CN110377719B - Medical question and answer method and device - Google Patents
Medical question and answer method and device Download PDFInfo
- Publication number
- CN110377719B CN110377719B CN201910677024.3A CN201910677024A CN110377719B CN 110377719 B CN110377719 B CN 110377719B CN 201910677024 A CN201910677024 A CN 201910677024A CN 110377719 B CN110377719 B CN 110377719B
- Authority
- CN
- China
- Prior art keywords
- vector
- question
- medical
- information
- word
- 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.)
- Active
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
- G06F16/33—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
- G06F16/33—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
- 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
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Epidemiology (AREA)
- Public Health (AREA)
- Primary Health Care (AREA)
- Medical Informatics (AREA)
- Human Computer Interaction (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention provides a medical question-answering method, which comprises the following steps: acquiring query information, and performing information processing on the query information to obtain a word vector corresponding to the query information; acquiring a personal health file of a user, and associating the personal health file with each key word to obtain associated content; searching each question information corresponding to the associated content in the medical knowledge map, and obtaining a question vector corresponding to each question information; similarity calculation is carried out on the word vectors and each question vector, and a target question vector with the highest similarity to the word vectors is determined; and outputting the medical answer corresponding to the target question vector to the user. By applying the method provided by the invention, the query information is converted into a word vector form, each question information is converted into a question vector, the similarity of each question vector and the word vector is calculated, and the medical answer corresponding to the question vector with the highest similarity is obtained and fed back to the user, so that the user can be ensured to perform timely prevention and treatment according to the disease.
Description
Technical Field
The invention relates to the technical field of computers, in particular to a medical question answering method and device.
Background
With the rapid development of internet technology, since computer technology is applied to the medical field, people can make appointment registration or collect and record some medical information through the internet. When a user needs to know the relevant information of certain diseases, on-line medical questioning and answering can be carried out on some questioning and answering websites. The user only needs to input questions to be asked on the related website, and then on-line professional doctors answer the questions according to the questions asked by the user, and the user answers the questions and selects a proper prevention or treatment method for the questions asked.
However, as the number of users who ask questions online increases, the rate of increasing questions is much greater than the rate of answering, and when the number of online users is much greater than the number of professional doctors, the online doctors need to answer each question one by one, so that the waiting time of the users is too long to answer. Meanwhile, not all users know the related terms of all diseases, and the questions cannot be professionally described when being asked, so that different descriptions can be provided for the same questions, the same or similar questions can be answered by online doctors for many times, manpower and material resources are greatly wasted, and the timeliness of preventing or treating certain diseases by the users is also influenced.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a medical question and answer method, by which answers corresponding to query questions can be fed back to a user in time according to the query questions input by the user, so that the user can be ensured to perform prevention and treatment in time according to symptoms.
The invention also provides a medical question answering device which is used for ensuring the realization and the application of the method in practice.
A medical question-answering method, comprising:
acquiring query information input by a user, and performing information processing on the query information to obtain a vector group corresponding to the query information, wherein the vector group comprises an information category corresponding to the query information, each key word and an associated word corresponding to each key word;
calling a neural network model which is trained in advance to convert the vector groups to obtain a word vector corresponding to each vector group, wherein the word vector is semantic data corresponding to the vector groups;
acquiring a personal health file of the user, and associating the personal health file with each key word to obtain associated content;
searching each question information corresponding to the associated content in a preset medical knowledge graph, performing information processing on each question information to obtain a question vector group corresponding to each question information, inputting each question vector group into the neural network model, and obtaining a question vector corresponding to each question information;
similarity calculation is carried out on the word vectors and each question vector, and a target question vector with the highest similarity with the word vectors is determined;
acquiring target question information corresponding to the target question vector, taking a medical answer corresponding to the target question information as an answer of the query information, and displaying the answer of the query information on a preset display page so that the user can look up the answer of the query information.
Optionally, in the method, the performing information processing on the query information to obtain a vector group corresponding to the query information includes:
inputting the query information into a preset entity recognizer, and triggering the entity recognizer to output each key vocabulary corresponding to the query information;
determining the information category corresponding to each key word, and searching the associated word corresponding to each key word through a preset session description protocol SDP;
and generating a vector group corresponding to the query information according to the query information, each key word and the associated word corresponding to each key word.
The method described above, optionally, the training process of the neural network model includes:
acquiring a preset initial neural network model and a medical vocabulary data set, wherein the medical vocabulary data set comprises a plurality of medical vocabularies;
setting a standard word vector corresponding to each medical vocabulary, and executing a training process corresponding to the initial neural network model;
wherein the training process comprises: inputting each medical vocabulary in the medical vocabulary data set into the initial neural network model, converting each medical vocabulary through the initial neural network model, outputting a training word vector corresponding to each medical vocabulary, calculating an error value between the training word vector corresponding to each medical vocabulary and a standard word vector, and judging whether the error value between the training word vector corresponding to each medical vocabulary and the standard word vector is not less than a preset error threshold value or not;
when the error value between the training word vector and the standard word vector corresponding to each medical vocabulary is not less than a preset error threshold value, calling a preset loss function to adjust the model parameters of the initial neural network model;
and performing iterative training on the initial neural network model subjected to model parameter adjustment according to the training process, calling the loss function to adjust the model parameters of the initial neural network model after each training until the error value between the training word vector corresponding to each medical vocabulary currently output by the initial neural network model and the standard word vector is smaller than the error threshold value, finishing the training of the initial neural network model and obtaining the neural network model.
Optionally, the above method, where searching for each question information corresponding to the associated content in a preset medical knowledge graph includes:
determining each key word as a main node corresponding to the associated content;
searching each branch node corresponding to the main node based on the main node in the medical knowledge graph;
and acquiring user state information corresponding to the user according to the personal health record in the associated content, and searching each question information matched with the user state information in each branch node.
