CN117216221A - Intelligent question-answering system based on knowledge graph and construction method - Google Patents
Intelligent question-answering system based on knowledge graph and construction method Download PDFInfo
- Publication number
- CN117216221A CN117216221A CN202311258310.9A CN202311258310A CN117216221A CN 117216221 A CN117216221 A CN 117216221A CN 202311258310 A CN202311258310 A CN 202311258310A CN 117216221 A CN117216221 A CN 117216221A
- Authority
- CN
- China
- Prior art keywords
- question
- user
- candidate
- entity
- knowledge
- 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
- 238000010276 construction Methods 0.000 title claims abstract description 20
- 238000000034 method Methods 0.000 claims description 33
- 238000000605 extraction Methods 0.000 claims description 16
- 238000003062 neural network model Methods 0.000 claims description 8
- 230000011218 segmentation Effects 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 5
- 238000012986 modification Methods 0.000 claims description 4
- 230000004048 modification Effects 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000013135 deep learning Methods 0.000 abstract description 3
- 230000004044 response Effects 0.000 abstract description 3
- 238000002372 labelling Methods 0.000 description 4
- 238000013145 classification model Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000014509 gene expression Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides an intelligent question-answering system based on a knowledge graph and a construction method thereof, wherein the knowledge graph is introduced into the intelligent question-answering system for question-answering, and the knowledge construction of the knowledge graph is marked by an expert, or is obtained from a professional structured database through formatted grabbing, or is obtained by detecting and correcting data in the knowledge graph through a multiple deep learning algorithm, so that the high accuracy of the data in the knowledge graph is ensured. The structured data format and the distributed storage of the knowledge graph provide support for the rapid knowledge retrieval of the computer, the knowledge graph is introduced, the questions input by the user are retrieved and matched with the data in the knowledge graph, and the intelligent question-answering system compares the intelligent question-answering based on the question-answering pair by the high accuracy and the high retrieval rate of the knowledge graph, so that the answer accuracy exceeds 10% and the question-answering response time is reduced by 30%.
Description
Technical Field
The invention belongs to the field of question and answer only, and particularly relates to an intelligent question and answer system based on a knowledge graph and a construction method thereof.
Background
The intelligent question-answering system is a typical application in the field of natural language processing, and is realized based on massive data and deep semantic understanding technology on the Internet. The method is widely applied to fields closely related to people's lives, such as science and technology, education, shopping, life, medical treatment and the like, and greatly improves the efficiency of information acquisition.
The current intelligent question-answering system mainly builds a question-answering knowledge base by accumulating historical question-answering pair data or editing the question-answering pair data by manual experts based on the community question-answering realized by the question-answering pair, matches the questions in the knowledge base by the keywords in the questions proposed by the user, and returns the searched most relevant question-answering data as answers to the user.
The intelligent question-answering method has the problems that a question-answering knowledge base needs to spend a great deal of time cost or labor cost, the question cannot be answered by a cold door question, and the given answer is not standard.
Disclosure of Invention
The technical problem to be solved by the invention is how to quickly construct a question-answering system without spending a great deal of time cost and labor cost, and the answer accuracy is higher, and an intelligent question-answering system based on a knowledge graph and a construction method are provided.
In order to solve the technical problems, the invention adopts the following technical scheme:
the method for constructing the intelligent question-answering system based on the knowledge graph comprises the following steps:
step 1: constructing a question-answer knowledge graph in a certain knowledge field;
step 2: manually analyzing the historical question-answering information to obtain a question template set, a question class set, a query intention set and a keyword-superordinate class word set of user questions in the field of knowledge, and dividing the question template into one question class of the question class set according to the query intention expressed by each question template in the question template set;
step 3: entity identification is carried out on the user problems to obtain the entity, attribute and entity relation of the user problems;
step 4: extracting keywords from a user problem to obtain keywords, replacing the keywords in the user problem by using superior category words to obtain a problem template of the user problem and a problem category to which the problem template belongs, and finding out a query intention expressed by the problem template from the problem template set according to the problem template to which the user problem belongs as a query intention of the user problem;
step 5: generating a query statement in a knowledge graph according to a query intention of a user problem, a problem template and an entity, an attribute and an entity relation of the user problem, and carrying out graph query and graph reasoning from the knowledge graph to obtain a sub-graph in the knowledge graph related to the user problem, wherein a plurality of entities, relations and attributes included in the sub-graph form a candidate entity set, a candidate relation set and a candidate attribute set;
step 6: and carrying out matching degree calculation on the user questions and the candidate answers, and returning the candidate answers with high matching degree to the user as optimal answers, wherein the candidate answers are formed by combining a candidate entity set, a candidate relationship set and candidate entities, candidate relationships and candidate attributes in the candidate attribute set.
