenon, nodes related to the phenomenon are found from the phenomenon in a mode similar to a decision tree/binary tree algorithm (relationship line finding), backtracking is carried out to other related nodes, and a treatment scheme of well-developed diseases, whether the treatment scheme can treat other diseases or not can be inferred, and the like. S7021 explains S702: reasoning can be searched without limit through the relation line, and in the step S702, only a single example is used, and in one aspect, the reasoning can be terminated according to specific service requirements in specific situations.
The above embodiments are only preferred embodiments of the present invention, and any changes and modifications based on the technical solutions of the present invention in the technical field should not be excluded from the protection scope of the present invention.
Claims (10)
1. A data cleaning method is characterized by comprising the following steps:
step 1, acquiring and storing information given by each participant of diseases related to a user through manual customer service;
step 2, converting the information given by each participant of the disease related to the user into word vector data which can be identified by a computer and manually identifying the meaning of the information given by each participant of the disease related to the user;
step 3, inputting the word vector data into a recurrent neural network to understand the meaning expressed by each participant;
and 4, performing two-dimensional training on the word vector data and the artificial identification obtained in the step 2 so as to perform emotion classification on subsequently input information given by each participant of diseases related to the user.
2. The data cleansing method according to claim 1, wherein the step 4 further comprises a step 41 of performing two-dimensional training on the word vector data and the artificial identification obtained in the step 2 so as to perform keyword classification on subsequently input information given to each participant of the user-related disease.
3. The data cleansing method according to claim 2, wherein said step 2 further comprises a step 21 of performing keyword identification on information given to each participant of the user-related disease and converting the keyword into a number.
4. The data cleansing method according to claim 3, wherein the information given to each participant of the user-related disease includes one or all of: expert inquiry information, user response information related to diseases given by the user and expert judgment results.
5. The data cleaning method according to claim 4, further comprising a step 5 of constructing and storing triple data of keywords of the expert query information, keywords of the user response information and keywords of the expert judgment result.
6. The data cleansing method according to claim 5, wherein the information given to each participant of the user-related disease includes one or all of: symptoms of the disease, etiology of the disease, and treatment regimen for the disease.
7. The data cleaning method of claim 6, wherein the keyword classification is implemented by a Bi-LSTM + CRF model.
8. A method for constructing a disease diagnosis using a knowledge map, wherein the data cleansing method according to any one of claims 1 to 7 is used.
9. The method of claim 8, comprising step 6 of using the three sets of data to match existing medical atlas entity data to update the smart question-and-answer module background knowledge atlas data.
10. The method for constructing a disease diagnosis using a knowledge-graph according to claim 9, comprising a step 7 of determining a final diagnosis and providing a treatment plan in a query characteristic manner through the updated knowledge-graph data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210123200.0A CN114155962A (en) | 2022-02-10 | 2022-02-10 | Data cleaning method and method for constructing disease diagnosis by using knowledge graph |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210123200.0A CN114155962A (en) | 2022-02-10 | 2022-02-10 | Data cleaning method and method for constructing disease diagnosis by using knowledge graph |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114155962A true CN114155962A (en) | 2022-03-08 |
Family
ID=80450189
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210123200.0A Pending CN114155962A (en) | 2022-02-10 | 2022-02-10 | Data cleaning method and method for constructing disease diagnosis by using knowledge graph |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114155962A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114676258A (en) * | 2022-04-06 | 2022-06-28 | 北京航空航天大学 | Disease classification intelligent service method based on patient symptom description text |
| CN115563286A (en) * | 2022-11-10 | 2023-01-03 | 东北农业大学 | Knowledge-driven milk cow disease text classification method |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110377719A (en) * | 2019-07-25 | 2019-10-25 | 广东工业大学 | Medical answering method and device |
| CN110413798A (en) * | 2019-07-24 | 2019-11-05 | 厦门快商通科技股份有限公司 | A kind of medical and beauty treatment knowledge mapping method for auto constructing, system and storage medium |
| CN110415819A (en) * | 2019-08-05 | 2019-11-05 | 江苏康尚生物医疗科技有限公司 | The distributed medical big data analysis processing system of one kind and method |
| CN110911009A (en) * | 2019-11-14 | 2020-03-24 | 南京医科大学 | Clinical diagnosis aid decision-making system and medical knowledge map accumulation method |
| CN111339777A (en) * | 2020-02-24 | 2020-06-26 | 中国科学院自动化研究所 | Medical related intention identification method and system based on neural network |
| CN111667915A (en) * | 2020-06-05 | 2020-09-15 | 山东凯鑫宏业生物科技有限公司 | Intelligent medical system with disease reasoning and diagnosis method thereof |