Optionally, in the method, after the similarity calculation is performed on the word vector and each question vector, the method further includes:
judging whether the similarity of each question vector and the word vector is greater than a preset similarity threshold value or not;
when the similarity between at least one target question vector and the word vector is greater than the similarity threshold, determining target question information corresponding to each target question vector, and acquiring medical answers corresponding to each target question information;
and storing each medical answer into a preset answer list from high to low according to the similarity of each corresponding target question vector of each medical answer, and displaying the answer list on the display page so that the user can look up each medical answer according to the answer list.
A medical question-answering device comprising:
the processing unit is used for acquiring query information input by a user, performing information processing on the query information and acquiring a vector group corresponding to the query information, wherein the vector group comprises an information category corresponding to the query information, each key word and an associated word corresponding to each key word;
the conversion unit is used for calling a neural network model which is trained in advance to convert the vector groups to obtain a word vector corresponding to each vector group, and the word vectors are semantic data corresponding to the vector groups;
the first acquisition unit is used for acquiring the personal health file of the user and associating the personal health file with each key word to acquire associated content;
the searching unit is used for searching each question information corresponding to the associated content in a preset medical knowledge map, performing information processing on each question information to obtain a question vector group corresponding to each question information, inputting each question vector group into the neural network model, and obtaining a question vector corresponding to each question information;
the calculation unit is used for calculating the similarity between the word vector and each question vector and determining a target question vector with the highest similarity to the word vector;
the first display unit is used for acquiring target question information corresponding to the target question vector, taking a medical answer corresponding to the target question information as an answer of the query information, and displaying the answer of the query information on a preset display page so that the user can look up the answer of the query information.
The above apparatus, optionally, the processing unit includes:
the input subunit is used for inputting the query information into a preset entity recognizer and triggering the entity recognizer to output each key vocabulary corresponding to the query information;
the first searching subunit is used for determining the information category corresponding to each key word and searching the associated word corresponding to each key word through a preset session description protocol SDP;
and the generating subunit is used for generating a vector group corresponding to the query information according to the query information, each key word and the associated word corresponding to each key word.
The above apparatus, optionally, further comprises:
the second acquisition unit is used for acquiring a preset initial neural network model and a medical vocabulary data set, and the medical vocabulary data set comprises a plurality of medical vocabularies;
the setting unit is used for setting a standard word vector corresponding to each medical vocabulary and executing a training process corresponding to the initial neural network model; wherein the training process comprises: inputting each medical vocabulary in the medical vocabulary data set into the initial neural network model, converting each medical vocabulary through the initial neural network model, outputting a training word vector corresponding to each medical vocabulary, calculating an error value between the training word vector corresponding to each medical vocabulary and a standard word vector, and judging whether the error value between the training word vector corresponding to each medical vocabulary and the standard word vector is not less than a preset error threshold value or not;
the adjusting unit is used for calling a preset loss function to adjust the model parameters of the initial neural network model when the error value between the training word vector and the standard word vector corresponding to each medical vocabulary is not less than a preset error threshold value;
and the training unit is used for carrying out iterative training on the initial neural network model subjected to model parameter adjustment according to the training process, calling the loss function to adjust the model parameters of the initial neural network model after each training until the error value between the training word vector corresponding to each medical vocabulary currently output by the initial neural network model and the standard word vector is smaller than the error threshold value, finishing the training of the initial neural network model and obtaining the neural network model.
The above apparatus, optionally, the search unit includes:
the determining subunit is used for determining each key word as a main node corresponding to the associated content;
the second searching subunit is used for searching each branch node corresponding to the main node based on the main node in the medical knowledge graph;
and the third searching subunit is used for acquiring user state information corresponding to the user according to the personal health record in the associated content, and searching each question information matched with the user state information in each branch node.
The above apparatus, optionally, further comprises:
the judging unit is used for judging whether the similarity of each question vector and the word vector is greater than a preset similarity threshold value or not;
the third obtaining unit is used for determining target question information corresponding to each target question vector and obtaining a medical answer corresponding to each target question information when the similarity between at least one target question vector and the word vector is larger than the similarity threshold;
and the second display unit is used for storing each medical answer into a preset answer list from high to low according to the similarity of each corresponding target question vector of each medical answer, and displaying the answer list on the display page so that the user can look up each medical answer according to the answer list.
A storage medium comprising stored instructions, wherein the instructions, when executed, control a device on which the storage medium is located to perform the medical question answering method described above.
An electronic device comprising a memory, and one or more instructions, wherein the one or more instructions are stored in the memory and configured to be executed by one or more processors to perform the medical question answering method.