Further, the method for constructing the question-answer knowledge graph in the step 1 is as follows:
step 1.1: extracting entities from text corpus related to the field of question-answer knowledge to be constructed, acquiring association relations among the entities, and acquiring entity attribute names and entity attribute values related to the entities during entity extraction;
step 1.2: taking the extracted entities as nodes of the knowledge graph, and taking the association relationship between the extracted entities as edges between the nodes corresponding to the two entities;
step 1.3: the entity attribute name is used as an attribute contained in the node corresponding to the entity, and the entity attribute value is used as an attribute value corresponding to the entity node attribute;
step 1.4: and constructing a knowledge graph according to the nodes, the edges and the attributes and attribute values of the nodes.
Further, the method for entity recognition of the user problem in the step 3 is as follows:
step 3.1: word segmentation is carried out on a problem text input by a user to obtain a word sequence string, and keyword extraction is carried out on the word sequence string through a TF-IDF algorithm to obtain keywords and weight information;
step 3.2: carrying out syntactic dependency analysis on the question text input by the user to obtain binary dependency relations among words in the question text of the user;
step 3.3: and processing and identifying the keywords, the weight information and the binary dependency relationship among the words through a neural network model to obtain the entity, the attribute and the entity relationship of the user problem.
Further, before word segmentation is performed on the question text input by the user in step 3.1, nonsensical characters in the user question text need to be removed, and modification and completion are performed on wrongly written characters and omission in the user question text.
Further, the method for generating the query sentence in the knowledge graph in the step 5 is as follows:
a query sentence template is preset, and the query intention of the user problem, the problem template, and the entity, attribute and entity relation of the user problem are embedded into the corresponding position of the query sentence template to form a query sentence.
Further, the method for carrying out the map reasoning in the step 5 is as follows:
and reasoning in a knowledge graph based on a neural network model according to the entity, the attribute, the relation and the user intention obtained after the text processing of the user problem, wherein the plurality of entities, the relation and the attribute included in the subgraph form a candidate entity set, a candidate relation set and a candidate attribute set in the knowledge graph related to the user problem.
Further, in step 6, the method for calculating the matching degree between the user question and the candidate answer is as follows:
calculating the relevance between the user question and the candidate answers, calculating the relevance between the candidate answers and a question template corresponding to the user question, calculating the relevance between the candidate answers and a keyword in the user question, finally weighting and combining a plurality of relevance to obtain a comprehensive relevance, sorting the candidate answers according to the comprehensive relevance, comparing the comprehensive relevance with a preset threshold, discarding the candidate answers smaller than the threshold, and selecting the candidate answer with the highest comprehensive relevance as the answer matched with the user question.
The invention also provides an intelligent question-answering system based on the knowledge graph, which is constructed by the steps of the intelligent question-answering system construction method based on the knowledge graph.
By adopting the technical scheme, the invention has the following beneficial effects:
according to the knowledge graph-based intelligent question-answering system and the construction method, the knowledge graph is introduced into the intelligent question-answering system for knowledge question-answering, and the knowledge construction of the knowledge graph is obtained by labeling from an expert, or is obtained by formatted grabbing from a professional structured database, or is obtained by detecting and correcting data in the knowledge graph through a multiple deep learning algorithm, so that the high accuracy of the data in the knowledge graph is ensured. The invention provides support for rapid knowledge retrieval of a computer by introducing a knowledge graph and carrying out retrieval matching on questions input by a user and data in the knowledge graph, and the intelligent question-answering system compares intelligent question-answering based on question-answering pairs by the high accuracy and the high retrieval rate of the knowledge graph, so that the answer accuracy exceeds 10% and the question-answering response time is reduced by 30%.