| CN112507696A (en) * | 2021-02-04 | 2021-03-16 | 湖南大学 | Human-computer interaction diagnosis guiding method and system based on global attention intention recognition |
| CN112800244A (en) * | 2021-02-06 | 2021-05-14 | 成都中医药大学 | Method for constructing knowledge graph of traditional Chinese medicine and national medicine |
| CN113409907A (en) * | 2021-07-19 | 2021-09-17 | 广州方舟信息科技有限公司 | Intelligent pre-inquiry method and system based on Internet hospital |
| CN113889259A (en) * | 2021-09-06 | 2022-01-04 | 浙江工业大学 | Automatic diagnosis dialogue system under assistance of knowledge graph |
| CN113946649A (en) * | 2020-07-17 | 2022-01-18 | 阿里云计算有限公司 | Providing method of mediation plan, training method, related device and storage medium |
-
2022
- 2022-02-10 CN CN202210123200.0A patent/CN114155962A/en active Pending
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110413798A (en) * | 2019-07-24 | 2019-11-05 | 厦门快商通科技股份有限公司 | A kind of medical and beauty treatment knowledge mapping method for auto constructing, system and storage medium |
| CN110377719A (en) * | 2019-07-25 | 2019-10-25 | 广东工业大学 | Medical answering method and device |
| CN110415819A (en) * | 2019-08-05 | 2019-11-05 | 江苏康尚生物医疗科技有限公司 | The distributed medical big data analysis processing system of one kind and method |
| CN110911009A (en) * | 2019-11-14 | 2020-03-24 | 南京医科大学 | Clinical diagnosis aid decision-making system and medical knowledge map accumulation method |
| CN111339777A (en) * | 2020-02-24 | 2020-06-26 | 中国科学院自动化研究所 | Medical related intention identification method and system based on neural network |
| CN111667915A (en) * | 2020-06-05 | 2020-09-15 | 山东凯鑫宏业生物科技有限公司 | Intelligent medical system with disease reasoning and diagnosis method thereof |
| CN113946649A (en) * | 2020-07-17 | 2022-01-18 | 阿里云计算有限公司 | Providing method of mediation plan, training method, related device and storage medium |
| CN112507696A (en) * | 2021-02-04 | 2021-03-16 | 湖南大学 | Human-computer interaction diagnosis guiding method and system based on global attention intention recognition |
| CN112800244A (en) * | 2021-02-06 | 2021-05-14 | 成都中医药大学 | Method for constructing knowledge graph of traditional Chinese medicine and national medicine |
| CN113409907A (en) * | 2021-07-19 | 2021-09-17 | 广州方舟信息科技有限公司 | Intelligent pre-inquiry method and system based on Internet hospital |
| CN113889259A (en) * | 2021-09-06 | 2022-01-04 | 浙江工业大学 | Automatic diagnosis dialogue system under assistance of knowledge graph |
Non-Patent Citations (2)
| Title |
|---|
| 丁雅琴: "基于知识图谱的医疗问答系统研究与开发", 《中国优秀硕士学位论文全文数据库 医药卫生科技辑》 * |
| 李梦龙: "基于多数据源医疗知识图谱的知识对话模型研究", 《中国优秀硕士学位论文全文数据库 医药卫生科技辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114676258A (en) * | 2022-04-06 | 2022-06-28 | 北京航空航天大学 | Disease classification intelligent service method based on patient symptom description text |
| CN114676258B (en) * | 2022-04-06 | 2024-05-31 | 北京航空航天大学 | Disease classification method based on symptom description text and not aiming at diagnosis |
| CN115563286A (en) * | 2022-11-10 | 2023-01-03 | 东北农业大学 | Knowledge-driven milk cow disease text classification method |
| CN115563286B (en) * | 2022-11-10 | 2023-12-01 | 东北农业大学 | Knowledge-driven dairy cow disease text classification method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112035636B (en) | Question and answer management method, device and equipment of medical question and answer system and storage medium | |
| CN110377715A (en) | Reasoning type accurate intelligent answering method based on legal knowledge map | |
| WO2023029506A1 (en) | Illness state analysis method and apparatus, electronic device, and storage medium | |
| CN114155962A (en) | Data cleaning method and method for constructing disease diagnosis by using knowledge graph | |
| CN110675944A (en) | Triage method and device, computer equipment and medium | |
| CN110120001B (en) | Method and system for scoring based on combination of knowledge graph library and memory curve | |
| CN110175229B (en) | Method and system for on-line training based on natural language | |
| CN113724882B (en) | Method, device, equipment and medium for constructing user portrait based on inquiry session | |
| CN113223735B (en) | Diagnosis method, device, equipment and storage medium based on dialogue characterization | |
| CN113569023A (en) | Chinese medicine question-answering system and method based on knowledge graph | |
| CN109119160B (en) | Expert triage system with multiple reasoning modes and method thereof | |
| CN117056493A (en) | Large language model medical question-answering system based on medical record knowledge graph | |
| CN111339269A (en) | Knowledge graph question-answer training and application service system with automatically generated template | |
| CN117095827A (en) | Anesthesia large model training method and equipment | |
| CN112700865A (en) | Intelligent triage method based on comprehensive reasoning | |
| CN113764112A (en) | Online medical question and answer method | |
| CN116992002A (en) | Intelligent care scheme response method and system | |
| Xu | An integrated rule-and case-based approach to AIDS initial assessment | |
| CN116737911A (en) | Deep learning-based hypertension question-answering method and system | |
| CN111400453A (en) | Intelligent interaction system and method based on legal consultation | |
| CN117909466A (en) | Domain question-answering system, construction method, electronic device and storage medium | |
| CN116386805A (en) | Intelligent guided diagnosis report generation method | |
| CN113886555B (en) | Legal consultation service method and device | |
| CN117435714B (en) | Knowledge graph-based database and middleware problem intelligent diagnosis system | |
| CN117056536B (en) | Knowledge graph driving-based virtual doctor system and operation method thereof |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20220308 |