Compared with the prior art, the invention has the following advantages:
the invention provides a medical question-answering method, which comprises the following steps: acquiring query information input by a user, and performing information processing on the query information to obtain a vector group corresponding to the query information; calling a neural network model which is trained in advance to convert the vector groups to obtain a word vector corresponding to each vector group, wherein the word vector is semantic data corresponding to the vector group; acquiring a personal health file of a user, and associating the personal health file with each key word to obtain associated content; searching each question information corresponding to the associated content in a preset medical knowledge map, performing information processing on each question information to obtain a question vector group corresponding to each question information, inputting each question vector group into a neural network model, and obtaining a question vector corresponding to each question information; similarity calculation is carried out on the word vectors and each question vector, and a target question vector with the highest similarity to the word vectors is determined; acquiring target question information corresponding to the target question vector, taking medical answers corresponding to the target question information as answers of the query information, and displaying the answers of the query information on a preset display page so that a user can look up the answers of the query information. By applying the method provided by the invention, the query information is converted into a word vector form, each question information is obtained through the knowledge map, the question vector corresponding to each question information is determined, the similarity calculation is carried out on each question vector and the word vector, the medical answer corresponding to the question vector with the highest similarity is obtained and fed back to the user in time, and the doctor on the waiting line does not need to answer, so that the user can be ensured to carry out prevention and treatment in time according to the symptoms.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
Fig. 1 is a flowchart of a method for medical question answering according to an embodiment of the present invention;
FIG. 2 is a flowchart of another method for medical question answering according to an embodiment of the present invention;
FIG. 3 is a flowchart of another method for medical question answering according to an embodiment of the present invention;
FIG. 4 is a diagram illustrating an apparatus structure of a medical question answering apparatus according to an embodiment of the present invention;
fig. 5 is a device structure diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
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 only a part of the embodiments of the present invention, and not all of the embodiments. 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 this application, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions, and the terms "comprises", "comprising", or any other variation thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The invention is operational with numerous general purpose or special purpose computing device environments or configurations. For example: personal computers, server computers, hand-held or portable devices, tablet-type devices, multi-processor apparatus, distributed computing environments that include any of the above devices or equipment, and the like.
The embodiment of the invention provides a medical question-answering method, which can be applied to various system platforms, wherein an execution main body of the method can be a computer terminal or a processor of various mobile devices, and a flow chart of the method is shown in figure 1 and specifically comprises the following steps:
s101: acquiring query information input by a user, and performing information processing on the query information to obtain a vector group corresponding to the query information, wherein the vector group comprises an information category corresponding to the query information, each key word and an associated word corresponding to each key word;
in the embodiment of the invention, when a user inputs a medical problem needing to be inquired on a preset display interface, the inquiry information input by the user is acquired. Wherein, the query information is a query question input by a user. And after the query information is subjected to information processing, obtaining a vector group corresponding to the query information. The vector group comprises information categories corresponding to the query information, key words and associated words corresponding to the key words.
It should be noted that the specific form of the vector group may be < category, resource, entity >. Wherein, the category is an information category corresponding to the query information; the resource is a key vocabulary obtained after mapping with each key vocabulary; the entity is each key word.
S102: calling a neural network model which is trained in advance to convert the vector groups to obtain a word vector corresponding to each vector group, wherein the word vector is semantic data corresponding to the vector groups;
in the embodiment of the invention, the vector group is converted through a pre-trained neural network model, and the word vectors corresponding to the vector group are obtained. The vector group is in a word form, and after the vector group is converted by the neural network model, the word form is converted into a word vector in a character data form.
It should be noted that the neural network model may specifically be a deep neural network model, a convolutional neural network model, a bidirectional cyclic neural network model, or the like.
S103: acquiring a personal health file of the user, and associating the personal health file with each key word to obtain associated content;
in the embodiment of the present invention, a personal health profile of a user is obtained, where the personal health profile may specifically include the age, sex, physical condition, historical medication condition, historical treatment condition, and the like of the user. And associating each key word with the personal health file to obtain associated content. For example, the key vocabulary is the cold; when the user age is 50, the gender is male, and the user is allergic to the A-drug in the personal health profile, the related contents may be: 50 year old male has common cold and is allergic to the A medicine.
S104: searching each question information corresponding to the associated content in a preset medical knowledge graph, performing information processing on each question information to obtain a question vector group corresponding to each question information, inputting each question vector group into the neural network model, and obtaining a question vector corresponding to each question information;
in the implementation of the invention, according to the associated content, the question information corresponding to the associated content is searched in the medical knowledge map. For example, if the associated content is a 50 year old male cold and allergic to drug a, the questioning information may be: the answering method of the cold, how to take the A-drug allergic cold and other questioning information. And after the question information is subjected to information processing, obtaining a question vector group. Wherein, the information processing mode is consistent with the information processing procedure of the step S101. And inputting each question vector into the neural network model to obtain the question vector. The process of obtaining the question vector is the same as the process of obtaining each word vector in step S102.
It should be noted that the knowledge map stores the relevant information and solutions of various diseases.
S105: similarity calculation is carried out on the word vectors and the question vectors, and a target question vector with the highest similarity with the word vectors is determined;
in the embodiment of the invention, the word vector and each question vector are subjected to similarity calculation, and the target question vector with the highest calculation similarity result is determined.
S106: acquiring target question information corresponding to the target question vector, taking a medical answer corresponding to the target question information as an answer of the query information, and displaying the answer of the query information on a preset display page so that the user can look up the answer of the query information.
In the embodiment of the invention, the medical answer of the question information corresponding to the target question vector with the highest similarity is taken as the answer of the query information to be output. That is, the medical answer is displayed on a preset display page, and the user can look up the medical answer.
In the method provided by the embodiment of the invention, query information input by a user is firstly subjected to information processing to obtain a vector group corresponding to the query information, and the vector group is trained through a neural network model to obtain a word vector. And associating according to the personal health record of the user and each keyword in the vector group to obtain associated content. And searching each question information corresponding to the associated content in a preset knowledge graph, and performing information processing on each question information to obtain a question vector group. Then the question vector group is input into the neural network model and converted into question vectors. And carrying out similarity calculation on each question vector and the word vector, and taking the question vector with the highest similarity obtained by calculation as a target question vector. And feeding back the medical answer corresponding to the question information of the target question vector to the user as output. I.e. the medical answer is presented on a display page.
By applying the method provided by the embodiment of the invention, the key content required to be inquired by the user can be determined after the inquiry information input by the user is processed. After the query information is converted into a word vector form, similarity calculation is carried out on the query information and each question vector, medical answers corresponding to the question vectors with the highest similarity are obtained and fed back to the user in time, and the on-line doctor does not need to wait for answering, so that the user can be guaranteed to carry out prevention and treatment in time according to symptoms.