Drawings
FIG. 1 is a flow chart of the system of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made apparent and fully in view of the accompanying drawings, in which some, but not all embodiments of the invention are shown. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The invention discloses a knowledge graph-based intelligent question-answering system construction method, which is shown in figure 1 and comprises the following steps:
step 1: and constructing a question-answer knowledge graph in a certain knowledge field. In this embodiment, the method for constructing the question-answer knowledge graph is as follows:
step 1.1: extracting entities from text corpus related to the field of question-answer knowledge to be constructed, acquiring association relations among the entities, and acquiring entity attribute names and entity attribute values related to the entities during entity extraction;
step 1.2: taking the extracted entities as nodes of the knowledge graph, and taking the association relationship between the extracted entities as edges between the nodes corresponding to the two entities;
step 1.3: the entity attribute name is used as an attribute contained in the node corresponding to the entity, and the entity attribute value is used as an attribute value corresponding to the entity node attribute;
step 1.4: and constructing a knowledge graph according to the nodes, the edges and the attributes and attribute values of the nodes.
In this embodiment, the purpose of constructing the knowledge graph is to use the feature of knowledge graph structuring, so as to facilitate the query. Knowledge graph construction uses information extraction techniques to extract entities, relationships, and entity attributes from structured data existing in the knowledge domain to be constructed, as well as from some semi-structured and unstructured data. After the text corpus is extracted by the entities, a series of discrete named entities are obtained, and in order to obtain semantic information, the association relation among the entities is extracted from the related corpus, and the entities are connected through the relation, so that a meshed knowledge structure can be formed. Currently, the main stream relation extraction is mainly divided into two types: and (5) relationship classification and entity relationship joint extraction under the remote supervision labeling data. The purpose of attribute extraction is to collect attribute information of a specific entity from different information sources, for example, for a certain course, the attribute information of the course name, the time of giving lessons, teachers giving lessons, outline of the course and the like can be obtained from a data source. In comparison with relation extraction, attribute extraction identifies attribute values of entities in addition to attribute names of the entities, and the attribute value structure is also uncertain, so most researches are based on rule extraction. However, the method has certain defects in the data of the field of the professional courses, and because the knowledge graph of the professional courses is different from the knowledge graph of common sense, the method has high requirements on the quality of information, and has lower fault tolerance to the noise of the information. And extracting knowledge elements such as entities, relations, attributes and the like from the original corpus through information extraction, and eliminating ambiguity between entity names and entity objects through knowledge fusion to obtain a series of basic fact expressions. However, the fact itself is not equal to knowledge, a structured, networked knowledge system is required, and knowledge is also required to be processed.
After the knowledge graph is constructed, the structure of the knowledge graph and the data value in the knowledge graph are required to be stored, and when the knowledge graph is stored, a neo4j graph database and an elastic search data storage service are used for storing. The neo4j graph database stores data of a graph structure, and the elastic search stores data which can be used in an intermediate process, such as relationships and directions among nodes in a graph structure, attributes contained in the nodes and data values corresponding to the attributes.
Step 2: and manually analyzing the historical question-answering information to obtain a question template set, a question class set, a query intention set and a keyword upper-level class word set of the user questions in the belonging knowledge field, and dividing the question template into one question class of the question class set according to the query intention expressed by each question template in the question template set.
In this embodiment, the historical question and answer information is fully utilized, so as to analyze what the question template of the frequently asked question is, what the query intention is to be expressed is, the keyword information in the question is replaced by the upper category word of the keyword through the analysis and labeling of the manual expert, the question template of the question is obtained, and the category classification is performed according to the query intention expressed by the question template, so that different question templates contained under each query intention can be obtained, and the query is conveniently performed from the knowledge graph by using the question template with a relatively standard.
Step 3: and carrying out entity identification on the user problem to obtain the entity, attribute and entity relation of the user problem. When a user inputs a text, a lot of nonsensical texts are added in the input text due to language expression habit, so that when the text type data is analyzed, the text itself needs to be processed, and negative influence on a system caused by the text itself is reduced. Therefore, before word segmentation is performed on the question text input by the user, nonsensical characters in the user question text are removed, and modification and complementation are performed on wrongly written characters and omission in the user question text. And then carrying out entity identification on the problem input by the user.
In this embodiment, the method for entity identification of the user problem is:
step 3.1: word segmentation is carried out on a problem text input by a user to obtain a word sequence string, and keyword extraction is carried out on the word sequence string through a TF-IDF algorithm to obtain keywords and weight information; the TF-IDF (term frequency-inverse document frequency) algorithm is a common weighting method for information retrieval and data mining, with which keywords and weights can be extracted.
Step 3.2: carrying out syntactic dependency analysis on the question text input by the user to obtain binary dependency relations among words in the question text of the user; the syntactic dependency structure is a tree-shaped graph structure, the words are associated through the dependency relationship, the words are mapped with the entities in the knowledge graph, and the dependency relationship is mapped with the edges in the knowledge graph.