In the method provided in the embodiment of the present invention, the process of performing information processing on the query information to obtain a vector group corresponding to the query information is shown in fig. 2, and specifically includes:
s201: inputting the query information into a preset entity recognizer, and triggering the entity recognizer to output each key vocabulary corresponding to the query information;
in the embodiment of the invention, the entity recognizer filters some irrelevant words of the query information input by the user, and only stores the key words corresponding to the query information.
S202: determining the information category corresponding to each key word, and searching the associated word corresponding to each key word through a preset session description protocol SDP;
in the embodiment of the invention, the category of each associated vocabulary is determined, and the associated vocabulary corresponding to the key vocabulary is searched in the SDP. In order to prevent other calling of a certain medical professional term, the SDP is used for searching the associated vocabulary corresponding to the key vocabulary, and all the possible query information input by the user is searched.
S203: and generating a vector group corresponding to the query information according to the query information, each key word and the associated word corresponding to each key word.
In the embodiment of the invention, the query information, each key word and the associated word form a vector group.
In the medical question answering method provided by the embodiment of the invention, the query information input by the user is processed by the preset entity identifier. After the query information is input into the entity recognizer, the entity recognizer outputs the key vocabulary of the query information. And determining the information category according to each key vocabulary, wherein the key vocabulary is high fever, and the information category is fever. And searching the associated vocabulary corresponding to the key vocabulary according to the SDP. For example, the associated words of fever include low fever, high fever, cold, common cold, etc. And generating a vector group corresponding to the query information by the query information, each key word and the associated word corresponding to each key word.
By applying the method provided by the embodiment of the invention, the query information is processed by the entity identifier and the SDP. When a user inputs some complex contents, key words can be extracted from the contents for information processing, and the accuracy of answer searching is ensured.
In the method provided by the embodiment of the present invention, the training process of the neural network model includes:
acquiring a preset initial neural network model and a medical vocabulary data set, wherein the medical vocabulary data set comprises a plurality of medical vocabularies;
setting a standard word vector corresponding to each medical vocabulary, and executing a training process corresponding to the initial neural network model;
wherein the training process comprises: inputting each medical vocabulary in the medical vocabulary data set into the initial neural network model, converting each medical vocabulary through the initial neural network model, outputting a training word vector corresponding to each medical vocabulary, calculating an error value between the training word vector corresponding to each medical vocabulary and a standard word vector, and judging whether the error value between the training word vector corresponding to each medical vocabulary and the standard word vector is not less than a preset error threshold value or not;
when the error value between the training word vector and the standard word vector corresponding to each medical vocabulary is not less than a preset error threshold value, calling a preset loss function to adjust the model parameters of the initial neural network model;
and performing iterative training on the initial neural network model subjected to model parameter adjustment according to the training process, calling the loss function to adjust the model parameters of the initial neural network model after each training until the error value between the training word vector corresponding to each medical vocabulary currently output by the initial neural network model and the standard word vector is smaller than the error threshold value, finishing the training of the initial neural network model and obtaining the neural network model.
In the method provided by the embodiment of the invention, the initial neural network model is trained by acquiring the initial neural network model and the medical vocabulary data set. The initial neural network model cannot accurately convert each medical vocabulary and output standard word vectors at the beginning, so that the training process needs to be continuously executed, and the parameters of the initial neural network model are adjusted after each training. The initial neural network model is subjected to parameter adjustment through a loss function, so that the capability of the initial neural network model for converting each medical vocabulary is improved. The process of training the initial neural network specifically comprises the following steps: inputting each medical vocabulary in the medical vocabulary data set into the initial neural network model, converting each medical vocabulary through the initial neural network model, outputting a training word vector corresponding to each medical vocabulary, calculating an error value between the training word vector corresponding to each medical vocabulary and a standard word vector, and judging whether the error value between the training word vector corresponding to each medical vocabulary and the standard word vector is not less than a preset error threshold value. When the error value between the training word vector corresponding to each medical vocabulary and the standard word vector is smaller than a preset error threshold value, the performance of the initial neural network model is proved to be the maximum, and therefore the training word vector output by the initial neural network model is considered to be basically consistent with the standard word vector. At this time, the training of the initial neural network model is completed, and the neural network model is obtained.
It should be noted that each medical vocabulary in the medical vocabulary data set can be obtained from each data or medical knowledge base of the network through the web crawler.
By applying the method provided by the embodiment of the invention, the word vector obtained by the neural network model when the vector group is converted can be more accurate through training the extension network model, and each obtained similarity is more accurate in subsequent similarity calculation, so that accurate answers can be searched for a user.
In the method provided by the embodiment of the present invention, the process of searching each question information corresponding to the associated content in the preset medical knowledge graph is shown in fig. 3, and specifically includes:
s301: determining each key word as a main node corresponding to the associated content;
in the embodiment of the invention, each key word is taken as a main node. The master node is the basis for searching each question message.
S302: searching each branch node corresponding to the main node based on the main node in the medical knowledge graph;
in the embodiment of the invention, the branch node corresponding to the main node is searched in the medical map, that is, all relevant questioning information related to each keyword is searched in the medical knowledge map.
It should be noted that the medical instruction map stores question and answer questions for such medical treatment, and the medical knowledge map is composed of various questions and answers, in which questions corresponding to each symptom are acquired from a plurality of medical databases. The knowledge map has words with various possible symptoms and various related information to form questions, and each question has corresponding answer content.
S303: and acquiring user state information corresponding to the user according to the personal health record in the associated content, and searching each question information matched with the user state information in each branch node.
In the embodiment of the invention, after each branch node is found, the question information matched with the user state information is searched in each corresponding branch node in the medical knowledge graph by combining the user state information in the personal health record of the user.
In the method provided by the embodiment of the invention, the key words are used as main nodes for searching the question information, and corresponding branch nodes, namely the content related to the key words, are searched based on the key words in the medical knowledge graph. And further performing fine search by combining with the personal health profile of the user. And searching each question information matched with the user state information in each branch node according to the user state information of the personal profile of the user. The user status information may specifically include: whether the user is allergic to the medication, medication records, treatment records of the user, etc.
By applying the method provided by the embodiment of the invention, the questioning information is searched through the key words and the personal health records, the medical answer matched with the user can be searched more finely, and the user can be ensured to prevent and treat the disease symptoms in time.
In the method provided in the embodiment of the present invention, after calculating the similarity between the word vector and each question vector, the method may further include:
judging whether the similarity of each question vector and the word vector is greater than a preset similarity threshold value or not;
when the similarity between at least one target question vector and the word vector is greater than the similarity threshold, determining target question information corresponding to each target question vector, and acquiring medical answers corresponding to each target question information;
and storing each medical answer into a preset answer list from high to low according to the similarity of each corresponding target question vector of each medical answer, and displaying the answer list on the display page so that the user can look up each medical answer according to the answer list.
In the method provided by the embodiment of the invention, after the similarity calculation is carried out on the word vector and each question vector, the similarity between the word vector and each question vector is obtained. And judging whether the similarity of each question vector and the word vector is greater than a preset similarity threshold. For example, if the similarity threshold is 80%, each question vector having a similarity higher than 80% is determined as a target question vector, and question information corresponding to each target question vector is determined. And sequencing the medical answers corresponding to each piece of questioning information according to the similarity from high to low. And storing the medical answers corresponding to the sorted target question vectors in an answer list. And outputting the answer list to the user, namely displaying the answer list on a preset display page, wherein the user can select medical answers which are more consistent with the user by looking up all the answers in the answer list.
By applying the method provided by the embodiment of the invention, a plurality of medical answers are displayed to the user in a mode of generating the answer list, so that the user can select more standard answers or more comprehensively know all solutions corresponding to the query information, and the user can be ensured to perform timely prevention and treatment on symptoms.
Based on the medical question-answering method provided by the embodiment, a specific embodiment corresponding to the method is provided, and the process is as follows:
when the user inputs "how to take a fever 38? "when inquiring information, the information processing is performed on the inquired information to obtain the vector group < fever type, fever, 38 degrees, medication, high fever, low fever, cold > corresponding to the inquired information. Wherein, the fever is classified as information; the key words are fever, 38 degrees and medication; high fever, low fever and cold are related words;
inputting the vector group into a neural network model to obtain a word vector corresponding to the vector group;
acquiring a personal health profile of the user; wherein the personal health record comprises the age of the user is 50 years old, the gender is male, the physical condition is good, the historical administration condition is absent, the historical treatment condition is absent, and the drug allergy condition is allergy to the A drug;
the health file is associated with an individual according to the key words, and the associated content '50-year-old male fever 38 degrees in medication mode and allergy to A medicine' is obtained;
searching question information corresponding to the related content in the medical knowledge map, for example, question information such as 'medication mode corresponding to A-drug allergy when fever occurs', 'how to take medication at high fever 38 degrees when male is allergic to A-drug', 'how to deal with fever and cold of middle-aged people' and the like;
similarly, each question information is processed to obtain a question vector group, such as < fever type, fever, A-drug allergy, medication, high fever, low fever, cold >, < fever type, A-drug allergy, high fever, 38 degrees, medication, fever, cold >, < fever type, fever, cold, low fever, high fever, inflammation > and the like;
training each question vector group through a neural network model to obtain a question vector corresponding to each question vector group;
similarity calculation is carried out on each question vector and the word vectors, the question vector with the highest similarity is obtained, and question information corresponding to the question vector is determined;
and feeding back the medical answer corresponding to the questioning information to the user as the answer of the query information.
The specific implementation procedures and derivatives thereof of the above embodiments are within the scope of the present invention.
Corresponding to the method described in fig. 1, an embodiment of the present invention further provides a medical question-answering device, which is used for implementing the method in fig. 1 specifically, the medical question-answering device provided in the embodiment of the present invention may be applied to a computer terminal or various mobile devices, and a schematic structural diagram of the medical question-answering device is shown in fig. 4, and specifically includes:
a processing unit 401, configured to obtain query information input by a user, perform information processing on the query information, and obtain a vector group corresponding to the query information, where the vector group includes an information category corresponding to the query information, each key word, and an associated word corresponding to each key word;
a conversion unit 402, configured to invoke a neural network model that is trained in advance to convert the vector groups, so as to obtain a word vector corresponding to each vector group, where the word vector is semantic data corresponding to the vector group;
a first obtaining unit 403, configured to obtain a personal health profile of the user, and associate the personal health profile with each of the key words to obtain associated content;
a searching unit 404, configured to search each piece of question information corresponding to the associated content in a preset medical knowledge graph, perform information processing on each piece of question information, obtain a question vector group corresponding to each piece of question information, input each question vector group into the neural network model, and obtain a question vector corresponding to each piece of question information;
a calculating unit 405, configured to perform similarity calculation on the word vector and each question vector, and determine a target question vector with the highest similarity to the word vector;
the first display unit 406 is configured to obtain target question information corresponding to the target question vector, use a medical answer corresponding to the target question information as an answer to the query information, and display the answer to the query information on a preset display page, so that the user refers to the answer to the query information.
In the apparatus provided by the present invention, the processing unit includes:
the input subunit is used for inputting the query information into a preset entity recognizer and triggering the entity recognizer to output each key vocabulary corresponding to the query information;
the first searching subunit is used for determining the information category corresponding to each key word and searching the associated word corresponding to each key word through a preset session description protocol SDP;
and the generating subunit is used for generating a vector group corresponding to the query information according to the query information, each key word and the associated word corresponding to each key word.