Step 3.3: and processing and identifying the keywords, the weight information and the binary dependency relationship among the words through a neural network model to obtain the entity, the attribute and the entity relationship of the user problem.
By processing the user problems and extracting the information, the effective structured information is used for inquiring, so that the information can be inquired in the knowledge graph.
Step 4: extracting keywords from the user questions to obtain keywords, replacing the keywords in the user questions by using superior category words to obtain question templates of the user questions and question categories to which the question templates belong, and finding out query intentions expressed by the question templates from the question templates according to the question templates to which the user questions belong to as query intentions of the user questions.
In this embodiment, the method for identifying the intention of the user problem is that the method for identifying the intention of the user is that the neural network classification model identifies the intention of the user to obtain the intention of the user. Specifically used in this embodiment is user intention recognition based on a Bert-based text classification model.
In the embodiment, the keywords in the user questions are replaced by the class words of the upper class, wherein the class words of the upper class refer to class words obtained by analyzing and summarizing the historical question and answer questions in the manual analysis process in the step 2, and the question templates are obtained by replacement. Because the question templates represent a class of questions, a large number of user question instances can be included, so that the user questions can be mapped to the question templates only after keyword extraction and intention recognition are performed on the user questions. The keyword extraction method in the user problem is to divide words of a problem text input by a user to obtain word sequence strings, and extract keywords from the word sequence strings through a TF-IDF algorithm. Examples of keywords such as question templates:
user question text: who is the teacher of the linear algebra course?
Vocabulary and generic class: : linear algebra [ curriculum ] and teacher [ teaching teacher ]
Problems after treatment: who is the teaching teacher?
Meanwhile, the user questions can be classified into categories in which the query intention is query [ teaching teacher ].
Step 5: generating a query statement in a knowledge graph according to the query intention of the user problem, the problem template and the entity, attribute and entity relation of the user problem, and carrying out graph query and graph reasoning from the knowledge graph to obtain a candidate entity, a candidate relation and a candidate attribute.
In this embodiment, the method for generating the query sentence in the knowledge graph is:
a query sentence template is preset, and the query intention of the user problem, the problem template, and the entity, attribute and entity relation of the user problem are embedded into the corresponding position of the query sentence template to form a query sentence.
The query is performed by using a query method provided by a knowledge graph storage database.
In this embodiment, the method for performing map reasoning is:
and according to the entity, the attribute, the relation and the user intention obtained after the text processing of the user problem, reasoning is carried out in the knowledge graph based on the neural network model, so as to obtain a sub-graph in the knowledge graph related to the user problem, wherein a plurality of entities, the relation and the attribute included in the sub-graph form a candidate entity set, a candidate relation set and a candidate attribute set. The neural network model mainly comprises a memory network-based method Key-Value Memory Network, a vector matching method based on a representation learning method, multi-Column Convolutional Neural Networks and the like.
Step 6: and carrying out matching degree calculation on the user questions and the candidate answers, and returning the candidate answers with high matching degree to the user as optimal answers, wherein the candidate answers are formed by combining a candidate entity set, a candidate relationship set and candidate entities, candidate relationships and candidate attributes in the candidate attribute set.
In this embodiment, the method for calculating the matching degree between the user question and the candidate answer is as follows:
calculating the relevance between the user question and the candidate answers, calculating the relevance between the candidate answers and a question template corresponding to the user question, calculating the relevance between the candidate answers and a keyword in the user question, finally weighting and combining a plurality of relevance to obtain a comprehensive relevance, sorting the candidate answers according to the comprehensive relevance, comparing the comprehensive relevance with a preset threshold, discarding the candidate answers smaller than the threshold, and selecting the candidate answer with the highest comprehensive relevance as the answer matched with the user question. And finally, converting the candidate entity, the candidate relation and the candidate attribute in the answer matched with the user question into corresponding natural language text.
In this embodiment, the method for calculating the correlation degree between the user question and the candidate answer is to convert the text into a vector, and calculate the correlation degree between the question and the candidate answer through the neural network model.
The invention also provides an intelligent question-answering system based on the knowledge graph, which is constructed by the steps of the intelligent question-answering system construction method based on the knowledge graph.