The device provided by the implementation of the invention further comprises:
the second acquisition unit is used for acquiring a preset initial neural network model and a medical vocabulary data set, and the medical vocabulary data set comprises a plurality of medical vocabularies;
the setting unit is used for setting a standard word vector corresponding to each medical vocabulary and executing a training process corresponding to the initial neural network model; wherein the training process comprises: inputting each medical vocabulary in the medical vocabulary data set into the initial neural network model, converting each medical vocabulary through the initial neural network model, outputting a training word vector corresponding to each medical vocabulary, calculating an error value between the training word vector corresponding to each medical vocabulary and a standard word vector, and judging whether the error value between the training word vector corresponding to each medical vocabulary and the standard word vector is not less than a preset error threshold value or not;
the adjusting unit is used for calling a preset loss function to adjust the model parameters of the initial neural network model when the error value between the training word vector and the standard word vector corresponding to each medical vocabulary is not less than a preset error threshold value;
and the training unit is used for carrying out iterative training on the initial neural network model subjected to model parameter adjustment according to the training process, calling the loss function to adjust the model parameters of the initial neural network model after each training until the error value between the training word vector corresponding to each medical vocabulary currently output by the initial neural network model and the standard word vector is smaller than the error threshold value, finishing the training of the initial neural network model and obtaining the neural network model.
In the apparatus provided by the present invention, the search unit includes:
the determining subunit is used for determining each key word as a main node corresponding to the associated content;
the second searching subunit is used for searching each branch node corresponding to the main node based on the main node in the medical knowledge graph;
and the third searching subunit is used for acquiring user state information corresponding to the user according to the personal health record in the associated content, and searching each question information matched with the user state information in each branch node.
The device provided by the implementation of the invention further comprises:
the judging unit is used for judging whether the similarity of each question vector and the word vector is greater than a preset similarity threshold value or not;
the third obtaining unit is used for determining target question information corresponding to each target question vector and obtaining a medical answer corresponding to each target question information when the similarity between at least one target question vector and the word vector is larger than the similarity threshold;
and the second display unit is used for storing each medical answer into a preset answer list from high to low according to the similarity of each corresponding target question vector of each medical answer, and displaying the answer list on the display page so that the user can look up each medical answer according to the answer list.
The specific working processes of the processing unit 401, the converting unit 402, the first obtaining unit 403, the searching unit 404, the calculating unit 405, and the first displaying unit 406 in the medical question-answering device disclosed in the above embodiment of the present invention may refer to the corresponding contents in the medical question-answering method disclosed in the above embodiment of the present invention, and are not described herein again.
The embodiment of the invention also provides a storage medium, which comprises a stored instruction, wherein when the instruction runs, the device where the storage medium is located is controlled to execute the medical question-answering method.
An electronic device is provided in an embodiment of the present invention, and the structural diagram of the electronic device is shown in fig. 5, which specifically includes a memory 501 and one or more instructions 502, where the one or more instructions 502 are stored in the memory 501, and are configured to be executed by one or more processors 503 to perform the following operations according to the one or more instructions 502:
acquiring query information input by a user, and performing information processing on the query information to obtain a vector group corresponding to the query information, wherein the vector group comprises an information category corresponding to the query information, each key word and an associated word corresponding to each key word;
calling a neural network model which is trained in advance to convert the vector groups to obtain a word vector corresponding to each vector group, wherein the word vector is semantic data corresponding to the vector groups;
acquiring a personal health file of the user, and associating the personal health file with each key word to obtain associated content;
searching each question information corresponding to the associated content in a preset medical knowledge graph, performing information processing on each question information to obtain a question vector group corresponding to each question information, inputting each question vector group into the neural network model, and obtaining a question vector corresponding to each question information;
similarity calculation is carried out on the word vectors and the question vectors, and a target question vector with the highest similarity with the word vectors is determined;
acquiring target question information corresponding to the target question vector, taking a medical answer corresponding to the target question information as an answer of the query information, and displaying the answer of the query information on a preset display page so that the user can look up the answer of the query information.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, the system or system embodiments are substantially similar to the method embodiments and therefore are described in a relatively simple manner, and reference may be made to some of the descriptions of the method embodiments for related points. The above-described system and system embodiments are only illustrative, wherein the units described as separate parts may or may not be physically separate, and the parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both,
to clearly illustrate this interchangeability of hardware and software, various illustrative components and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (8)
1. A medical question-answering method, comprising:
acquiring query information input by a user, and performing information processing on the query information to obtain a vector group corresponding to the query information, wherein the vector group comprises an information category corresponding to the query information, each key word and an associated word corresponding to each key word;
calling a neural network model which is trained in advance to convert the vector groups to obtain a word vector corresponding to each vector group, wherein the word vector is semantic data corresponding to the vector groups;
acquiring a personal health file of the user, and associating the personal health file with each key word to obtain associated content;
searching each question information corresponding to the associated content in a preset medical knowledge graph, performing information processing on each question information to obtain a question vector group corresponding to each question information, inputting each question vector group into the neural network model, and obtaining a question vector corresponding to each question information;
similarity calculation is carried out on the word vectors and each question vector, and a target question vector with the highest similarity with the word vectors is determined;
acquiring target question information corresponding to the target question vector, taking a medical answer corresponding to the target question information as an answer of the query information, and displaying the answer of the query information on a preset display page so that the user can look up the answer of the query information;
wherein the processing the query information to obtain the vector group corresponding to the query information includes:
inputting the query information into a preset entity recognizer, and triggering the entity recognizer to output each key vocabulary corresponding to the query information;
determining the information category corresponding to each key word, and searching the associated word corresponding to each key word through a preset session description protocol SDP;
and generating a vector group corresponding to the query information according to the query information, each key word and the associated word corresponding to each key word.
2. The method of claim 1, wherein the training process of the neural network model comprises:
acquiring a preset initial neural network model and a medical vocabulary data set, wherein the medical vocabulary data set comprises a plurality of medical vocabularies;
setting a standard word vector corresponding to each medical vocabulary, and executing a training process corresponding to the initial neural network model;
wherein the training process comprises: inputting each medical vocabulary in the medical vocabulary data set into the initial neural network model, converting each medical vocabulary through the initial neural network model, outputting a training word vector corresponding to each medical vocabulary, calculating an error value between the training word vector corresponding to each medical vocabulary and a standard word vector, and judging whether the error value between the training word vector corresponding to each medical vocabulary and the standard word vector is not less than a preset error threshold value or not;
when the error value between the training word vector and the standard word vector corresponding to each medical vocabulary is not less than a preset error threshold value, calling a preset loss function to adjust the model parameters of the initial neural network model;
and performing iterative training on the initial neural network model subjected to model parameter adjustment according to the training process, calling the loss function to adjust the model parameters of the initial neural network model after each training until the error value between the training word vector corresponding to each medical vocabulary currently output by the initial neural network model and the standard word vector is smaller than the error threshold value, finishing the training of the initial neural network model and obtaining the neural network model.
3. The method according to claim 1, wherein the searching for each question information corresponding to the associated content in a preset medical knowledge map comprises:
determining each key word as a main node corresponding to the associated content;
searching each branch node corresponding to the main node based on the main node in the medical knowledge graph;
and acquiring user state information corresponding to the user according to the personal health record in the associated content, and searching each question information matched with the user state information in each branch node.
4. The method of claim 1, wherein after calculating the similarity between the word vector and each of the question vectors, further comprising:
judging whether the similarity of each question vector and the word vector is greater than a preset similarity threshold value or not;
when the similarity between at least one target question vector and the word vector is greater than the similarity threshold, determining target question information corresponding to each target question vector, and acquiring medical answers corresponding to each target question information;
and storing each medical answer into a preset answer list from high to low according to the similarity of each corresponding target question vector of each medical answer, and displaying the answer list on the display page so that the user can look up each medical answer according to the answer list.
5. A medical question-answering device, comprising:
the processing unit is used for acquiring query information input by a user, performing information processing on the query information and acquiring a vector group corresponding to the query information, wherein the vector group comprises an information category corresponding to the query information, each key word and an associated word corresponding to each key word;
the conversion unit is used for calling a neural network model which is trained in advance to convert the vector groups to obtain a word vector corresponding to each vector group, and the word vectors are semantic data corresponding to the vector groups;
the first acquisition unit is used for acquiring the personal health file of the user and associating the personal health file with each key word to acquire associated content;
the searching unit is used for searching each question information corresponding to the associated content in a preset medical knowledge map, performing information processing on each question information to obtain a question vector group corresponding to each question information, inputting each question vector group into the neural network model, and obtaining a question vector corresponding to each question information;
the calculation unit is used for calculating the similarity between the word vector and each question vector and determining a target question vector with the highest similarity to the word vector;
the first display unit is used for acquiring target question information corresponding to the target question vector, taking a medical answer corresponding to the target question information as an answer of the query information, and displaying the answer of the query information on a preset display page so that the user can look up the answer of the query information;
wherein the processing unit comprises:
the input subunit is used for inputting the query information into a preset entity recognizer and triggering the entity recognizer to output each key vocabulary corresponding to the query information;
the first searching subunit is used for determining the information category corresponding to each key word and searching the associated word corresponding to each key word through a preset session description protocol SDP;
and the generating subunit is used for generating a vector group corresponding to the query information according to the query information, each key word and the associated word corresponding to each key word.
6. The apparatus of claim 5, further comprising:
the second acquisition unit is used for acquiring a preset initial neural network model and a medical vocabulary data set, and the medical vocabulary data set comprises a plurality of medical vocabularies;
the setting unit is used for setting a standard word vector corresponding to each medical vocabulary and executing a training process corresponding to the initial neural network model; wherein the training process comprises: inputting each medical vocabulary in the medical vocabulary data set into the initial neural network model, converting each medical vocabulary through the initial neural network model, outputting a training word vector corresponding to each medical vocabulary, calculating an error value between the training word vector corresponding to each medical vocabulary and a standard word vector, and judging whether the error value between the training word vector corresponding to each medical vocabulary and the standard word vector is not less than a preset error threshold value or not;
the adjusting unit is used for calling a preset loss function to adjust the model parameters of the initial neural network model when the error value between the training word vector and the standard word vector corresponding to each medical vocabulary is not less than a preset error threshold value;
and the training unit is used for carrying out iterative training on the initial neural network model subjected to model parameter adjustment according to the training process, calling the loss function to adjust the model parameters of the initial neural network model after each training until the error value between the training word vector corresponding to each medical vocabulary currently output by the initial neural network model and the standard word vector is smaller than the error threshold value, finishing the training of the initial neural network model and obtaining the neural network model.
7. The apparatus of claim 5, wherein the lookup unit comprises:
the determining subunit is used for determining each key word as a main node corresponding to the associated content;
the second searching subunit is used for searching each branch node corresponding to the main node based on the main node in the medical knowledge graph;
and the third searching subunit is used for acquiring user state information corresponding to the user according to the personal health record in the associated content, and searching each question information matched with the user state information in each branch node.