According to the knowledge graph-based intelligent question-answering system and the construction method, the knowledge graph is introduced into the intelligent question-answering system for knowledge question-answering, and the knowledge construction of the knowledge graph is obtained by labeling from an expert, or is obtained by formatted grabbing from a professional structured database, or is obtained by detecting and correcting data in the knowledge graph through a multiple deep learning algorithm, so that the high accuracy of the data in the knowledge graph is ensured. The invention provides support for rapid knowledge retrieval of a computer by introducing a knowledge graph and carrying out retrieval matching on questions input by a user and data in the knowledge graph, and the intelligent question-answering system compares intelligent question-answering based on question-answering pairs by the high accuracy and the high retrieval rate of the knowledge graph, so that the answer accuracy exceeds 10% and the question-answering response time is reduced by 30%.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention.
Claims (8)
1. The method for constructing the intelligent question-answering system based on the knowledge graph is characterized by comprising the following steps of:
step 1: constructing a question-answer knowledge graph in a certain knowledge field;
step 2: manually analyzing the historical question-answering information to obtain a question template set, a question class set, a query intention set and a keyword-superordinate class word set of user questions in the field of knowledge, and dividing the question template into one question class of the question class set according to the query intention expressed by each question template in the question template set;
step 3: entity identification is carried out on the user problems to obtain the entity, attribute and entity relation of the user problems;
step 4: extracting keywords from a user problem to obtain keywords, replacing the keywords in the user problem by using superior category words to obtain a problem template of the user problem and a problem category to which the problem template belongs, and finding out a query intention expressed by the problem template from the problem template set according to the problem template to which the user problem belongs as a query intention of the user problem;
step 5: generating a query statement in a knowledge graph according to a query intention of a user problem, a problem template and an entity, an attribute and an entity relation of the user problem, and carrying out graph query or graph reasoning from the knowledge graph to obtain a sub-graph in the knowledge graph related to the user problem, wherein a plurality of entities, relations and attributes included in the sub-graph form a candidate entity set, a candidate relation set and a candidate attribute set;
step 6: and carrying out matching degree calculation on the user questions and the candidate answers, and returning the candidate answers with high matching degree to the user as optimal answers, wherein the candidate answers are formed by combining a candidate entity set, a candidate relationship set and candidate entities, candidate relationships and candidate attributes in the candidate attribute set.
2. The method for constructing an intelligent question-answering system based on a knowledge graph according to claim 1, wherein the method for constructing a question-answering knowledge graph in step 1 is as follows:
step 1.1: extracting entities from text corpus related to the field of question-answer knowledge to be constructed, acquiring association relations among the entities, and acquiring entity attribute names and entity attribute values related to the entities during entity extraction;
step 1.2: taking the extracted entities as nodes of the knowledge graph, and taking the association relationship between the extracted entities as edges between the nodes corresponding to the two entities;
step 1.3: the entity attribute name is used as an attribute contained in the node corresponding to the entity, and the entity attribute value is used as an attribute value corresponding to the entity node attribute;
step 1.4: and constructing a knowledge graph according to the nodes, the edges and the attributes and attribute values of the nodes.
3. The knowledge graph-based intelligent question-answering system construction method according to claim 2, wherein the method for entity recognition of the user questions in step 3 is as follows:
step 3.1: word segmentation is carried out on a problem text input by a user to obtain a word sequence string, and keyword extraction is carried out on the word sequence string through a TF-IDF algorithm to obtain keywords and weight information;
step 3.2: carrying out syntactic dependency analysis on the question text input by the user to obtain binary dependency relations among words in the question text of the user;
step 3.3: and processing and identifying the keywords, the weight information and the binary dependency relationship among the words through a neural network model to obtain the entity, the attribute and the entity relationship of the user problem.
4. The knowledge-graph-based intelligent question-answering system construction method according to claim 3, wherein in step 3.1, before word segmentation is performed on the question text input by the user, meaningless characters in the user question text are removed, and modification and complementation are performed on wrongly written characters and omission in the user question text.
5. The knowledge-based intelligent question-answering system construction method according to claim 1, wherein the method for generating the query sentence in the knowledge-based is as follows:
a query sentence template is preset, and the query intention of the user problem, the problem template, and the entity, attribute and entity relation of the user problem are embedded into the corresponding position of the query sentence template to form a query sentence.
6. The knowledge graph-based intelligent question-answering system construction method according to claim 1, wherein the method for graph reasoning in step 5 is as follows:
and reasoning in a knowledge graph based on a neural network model according to the entity, the attribute, the relation and the user intention obtained after the text processing of the user problem, wherein the plurality of entities, the relation and the attribute included in the subgraph form a candidate entity set, a candidate relation set and a candidate attribute set in the knowledge graph related to the user problem.
7. The knowledge-graph-based intelligent question-answering system construction method according to claim 6, wherein the method for calculating the matching degree between the user question and the candidate answer in step 6 is as follows:
calculating the correlation degree between the user question and the candidate answers, matching the candidate answers with a question template corresponding to the user question, correlating the candidate answers with keywords in the user question, finally weighting and combining the multiple correlations to obtain a comprehensive correlation, sorting the candidate answers according to the comprehensive correlation, comparing the comprehensive correlation with a preset threshold, discarding the candidate answers smaller than the threshold, and selecting the candidate answer with the highest comprehensive correlation as an answer matched with the user question, wherein the candidate answer is obtained by converting a candidate entity, a candidate relation and a candidate attribute into a corresponding natural language text.
8. A knowledge-graph-based intelligent question-answering system, characterized in that the system constructed by the steps of the knowledge-graph-based intelligent question-answering system construction method according to any one of claims 1 to 7 is used.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311258310.9A CN117216221A (en) | 2023-09-27 | 2023-09-27 | Intelligent question-answering system based on knowledge graph and construction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311258310.9A CN117216221A (en) | 2023-09-27 | 2023-09-27 | Intelligent question-answering system based on knowledge graph and construction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117216221A true CN117216221A (en) | 2023-12-12 |
Family
ID=89042278
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311258310.9A Pending CN117216221A (en) | 2023-09-27 | 2023-09-27 | Intelligent question-answering system based on knowledge graph and construction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117216221A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117540004A (en) * | 2024-01-10 | 2024-02-09 | 安徽省优质采科技发展有限责任公司 | Industrial domain intelligent question-answering method and system based on knowledge graph and user behavior |
-
2023
- 2023-09-27 CN CN202311258310.9A patent/CN117216221A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117540004A (en) * | 2024-01-10 | 2024-02-09 | 安徽省优质采科技发展有限责任公司 | Industrial domain intelligent question-answering method and system based on knowledge graph and user behavior |
| CN117540004B (en) * | 2024-01-10 | 2024-03-22 | 安徽省优质采科技发展有限责任公司 | Industrial domain intelligent question-answering method and system based on knowledge graph and user behavior |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111475623B (en) | Case Information Semantic Retrieval Method and Device Based on Knowledge Graph | |
| CN110765257B (en) | Intelligent consulting system of law of knowledge map driving type | |
| CN109255031B (en) | Data processing method based on knowledge graph | |
| Mondal et al. | A course recommendation system based on grades | |
| CN113806563B (en) | Architect knowledge graph construction method for multi-source heterogeneous building humanistic historical material | |
| CN112163077B (en) | Knowledge graph construction method for field question and answer | |
| CN110990590A (en) | Dynamic financial knowledge map construction method based on reinforcement learning and transfer learning | |
| CN107491655B (en) | Liver disease information intelligent consultation system based on machine learning | |
| CN113191148B (en) | Rail transit entity identification method based on semi-supervised learning and clustering | |
| CN112597316B (en) | Method and device for interpretive reasoning question-answering | |
| CN114036281B (en) | Knowledge graph-based citrus control question-answering module construction method and question-answering system | |
| CN111159356B (en) | Knowledge graph construction method based on teaching content | |
| Lalata et al. | A sentiment analysis model for faculty comment evaluation using ensemble machine learning algorithms | |
| CN113168499A (en) | Method for searching patent document | |
| CN112328800A (en) | System and method for automatically generating programming specification question answers | |
| CN114238653B (en) | Method for constructing programming education knowledge graph, completing and intelligently asking and answering | |
| CN116860978B (en) | Primary school Chinese personalized learning system based on knowledge graph and large model | |
| CN113196277A (en) | System for retrieving natural language documents | |
| CN109522396B (en) | Knowledge processing method and system for national defense science and technology field | |
| CN113196278A (en) | Method for training a natural language search system, search system and corresponding use | |
| CN117216221A (en) | Intelligent question-answering system based on knowledge graph and construction method | |
| CN116737967A (en) | Knowledge graph construction and perfecting system and method based on natural language | |
| Zhao | RETRACTED ARTICLE: Application of deep learning algorithm in college English teaching process evaluation | |
| CN113868406B (en) | Search method, search system, and computer-readable storage medium | |
| CN114840685A (en) | Emergency plan knowledge graph construction method |
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 |