8. The apparatus of claim 5, further comprising:
the judging unit is used for judging whether the similarity of each question vector and the word vector is greater than a preset similarity threshold value or not;
the third obtaining unit is used for determining target question information corresponding to each target question vector and obtaining a medical answer corresponding to each target question information when the similarity between at least one target question vector and the word vector is larger than the similarity threshold;
and the second display unit is used for storing each medical answer into a preset answer list from high to low according to the similarity of each corresponding target question vector of each medical answer, and displaying the answer list on the display page so that the user can look up each medical answer according to the answer list.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910677024.3A CN110377719B (en) | 2019-07-25 | 2019-07-25 | Medical question and answer method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910677024.3A CN110377719B (en) | 2019-07-25 | 2019-07-25 | Medical question and answer method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110377719A CN110377719A (en) | 2019-10-25 |
| CN110377719B true CN110377719B (en) | 2022-02-15 |
Family
ID=68255912
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910677024.3A Active CN110377719B (en) | 2019-07-25 | 2019-07-25 | Medical question and answer method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110377719B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111552818A (en) * | 2020-04-27 | 2020-08-18 | 中国银行股份有限公司 | Customer service knowledge base query method and device |
| CN112199482B (en) * | 2020-09-30 | 2023-07-21 | 平安科技（深圳）有限公司 | Dialogue generation method, device, equipment and readable storage medium |
| CN112364065B (en) * | 2020-10-27 | 2022-04-15 | 刘锋 | Method and system for obtaining big data transfer and feedback |
| CN113077895B (en) * | 2021-04-27 | 2022-10-28 | 上海德衡数据科技有限公司 | Software definition-based intelligent HIE platform construction method and electronic equipment |
| CN114155962A (en) * | 2022-02-10 | 2022-03-08 | 北京妙医佳健康科技集团有限公司 | Data cleaning method and method for constructing disease diagnosis by using knowledge graph |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107247868A (en) * | 2017-05-18 | 2017-10-13 | 深思考人工智能机器人科技（北京）有限公司 | A kind of artificial intelligence aids in interrogation system |
| CN108108449A (en) * | 2017-12-27 | 2018-06-01 | 哈尔滨福满科技有限责任公司 | A kind of implementation method based on multi-source heterogeneous data question answering system and the system towards medical field |
| CN109065172A (en) * | 2018-07-04 | 2018-12-21 | 平安科技（深圳）有限公司 | Illness information acquisition method, device, computer equipment and storage medium |
| CN109147934A (en) * | 2018-07-04 | 2019-01-04 | 平安科技（深圳）有限公司 | Interrogation data recommendation method, device, computer equipment and storage medium |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10621218B2 (en) * | 2015-03-30 | 2020-04-14 | Avaya Inc. | Systems and methods for compiling and dynamically updating a collection of frequently asked questions |
| CN107368547A (en) * | 2017-06-28 | 2017-11-21 | 西安交通大学 | A kind of intelligent medical automatic question-answering method based on deep learning |
-
2019
- 2019-07-25 CN CN201910677024.3A patent/CN110377719B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107247868A (en) * | 2017-05-18 | 2017-10-13 | 深思考人工智能机器人科技（北京）有限公司 | A kind of artificial intelligence aids in interrogation system |
| CN108108449A (en) * | 2017-12-27 | 2018-06-01 | 哈尔滨福满科技有限责任公司 | A kind of implementation method based on multi-source heterogeneous data question answering system and the system towards medical field |
| CN109065172A (en) * | 2018-07-04 | 2018-12-21 | 平安科技（深圳）有限公司 | Illness information acquisition method, device, computer equipment and storage medium |
| CN109147934A (en) * | 2018-07-04 | 2019-01-04 | 平安科技（深圳）有限公司 | Interrogation data recommendation method, device, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110377719A (en) | 2019-10-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110377719B (en) | Medical question and answer method and device | |
| CN107609163B (en) | Medical knowledge map generation method, storage medium and server | |
| CN108829822B (en) | Media content recommendation method and device, storage medium and electronic device | |
| US10885278B2 (en) | Auto tele-interview solution | |
| US20180322954A1 (en) | Method and device for constructing medical knowledge graph and assistant diagnosis method | |
| US20180174055A1 (en) | Intelligent conversation system | |
| WO2021139232A1 (en) | Medical knowledge graph-based triage method and apparatus, device, and storage medium | |
| US20150154721A1 (en) | System, apparatus and method for user to obtain service from professional | |
| US20180218127A1 (en) | Generating a Knowledge Graph for Determining Patient Symptoms and Medical Recommendations Based on Medical Information | |
| US20180218126A1 (en) | Determining Patient Symptoms and Medical Recommendations Based on Medical Information | |
| CN110289068A (en) | Drug recommended method and equipment | |
| KR20190016653A (en) | System and method for providing intelligent counselling service | |
| CN112182252B (en) | Intelligent medication question-answering method and device based on medicine knowledge graph | |
| WO2019000905A1 (en) | Triage dialogue method, device, and system | |
| CN112016295A (en) | Symptom data processing method and device, computer equipment and storage medium | |
| CN111933291A (en) | Medical information recommendation device, method, system, equipment and readable storage medium | |
| CN109522397B (en) | Information processing method and device | |
| US20160180051A1 (en) | Method and arrangement for matching of diseases and detection of changes for a disease by the use of mathematical models | |
| WO2020040883A1 (en) | Conversational assistant using extracted guidance knowledge | |
| CN115169364B (en) | Intelligent question-answering method, device, equipment and storage medium | |
| CN111063429A (en) | Medical consultation method, device, equipment and computer-readable storage medium | |
| CN116796857A (en) | LLM model training method, device, equipment and storage medium thereof | |
| Christopherjames et al. | Natural language processing based human assistive health conversational agent for multi-users | |
| CN110929014B (en) | Information processing method, information processing device, electronic equipment and storage medium | |
| Ni et al. | A framework for domain-specific natural language information brokerage |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |