CN111291135A - Knowledge graph construction method and device, server and computer readable storage medium - Google Patents
Knowledge graph construction method and device, server and computer readable storage medium Download PDFInfo
- Publication number
- CN111291135A CN111291135A CN202010068813.XA CN202010068813A CN111291135A CN 111291135 A CN111291135 A CN 111291135A CN 202010068813 A CN202010068813 A CN 202010068813A CN 111291135 A CN111291135 A CN 111291135A
- Authority
- CN
- China
- Prior art keywords
- nodes
- entity
- data
- literal
- node
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/288—Entity relationship models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Computational Linguistics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application relates to a knowledge graph construction method and device, a server and a computer readable storage medium, comprising the following steps: and acquiring resource data of the preset field, and preprocessing the resource data of the preset field to obtain preprocessed resource data. And carrying out knowledge modeling on the preprocessed resource data to obtain a knowledge graph body, wherein the knowledge graph body comprises entity nodes, literal volume nodes, relations among the entity nodes and the literal volume nodes. Extracting triple data which accord with the knowledge graph body from the preprocessed resource data, wherein the triple data is composed of entity nodes, literal quantity nodes, the relation among the entity nodes and the relation between the entity nodes and the literal quantity nodes, and the knowledge graph is constructed according to the triple data. And the entity nodes and the literal nodes are respectively extracted from the resource data, so that the resource data can be better classified, and the retrieval efficiency of the constructed knowledge graph is improved.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to a method and an apparatus for constructing a knowledge graph, a server, and a computer-readable storage medium.
Background
A knowledge graph is a graph-based data structure that is semantic network in nature. Unlike general knowledge maps, industry knowledge maps are domain-specific knowledge maps. The resource data in a specific field has stronger specialization, so that the constructed industry knowledge graph needs to have higher availability and retrieval efficiency. The application capability of the established knowledge graph is directly determined by the effective modeling method, and the subsequent application capability of the knowledge graph can be enabled to a great extent. In contrast, a poorly modeled knowledge profile will encounter several bottlenecks in later use. The retrieval efficiency of the industry knowledge graph constructed by the traditional modeling method is low. Therefore, the problem of low efficiency of traditional industry knowledge graph retrieval is urgently to be solved.
Disclosure of Invention
The embodiment of the application provides a knowledge graph construction method, a knowledge graph construction device, a server and a computer readable storage medium, and the retrieval efficiency of the constructed knowledge graph can be improved.
A knowledge graph construction method comprises the following steps:
acquiring resource data of a preset field;
preprocessing the resource data of the preset field to obtain preprocessed resource data;
performing knowledge modeling on the preprocessed resource data to obtain a knowledge graph body, wherein the knowledge graph body comprises entity nodes, literal volume nodes, relations among the entity nodes, and relations among the entity nodes and the literal volume nodes;
extracting triple data which accord with the knowledge graph body from the preprocessed resource data, wherein the triple data consists of the entity nodes, the literal volume nodes, the relationship among the entity nodes and the relationship between the entity nodes and the literal volume nodes;
and constructing a knowledge graph according to the triple data.
In one embodiment, the performing knowledge modeling on the preprocessed resource data to obtain a knowledge graph ontology, where the knowledge graph ontology includes entity nodes, literal volume nodes, relationships between the entity nodes, and relationships between the entity nodes and the literal volume nodes, and includes:
extracting nodes from the preprocessed resource data;
acquiring nodes with the importance degree higher than a preset threshold value from the nodes as entity nodes, and acquiring nodes with the importance degree smaller than or equal to the preset threshold value from the nodes as literal nodes;
acquiring the relationship between the entity nodes and the literal volume nodes;
and obtaining a knowledge graph body according to the entity nodes, the literal quantity nodes, the relationship among the entity nodes and the literal quantity nodes.
In one embodiment, the obtaining the relationship between the entity nodes and the literal volume node includes:
and extracting the relationship between the entity nodes and the literal volume nodes based on the context information of the preprocessed resource data.
In one embodiment, the triple data conforming to the ontology of the knowledge-graph comprises a first triple data and a second triple data, the first triple data comprises a first entity node, a second entity node and a relationship between the first entity node and the second entity node, and the second triple data comprises a third entity node, a first literal node, a relationship between the third entity node and the first literal node;
extracting triple data conforming to the knowledge graph ontology from the preprocessed resource data comprises the following steps:
and extracting the first ternary group data and the second ternary group data from the preprocessed resource data.
In one embodiment, the method further comprises:
setting built-in attributes for the entity node, the literal measure node and the relationship respectively;
the extracting the first triple data and the second triple data from the preprocessed resource data includes:
extracting contents corresponding to built-in attributes of a first entity node and a second entity node in the first triple data and extracting contents corresponding to built-in attributes of a relationship between the first entity node and the second entity node in the first triple data from the preprocessed resource data;
and extracting the content corresponding to the built-in attributes of the third entity node and the first literal volume node in the second ternary group data and extracting the content corresponding to the built-in attributes of the relationship between the third entity node and the first literal volume node in the second ternary group data from the preprocessed resource data.
In one embodiment, the built-in properties of the entity node, the literal volume node, and the relationship include built-in date properties including a creation date, an update date, and an expiration date.
In one embodiment, the preprocessing the resource data in the preset domain to obtain preprocessed resource data includes:
and distinguishing and processing the homonymous entity nodes and homonymous denomination nodes in the resource data of the preset field to obtain distinguished and processed resource data.
In one embodiment, the method further comprises:
and identifying the relationship between the entity nodes by adopting a first identifier, and identifying the relationship between the entity nodes and the literal volume nodes by adopting a second identifier.
A knowledge-graph building apparatus comprising:
the resource data acquisition module is used for acquiring resource data of a preset field;
the resource data preprocessing module is used for preprocessing the resource data in the preset field to obtain preprocessed resource data;
the knowledge graph ontology generating module is used for carrying out knowledge modeling on the preprocessed resource data to obtain a knowledge graph ontology, and the knowledge graph ontology comprises entity nodes, literal volume nodes, relations among the entity nodes and relations among the entity nodes;
the triple data extraction module is used for extracting triple data required by the knowledge graph body from the preprocessed resource data;
and the knowledge graph building module is used for building a knowledge graph according to the triple data.
A server comprising a memory and a processor, the memory having stored therein a computer program which, when executed by the processor, causes the processor to carry out the steps of the above method.
A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method as above.
The knowledge graph construction method, the knowledge graph construction device, the server and the computer readable storage medium acquire resource data of a preset field, and preprocess the resource data of the preset field to obtain preprocessed resource data. And carrying out knowledge modeling on the preprocessed resource data to obtain a knowledge graph body, wherein the knowledge graph body comprises entity nodes, literal volume nodes, relations among the entity nodes and the literal volume nodes. Extracting triple data which accord with the knowledge graph body from the preprocessed resource data, wherein the triple data is composed of entity nodes, literal quantity nodes, the relation among the entity nodes and the relation between the entity nodes and the literal quantity nodes, and the knowledge graph is constructed according to the triple data. Compared with the traditional knowledge graph which only comprises entity nodes and the relation between the entity nodes, the knowledge graph body obtained by performing knowledge modeling on the preprocessed resource data in the application also comprises the literal nodes and the relation between the entity nodes and the literal nodes. And the entity nodes and the literal volume nodes are respectively extracted from the resource data, the knowledge in the resource data can be classified through the two different nodes of the entity nodes and the literal volume nodes, and meanwhile, the relationship between the entity nodes and the literal volume nodes is correspondingly increased. Therefore, the resource data can be better classified, and the retrieval efficiency of the constructed knowledge graph is improved.
Drawings
In order to more clearly illustrate the embodiments of the present application 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 some embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a diagram of an application environment in one embodiment;
FIG. 2 is a flow diagram of a method for knowledge graph construction in one embodiment;
FIG. 3 is a flowchart of the method of FIG. 2 for obtaining a knowledge graph ontology by performing knowledge modeling on the preprocessed resource data;
FIG. 4 is a flow diagram of a method of knowledge-graph construction in another embodiment;
FIG. 5 is an example of a knowledge graph in a particular embodiment;
FIG. 6 is a block diagram showing the structure of a knowledge-graph constructing apparatus according to an embodiment;
FIG. 7 is a block diagram of the knowledge-graph ontology generating module of FIG. 6;
fig. 8 is a schematic diagram of an internal configuration of a server in one embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
It will be understood that, as used herein, the terms "first," "second," "third," and the like may be used herein to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish one element from another.
The knowledge graph construction method provided by the application can be applied to an application environment graph shown in FIG. 1. The application environment includes a terminal 120 and a server 140, and the terminal 120 and the server 140 are connected via a network. The server 140 may obtain the resource data in the preset field, and preprocess the resource data in the preset field to obtain the preprocessed resource data. And carrying out knowledge modeling on the preprocessed resource data to obtain a knowledge graph body, wherein the knowledge graph body comprises entity nodes, literal volume nodes, relations among the entity nodes and the literal volume nodes. Extracting triple data which accord with the knowledge graph body from the preprocessed resource data, wherein the triple data is composed of entity nodes, literal quantity nodes, the relation among the entity nodes and the relation between the entity nodes and the literal quantity nodes, and the knowledge graph is constructed according to the triple data. The terminal 120 may obtain the retrieval entry input by the user, send the retrieval entry to the server 140, the server 140 finds the relevant knowledge data from the constructed knowledge graph according to the retrieval entry and returns the relevant knowledge data to the terminal 120, and the terminal 120 displays the retrieval result to the user.
Fig. 2 is a flowchart of a method for constructing a knowledge graph in an embodiment, and as shown in fig. 2, a method for constructing a knowledge graph is provided, which is applied to a server and includes steps 210 to 290.
The preset field may refer to a professional field, and is relative to a general field. For example, the predetermined fields include professional fields such as medical fields, insurance fields, and judicial fields. The method is mainly applied to the domain knowledge graph under the vertical knowledge graph. The resource data refers to the collected knowledge data in the field, and can be collected in various ways. For example, the data may be collected from a relational database in the field, collected from an internet site, and the like, and the present application is not limited thereto. And then resource data in the preset field is obtained through various ways.
And step 230, preprocessing the resource data of the preset field to obtain preprocessed resource data.
The preprocessing refers to a process of removing noise from the resource data in the preset field and converting the resource data into a uniform format. The collected resource data inevitably has error, repeated or invalid information, so that the noise of the resource data in the preset field is removed, the accuracy of the resource data is improved, the resource data after the noise is removed is converted into a uniform format, and the knowledge modeling is conveniently carried out according to the preprocessed resource data.
And 250, performing knowledge modeling on the preprocessed resource data to obtain a knowledge graph body, wherein the knowledge graph body comprises entity nodes, literal quantity nodes, relations among the entity nodes, and relations among the entity nodes and the literal quantity nodes.
The ontology is a term set for describing a domain, and the organizational structure of the ontology is hierarchical and can be used as a framework and a foundation of a knowledge base. The knowledge graph ontology is a skeleton formed by knowledge concepts and the relations among the knowledge concepts. Essentially, a knowledge graph is intended to describe various entities or concepts and their relationships that exist in the real world, which constitute a huge semantic network graph. Nodes in the semantic network graph represent entities or concepts, and edges in the semantic network graph are formed by relationships between the entities or concepts.
Specifically, an Entity (Entity) refers to something that is distinct and independent. Such as a person, a city, a plant, etc., a commodity, etc., for example, "china", "beijing", "shanghai", etc. An entity is the most basic element in a knowledge graph, and different relationships exist among different entities. Some entities are screened out as Literal volumes according to needs or scenes, and the Literal volumes (Literal) refer to something which is distinctive and independent, but the Literal volumes focus more on describing attributes. The knowledge in the resource data can be classified through two different nodes, namely the entity node and the literal volume node, and the efficiency of subsequent retrieval according to the knowledge graph can be greatly improved.
Specifically, when knowledge modeling is performed on the preprocessed resource data, firstly, nodes are extracted from the preprocessed resource data; secondly, dividing the nodes into entity nodes and literal nodes; and finally, acquiring the relationship between the entity nodes and the literal volume nodes. The knowledge graph ontology is formed by all entity nodes, literal volume nodes, the relationships among the entity nodes and the literal volume nodes.
And 270, extracting triple data which accord with the knowledge graph body from the preprocessed resource data, wherein the triple data consists of entity nodes, literal quantity nodes, the relationship among the entity nodes and the literal quantity nodes.
After the knowledge graph ontology is constructed, the preprocessed resource data needs to be filled into the knowledge graph ontology according to the format of the knowledge graph ontology, so that the knowledge graph in the field is obtained. The format of the knowledge-graph ontology herein generally refers to triple data composed of entity nodes, literal volume nodes, relationships between entity nodes, and relationships between entity nodes and literal volume nodes, such as triple data composed of entity node-relationship-entity node or entity node-relationship-literal volume node.
For example, when a knowledge map is constructed in the automotive field, if one entity is a vehicle model and the other entity is an engine, the two entities are connected by the relationship of the engine model. If one entity is the vehicle type and the other literal quantity is the vehicle amount, the relation between the entity and the literal quantity is connected according to the price instructed by the manufacturer. The resource data is extracted according to the format of the vehicle type, the engine, the vehicle type, the manufacturer instruction price and the vehicle amount to form three groups of data.
And 290, constructing a knowledge graph according to the triple data.
After the resource data are all converted into triple data which accord with the knowledge graph ontology, the knowledge graph can be constructed according to the triple data. Therefore, the knowledge graph is formed by combing the resource data, and the knowledge graph not only comprises the knowledge nodes such as the entity nodes and the literal quantity nodes, but also comprises the relationship between the entity nodes and the literal quantity nodes, so that the knowledge graph can help a user to analyze problems from the perspective of relationship.
In the embodiment of the application, a knowledge graph construction method is provided, and resource data in a preset field are preprocessed by acquiring the resource data in the preset field to obtain preprocessed resource data. And carrying out knowledge modeling on the preprocessed resource data to obtain a knowledge graph body, wherein the knowledge graph body comprises entity nodes, literal volume nodes, relations among the entity nodes and the literal volume nodes. Extracting triple data which accord with the knowledge graph body from the preprocessed resource data, wherein the triple data is composed of entity nodes, literal quantity nodes, the relation among the entity nodes and the relation between the entity nodes and the literal quantity nodes, and the knowledge graph is constructed according to the triple data. Compared with the traditional knowledge graph which only comprises entity nodes and the relation between the entity nodes, the knowledge graph body obtained by performing knowledge modeling on the preprocessed resource data in the application also comprises the literal nodes and the relation between the entity nodes and the literal nodes. And the entity nodes and the literal volume nodes are respectively extracted from the resource data, the knowledge in the resource data can be classified through the two different nodes of the entity nodes and the literal volume nodes, and meanwhile, the relationship between the entity nodes and the literal volume nodes is correspondingly increased. Therefore, the resource data can be better classified, and the retrieval efficiency of the constructed knowledge graph is improved.
In one embodiment, as shown in fig. 3, in step 250, performing knowledge modeling on the preprocessed resource data to obtain a knowledge graph ontology, where the knowledge graph ontology includes entity nodes, literal volume nodes, relationships between the entity nodes, and relationships between the entity nodes and the literal volume nodes, and includes:
at step 252, nodes are extracted from the preprocessed resource data.
Nodes in the knowledge graph represent entities or concepts, and thus classification algorithms can be employed to extract nodes from the preprocessed resource data. For example, a knowledge graph is constructed in the automobile field, and a classification algorithm is adopted to extract nodes from the preprocessed resource data. For example, nodes such as a vehicle type, an engine, a vehicle amount, and the like are extracted, which is certainly not exhaustive, and the present application does not limit this.
And 254, acquiring nodes with the importance degrees higher than the preset threshold value from the nodes as entity nodes, and acquiring nodes with the importance degrees smaller than or equal to the preset threshold value from the nodes as literal nodes.
After all nodes are extracted from the preprocessed resource data, classifying all the nodes according to the importance degree according to the requirements and scenes for constructing the domain knowledge graph. Specifically, a value corresponding to the importance degree of each node is calculated, the node with the importance degree higher than a preset threshold value is used as an entity node, and the node with the importance degree smaller than or equal to the preset threshold value is used as a literal node. For example, if the requirement for constructing the domain knowledge map is more inclined to the performance comparison of the engine, the node of the engine can be used as the entity node, and the node of the vehicle amount can be used as the literal node.
And 256, acquiring the relationship between the entity nodes and the literal volume nodes.
Wherein the edges in the knowledge-graph are formed by the relationships between these entities or concepts. After the nodes are divided into the entity nodes and the literal nodes, the relationships between the entity nodes and the literal nodes are further obtained. There is no correlation between literal nodes, and there is no next node for a literal node.
And step 258, obtaining a knowledge graph ontology according to the entity nodes, the literal volume nodes, the relationship among the entity nodes and the literal volume nodes.
Specifically, after the entity nodes and the literal volume nodes are defined, and the relationships between the entity nodes and the literal volume nodes are obtained, the knowledge graph ontology can be constructed by simultaneously having points and edges in the knowledge graph.
In the embodiment of the application, after the nodes are extracted from the preprocessed resource data, the importance degrees of all the nodes are calculated according to the requirements and scenes for constructing the knowledge graph in the field, and the nodes are classified according to the importance degrees. Specifically, a value corresponding to the importance degree of each node is calculated, the node with the importance degree higher than a preset threshold value is used as an entity node, and the node with the importance degree smaller than or equal to the preset threshold value is used as a literal node. Therefore, the nodes are divided into different categories, and when the data required to be inquired by the user belongs to the entity node type, the data only needs to be searched in the entity node type data in the knowledge graph, and the data does not need to be searched in all the nodes like the traditional method. And each entity node has a label, entities which are not in the searched category can be quickly filtered out through the labels during knowledge graph retrieval, and then the entities are searched through names in the searched label range. And the nodes are indexed in advance according to the names, and the retrieval speed can be further improved by retrieving through the indexes.
Obviously, the knowledge graph constructed in the application is adopted for retrieval, so that the retrieval difficulty is reduced, and the retrieval efficiency is improved. And the nodes are divided according to the importance degree, so that the division of primary and secondary nodes in the resource data is realized, and the division of primary and secondary relations is also realized, so that the core important nodes and relations are highlighted, the venation of the knowledge graph is clearer, and the user can grasp key knowledge more conveniently.
In one embodiment, obtaining the relationship between the entity nodes and the literal volume node includes:
and extracting the relationship between the entity nodes and the literal nodes based on the context information of the preprocessed resource data.
Specifically, based on the context information of the preprocessed resource data, the relationship between the entity nodes and the two relationships between the entity nodes and the literal nodes are extracted. Because not every entity in the ontology of the knowledge graph has a relationship, and the ontology of the knowledge graph mainly represents the relationship between the entities which is more important and meets the expected requirement, but does not represent some secondary relationships which are irrelevant to the expected requirement. Therefore, after the relation between the entity nodes or the relation between the entity nodes and the literal volume node is extracted, whether the extracted relation belongs to the relation in the knowledge graph body or not is judged, if yes, the extracted relation is information required for constructing the knowledge graph, and the extracted relation is reserved. And if the relation between the extracted entity nodes or the relation between the entity nodes and the literal volume nodes is judged not to belong to the relation in the knowledge graph body, the extracted relation is not the information required for constructing the knowledge graph, and the extracted relation is discarded.
For example, since china is an entity and beijing is an entity, the capital is the relationship between the entity china and the entity beijing, china-capital-beijing constitutes a (entity-relationship-entity) triple data sample. Beijing is an entity, population is a relation, quantity is a literal quantity, and Beijing-population-quantity forms a (entity-relation-literal quantity) triple data sample.
In the embodiment of the application, based on the context information of the preprocessed resource data, the relationship between the entity nodes and the literal nodes can be accurately extracted. And abandoning the extracted relation which does not belong to the knowledge graph ontology, thereby only keeping the relation which belongs to the knowledge graph ontology. Therefore, the field knowledge graph has stronger specialty and better meets the requirements of users, so that the retrieval efficiency is improved.
In one embodiment, the triple data conforming to the knowledge-graph ontology comprises a first triple data and a second triple data, the first triple data comprises a first entity node, a second entity node and a relation between the first entity node and the second entity node, and the second triple data comprises a third entity node, a first literal node, a relation between the third entity node and the first literal node;
extracting triple data conforming to the knowledge graph ontology from the preprocessed resource data, wherein the triple data comprises the following steps:
and extracting the first ternary group data and the second ternary group data from the preprocessed resource data.
Specifically, the resource data in the preset field is acquired, the noise of the resource data in the preset field is removed, and the resource data after the noise is removed is converted into a uniform format, so that the preprocessed resource data is obtained. The knowledge graph ontology is composed of all entity nodes, literal volume nodes, relationships among the entity nodes and the literal volume nodes. Therefore, the triple data which accord with the knowledge graph ontology are extracted from the preprocessed resource data, and the triple data comprise the first triple data and the second triple data. The first triple data is composed of a first entity node, a second entity node, and a relation between the first entity node and the second entity node, and the second triple data is composed of a third entity node, a first literal node, a relation between the third entity node and the first literal node.
For example, china is an entity, and beijing is an entity, and the capital is the relationship between the entity china and the entity beijing, so that china-capital-beijing constitutes a (entity-relationship-entity) triple data sample, i.e. the first triple data. Beijing is an entity, population is a relation, quantity is a literal quantity, and Beijing-population-quantity forms a (entity-relation-literal quantity) triple data sample, namely, second triple data. And extracting the ternary group data which accord with the knowledge graph ontology from all the resource data in the field, thereby completing the process of constructing the knowledge graph by the ternary group data.
In the embodiment of the application, all nodes are divided into entity nodes and literal nodes according to the importance degree or the core degree according to the requirements and scenes for constructing the knowledge graph in the field. The relationships in the knowledge graph are thus also divided into relationships between entity nodes and between entity nodes, and relationships between entity nodes and literal quantities. Therefore, the data in the constructed knowledge graph is composed of two triples of data, one is triple data composed of entity nodes, entity nodes and relations among the entity nodes, and the other is triple data composed of entity nodes, literal volume nodes and relations among the entity nodes and the literal volume nodes. Compared with the traditional knowledge graph which only comprises one triple data, the two triple data in the application classify the resource data more finely and highlight entity nodes which are more core-important, so that the retrieval difficulty is reduced, and the retrieval efficiency is improved.
In one embodiment, the knowledge-graph construction method further comprises:
respectively setting built-in attributes for the entity node, the literal quantity node and the relationship;
extracting first ternary group data and second ternary group data from the preprocessed resource data, wherein the first ternary group data and the second ternary group data comprise:
extracting contents corresponding to built-in attributes of a first entity node and a second entity node in the first triple data and extracting contents corresponding to built-in attributes of the relationship between the first entity node and the second entity node in the first triple data from the preprocessed resource data;
and extracting the content corresponding to the built-in attributes of the third entity node and the first literal volume node in the second type of ternary group data and extracting the content corresponding to the built-in attributes of the relationship between the third entity node and the first literal volume node in the second type of ternary group data from the preprocessed resource data.
Specifically, as shown in fig. 4, a method for constructing a knowledge graph is provided, which includes:
and step 410, acquiring resource data of the preset field.
And step 420, preprocessing the resource data in the preset field to obtain preprocessed resource data.
And 430, performing knowledge modeling on the preprocessed resource data to obtain a knowledge graph body, wherein the knowledge graph body comprises entity nodes, literal quantity nodes, relations among the entity nodes, and relations among the entity nodes and the literal quantity nodes.
Specifically, the resource data in the preset field is acquired through multiple ways, the noise of the resource data in the preset field is removed, and the resource data after the noise is removed is converted into a uniform format. Then, when knowledge modeling is carried out on the preprocessed resource data, firstly, nodes are extracted from the preprocessed resource data; secondly, dividing the nodes into entity nodes and literal nodes; and finally, acquiring the relationship between the entity nodes and the literal volume nodes. The knowledge graph ontology is formed by all entity nodes, literal volume nodes, the relationships among the entity nodes and the literal volume nodes.
Wherein the built-in attributes are used to describe properties, characteristics, etc. of the nodes. The entity node, the literal node and the relationship respectively comprise a plurality of built-in attributes. For example, built-in attributes include a name attribute, a category attribute, a date attribute, a value, a unit, and the like. The built-in attributes under different entity nodes may be the same or different. Similarly, the built-in attributes in different relationships may be the same or different.
And step 450, extracting the content corresponding to the built-in attributes of the first entity node and the second entity node in the first triple data and extracting the content corresponding to the built-in attribute of the relationship between the first entity node and the second entity node in the first triple data from the preprocessed resource data.
After the first type of ternary group data and the second type of ternary group data are extracted from the preprocessed resource data, the content corresponding to the built-in attributes of the first entity node and the second entity node in the first type of ternary group data and the content corresponding to the built-in attributes of the relationship between the first entity node and the second entity node in the first type of ternary group data are extracted from the preprocessed resource data.
And extracting the content corresponding to the built-in attributes of the third entity node and the first literal volume node in the second type of ternary group data and extracting the content corresponding to the built-in attributes of the relationship between the third entity node and the first literal volume node in the second type of ternary group data from the preprocessed resource data. Therefore, the resource data is filled into the knowledge graph body according to the format of the triple data and the built-in attribute.
And 470, constructing a knowledge graph according to the first ternary group data and the second ternary group data.
And constructing a knowledge graph by using all the first triple data and the second triple data in the resource data.
In the embodiment of the application, built-in attributes are respectively set for the entity node, the literal volume node and the relationship, and when the triple data are extracted from the preprocessed resource data, the built-in attributes under the triple data are extracted together. The built-in attributes contain more and more detailed information, so that the contained information is more comprehensive and more logical according to the triple data and the knowledge graph constructed by the built-in attributes under the triple data, the retrieval difficulty can be reduced, and the retrieval efficiency can be improved.
In one embodiment, the built-in properties of the entity node, the literal node, and the relationship include built-in date properties including a creation date, an update date, and an effective date.
Specifically, the built-in attributes of the entity node, the literal volume node, and the relationship include a built-in date attribute, and the built-in date attribute includes a creation date "createdate", an update date "lastmodified date", an effective date "expireDate", and the like. Wherein "createdate" represents the creation date of this node, "lastModifiedDate" is the modification date of this node, and "expireDate" is the date that the node should be destroyed. The three built-in date attributes can be used for maintaining the knowledge graph, and data which are out of date in the knowledge graph can be cleaned up quickly through a program. The data types supported by the literal nodes are classified into INTEGER, DOUBLE, FLOAT, decode, bolt, date, STRING, ENUM, etc., which are not exhaustive.
In addition, when the knowledge map is used in a field where the update frequency is fast, the three date-embedded attributes "creationDate", "lastModifiedDate", and expireDate "may be modified to" creationDate "," lastModifiedDatetime ", and" expireDatetime ", respectively, to hold time information with higher accuracy. For example, creationDate can only be accurate to date (e.g., 2019-01-06), while creationDatetime not only has a date, but also can be accurate to up to minutes of seconds or even to thousandths of a second (e.g., 2020-01-06T03:15:40.415), similarly analogizes to "lastModifiedDate", "expireDate". Therefore, the requirement that the knowledge graph in the field can timely clear out-of-date data can be met through the modified date built-in attribute, and the updating rate of the information in the knowledge graph body is improved.
In one embodiment, the preprocessing the resource data in the preset field to obtain the preprocessed resource data includes:
and distinguishing and processing the homonymous entity nodes and homonymous denomination nodes in the resource data of the preset field to obtain the distinguished and processed resource data.
Specifically, the preprocessing of the resource data in the preset field includes data deduplication processing, data completion processing, data unification processing, and the like. And acquiring resource data with a uniform format after removing noise after preprocessing. The data deduplication processing comprises the steps of judging whether meanings represented by the same-name entity nodes and the same-name volume nodes in the resource data of the preset field are consistent, and fusing the same-name entity nodes and the same-name volume nodes if the meanings are consistent. And if the meanings of the homonymous entity nodes and homonymous denomination nodes in the resource data of the preset field are judged to be inconsistent, deleting ambiguous nodes or distinguishing the ambiguous nodes. Therefore, a plurality of homonymous nodes do not exist in the knowledge graph, namely each node is unique and has unique definition.
In the embodiment of the application, when the resource data in the preset field is preprocessed, the homonymic entity nodes and homonymic denomination nodes in the resource data in the preset field are distinguished and processed. Therefore, each node in the knowledge graph is guaranteed to have uniqueness, correct information can be retrieved more quickly and accurately by retrieving through the knowledge graph, and too much noise is avoided being retrieved. And the same node is only stored once in the knowledge graph, so that the storage space is greatly saved.
In one embodiment, the knowledge-graph construction method further comprises:
the relationship between the entity nodes is identified by a first identifier, and the relationship between the entity nodes and the literal node is identified by a second identifier.
Specifically, in order to distinguish the relationship between the entity nodes from the relationship between the entity nodes and the literal volume node, the relationship between the entity nodes is identified by using the first identifier, and the relationship between the entity nodes and the literal volume node is identified by using the second identifier. Specifically, the first identifier means that the type in the built-in property is ObjectProperty, that is, the type in the built-in property of the relationship between the entity nodes is represented by ObjectProperty. The first identifier means that the type in the built-in property is DateProperty, i.e., the type in the built-in property of the relationship between the entity node and the literal node is represented by DateProperty. Of course, the relationship between entity nodes may also be referred to as relationship, and the relationship between an entity node and a literal node may also be referred to as Property.
In the embodiment of the application, the relation between the entity nodes and the literal volume nodes are distinguished in the knowledge graph in the form of identifiers, so that the resource data can be better classified, and the retrieval efficiency of the constructed knowledge graph is improved.
In a specific embodiment, as shown in fig. 5, an example of a knowledge graph of an automotive domain is shown, however, fig. 5 shows only a portion of the knowledge graph of the automotive domain and does not represent the entirety of the knowledge graph of the automotive domain. As shown, the triple data of the vehicle model-engine entity node-relation-entity node is included, and the triple data of the vehicle model-manufacturer guide price-amount entity node-attribute-literal volume node is also included.
The direction of the vector in the triple data of the entity node, the attribute and the literal volume node is from the vehicle type to the engine, and the direction of the vector in the triple data of the entity node, the attribute and the literal volume node is from the vehicle type to the amount. Vehicle type-manufacturer-instructed price-amount data, which means that the meaning of "the manufacturer-instructed price of vehicle type X6 is 76.69 ten thousand yuan".
The built-in attributes of this entity node of the vehicle model include a name attribute "name", an alias "aliases", a creation date "createdate", an update date "lastModifiedDate", an effective date "expireDate", and the like. Similarly, the built-in attributes of this physical node of the engine also include the 5-point information described above. The amount is a literal volume node whose built-in attributes include a name attribute "name", a value ", a unit", a creation date "createdate", an update date "lastModifiedDate", an effective date "expireDate", and the like.
The engine model is the relationship between the vehicle model and the engine in the triple data of vehicle model-engine. The built-in attributes in the relationship of the engine model include type, creation date "createdate", update date "lastModifiedDate", effective date "expireDate", and the like, and since the engine model is a relationship between the entity nodes, the type corresponding to the engine model is ObjectProperty. Similarly, the built-in attribute under the attribute of the vendor guide price also includes the 4-point information, and the vendor guide price is the relationship between the entity node and the literal volume node, so the type corresponding to the vendor guide price is the dataproperty.
In the embodiment of the application, firstly, the nodes in the resource data are divided into entity nodes and literal volume nodes according to the importance degree, and a knowledge graph ontology is constructed based on the entity nodes, the literal volume nodes, the relationship among the entity nodes and the literal volume nodes. Secondly, built-in attributes are set for the entity nodes, the literal nodes and the relationships respectively. And finally, extracting ternary group data which accords with the knowledge graph ontology from the preprocessed resource data, wherein the ternary group data comprises first ternary group data and second ternary group data. And constructing a knowledge graph of the field according to the two ternary sets of data. The nodes are divided according to the importance degree, so that the primary and secondary nodes in the resource data are divided, and the primary and secondary relationships are divided, so that the important nodes and relationships of the core are highlighted, the venation of the knowledge graph is clearer, and the user can grasp key knowledge more conveniently.
In one embodiment, a knowledge-graph building apparatus 600 is provided, as shown in fig. 6, comprising:
a resource data obtaining module 610, configured to obtain resource data of a preset domain;
the resource data preprocessing module 630 is configured to preprocess resource data in a preset field to obtain preprocessed resource data;
the knowledge graph ontology generating module 650 is configured to perform knowledge modeling on the preprocessed resource data to obtain a knowledge graph ontology, where the knowledge graph ontology includes entity nodes, literal nodes, relationships between the entity nodes, and relationships between the entity nodes and the literal nodes;
the triple data extraction module 670 is configured to extract triple data required by the knowledge graph ontology from the preprocessed resource data;
and a knowledge graph building module 690 for building a knowledge graph according to the triple data.
In one embodiment, as shown in fig. 7, the knowledge-graph ontology generating module 650 includes:
a node extraction unit 652 configured to extract a node from the preprocessed resource data;
the entity node and literal node dividing unit 654 is configured to obtain, from the nodes, a node whose importance degree is higher than a preset threshold as an entity node, and obtain, from the nodes, a node whose importance degree is less than or equal to the preset threshold as a literal node;
a relationship obtaining unit 656, configured to obtain relationships between entity nodes and literal nodes;
the knowledge graph ontology generating unit 658 is configured to obtain a knowledge graph ontology according to the entity nodes, the literal volume nodes, the relationships between the entity nodes, and the relationships between the entity nodes and the literal volume nodes.
In one embodiment, the relationship obtaining unit 656 is further configured to extract the relationship between the entity nodes and the literal node based on the context information of the preprocessed resource data.
In one embodiment, the triple data conforming to the knowledge-graph ontology comprises a first triple data and a second triple data, the first triple data comprises a first entity node, a second entity node and a relation between the first entity node and the second entity node, and the second triple data comprises a third entity node, a first literal node, a relation between the third entity node and the first literal node;
the triple data extracting module 670 is further configured to extract the first triple data and the second triple data from the preprocessed resource data.
In one embodiment, there is provided a knowledge-graph constructing apparatus, further comprising: the built-in attribute setting module is used for respectively setting built-in attributes for the entity node, the literal volume node and the relationship;
the triple data extraction module 670 is further configured to extract, from the preprocessed resource data, contents corresponding to built-in attributes of a first entity node and a second entity node in the first triple data, and contents corresponding to built-in attributes of a relationship between the first entity node and the second entity node in the first triple data; and extracting the content corresponding to the built-in attributes of the third entity node and the first literal volume node in the second type of ternary group data and extracting the content corresponding to the built-in attributes of the relationship between the third entity node and the first literal volume node in the second type of ternary group data from the preprocessed resource data.
In one embodiment, the built-in properties of the entity node, the literal node, and the relationship include built-in date properties including a creation date, an update date, and an effective date.
In one embodiment, the resource data preprocessing module 630 is further configured to perform distinguishing processing on the homonymous entity node and the homonymous denomination node in the resource data in the preset field to obtain the distinguished and processed resource data.
In one embodiment, a knowledge graph building apparatus is provided and is further configured to identify relationships between entity nodes using a first identifier and identify relationships between entity nodes and literal nodes using a second identifier.
The division of each module in the knowledge-graph constructing apparatus is only used for illustration, and in other embodiments, the knowledge-graph constructing apparatus may be divided into different modules as needed to complete all or part of the functions of the knowledge-graph constructing apparatus.
Fig. 8 is a schematic diagram of an internal configuration of a server in one embodiment. As shown in fig. 8, the server includes a processor and a memory connected by a system bus. Wherein, the processor is used for providing calculation and control capability and supporting the operation of the whole server. The memory may include a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The computer program can be executed by a processor for implementing a method for constructing a knowledge graph provided in the following embodiments. The internal memory provides a cached execution environment for the operating system computer programs in the non-volatile storage medium. The server may be a mobile phone, a tablet computer, or a personal digital assistant or a wearable device, etc.
The implementation of each module in the knowledge graph constructing apparatus provided in the embodiments of the present application may be in the form of a computer program. The computer program may be run on a terminal or a server. The program modules constituted by the computer program may be stored on the memory of the terminal or the server. Which when executed by a processor, performs the steps of the method described in the embodiments of the present application.
The embodiment of the application also provides a computer readable storage medium. One or more non-transitory computer-readable storage media containing computer-executable instructions that, when executed by one or more processors, cause the processors to perform the steps of the method of knowledge graph construction.
A computer program product containing instructions which, when run on a computer, cause the computer to perform a method of knowledge-graph construction.
Any reference to memory, storage, database, or other medium used by embodiments of the present application may include non-volatile and/or volatile memory. Suitable non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM), which acts as external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), Enhanced SDRAM (ESDRAM), synchronous Link (Synchlink) DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and bus dynamic RAM (RDRAM).
The above examples only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (11)
1. A knowledge graph construction method is characterized by comprising the following steps:
acquiring resource data of a preset field;
preprocessing the resource data of the preset field to obtain preprocessed resource data;
performing knowledge modeling on the preprocessed resource data to obtain a knowledge graph body, wherein the knowledge graph body comprises entity nodes, literal volume nodes, relations among the entity nodes, and relations among the entity nodes and the literal volume nodes;
extracting triple data which accord with the knowledge graph body from the preprocessed resource data, wherein the triple data consists of the entity nodes, the literal volume nodes, the relationship among the entity nodes and the relationship between the entity nodes and the literal volume nodes;
and constructing a knowledge graph according to the triple data.
2. The method of claim 1, wherein the knowledge modeling of the preprocessed resource data yields a knowledge-graph ontology that includes entity nodes, literal nodes, relationships between the entity nodes, and relationships between the entity nodes and the literal nodes, and comprises:
extracting nodes from the preprocessed resource data;
acquiring nodes with the importance degree higher than a preset threshold value from the nodes as entity nodes, and acquiring nodes with the importance degree smaller than or equal to the preset threshold value from the nodes as literal nodes;
acquiring the relationship between the entity nodes and the literal volume nodes;
and obtaining a knowledge graph body according to the entity nodes, the literal quantity nodes, the relationship among the entity nodes and the literal quantity nodes.
3. The method of claim 2, wherein obtaining the relationship between the entity nodes and the literal node comprises:
and extracting the relationship between the entity nodes and the literal volume nodes based on the context information of the preprocessed resource data.
4. The method according to any one of claims 1 to 3, wherein the triple data conforming to the ontology of the knowledge-graph comprises a first triple data and a second triple data, the first triple data comprises a first entity node, a second entity node, and a relationship between the first entity node and the second entity node, and the second triple data comprises a third entity node, a first literal node, a relationship between the third entity node and the first literal node;
extracting triple data conforming to the knowledge graph ontology from the preprocessed resource data comprises the following steps:
and extracting the first ternary group data and the second ternary group data from the preprocessed resource data.
5. The method of claim 4, further comprising:
setting built-in attributes for the entity node, the literal measure node and the relationship respectively;
the extracting the first triple data and the second triple data from the preprocessed resource data includes:
extracting contents corresponding to built-in attributes of a first entity node and a second entity node in the first triple data and extracting contents corresponding to built-in attributes of a relationship between the first entity node and the second entity node in the first triple data from the preprocessed resource data;
and extracting the content corresponding to the built-in attributes of the third entity node and the first literal volume node in the second ternary group data and extracting the content corresponding to the built-in attributes of the relationship between the third entity node and the first literal volume node in the second ternary group data from the preprocessed resource data.
6. The method of claim 5, wherein the built-in properties of the entity node, the literal node, and the relationship include built-in date properties including a creation date, an update date, and an expiration date.
7. The method according to claim 1, wherein the preprocessing the resource data in the preset domain to obtain preprocessed resource data includes:
and distinguishing and processing the homonymous entity nodes and homonymous denomination nodes in the resource data of the preset field to obtain distinguished and processed resource data.
8. The method of claim 1, further comprising:
and identifying the relationship between the entity nodes by adopting a first identifier, and identifying the relationship between the entity nodes and the literal volume nodes by adopting a second identifier.
9. A knowledge-graph building apparatus, comprising:
the resource data acquisition module is used for acquiring resource data of a preset field;
the resource data preprocessing module is used for preprocessing the resource data in the preset field to obtain preprocessed resource data;
the knowledge graph ontology generating module is used for carrying out knowledge modeling on the preprocessed resource data to obtain a knowledge graph ontology, and the knowledge graph ontology comprises entity nodes, literal volume nodes, relations among the entity nodes and relations among the entity nodes;
the triple data extraction module is used for extracting triple data required by the knowledge graph body from the preprocessed resource data;
and the knowledge graph building module is used for building a knowledge graph according to the triple data.
10. A server comprising a memory and a processor, the memory having stored thereon a computer program, wherein the computer program, when executed by the processor, causes the processor to perform the steps of the method of knowledge-graph construction according to any of claims 1 to 8.
11. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of knowledge-graph construction according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010068813.XA CN111291135A (en) | 2020-01-21 | 2020-01-21 | Knowledge graph construction method and device, server and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010068813.XA CN111291135A (en) | 2020-01-21 | 2020-01-21 | Knowledge graph construction method and device, server and computer readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111291135A true CN111291135A (en) | 2020-06-16 |
Family
ID=71029941
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010068813.XA Pending CN111291135A (en) | 2020-01-21 | 2020-01-21 | Knowledge graph construction method and device, server and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111291135A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112100392A (en) * | 2020-07-28 | 2020-12-18 | 浙江大学 | Equipment fault tracing method based on reinforcement learning and knowledge graph |
| CN112445919A (en) * | 2021-02-01 | 2021-03-05 | 深圳追一科技有限公司 | Knowledge graph construction method and device, server and computer readable storage medium |
| CN112612908A (en) * | 2021-01-05 | 2021-04-06 | 上海云扣科技发展有限公司 | Natural resource knowledge graph construction method and device, server and readable memory |
| CN112733874A (en) * | 2020-10-23 | 2021-04-30 | 招商局重庆交通科研设计院有限公司 | Suspicious vehicle discrimination method based on knowledge graph reasoning |
| CN113505889A (en) * | 2021-07-23 | 2021-10-15 | 中国平安人寿保险股份有限公司 | Processing method and device of atlas knowledge base, computer equipment and storage medium |
| CN113505889B (en) * | 2021-07-23 | 2024-08-02 | 中国平安人寿保险股份有限公司 | Processing method and device of mapping knowledge base, computer equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103493049A (en) * | 2011-02-28 | 2014-01-01 | 汇丰控股有限公司 | Computer system and database for reputation and knowledge management |
| US8799177B1 (en) * | 2010-07-29 | 2014-08-05 | Intuit Inc. | Method and apparatus for building small business graph from electronic business data |
| CN110489561A (en) * | 2019-07-12 | 2019-11-22 | 平安科技（深圳）有限公司 | Knowledge mapping construction method, device, computer equipment and storage medium |
-
2020
- 2020-01-21 CN CN202010068813.XA patent/CN111291135A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8799177B1 (en) * | 2010-07-29 | 2014-08-05 | Intuit Inc. | Method and apparatus for building small business graph from electronic business data |
| CN103493049A (en) * | 2011-02-28 | 2014-01-01 | 汇丰控股有限公司 | Computer system and database for reputation and knowledge management |
| CN110489561A (en) * | 2019-07-12 | 2019-11-22 | 平安科技（深圳）有限公司 | Knowledge mapping construction method, device, computer equipment and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| 张延星、王广祥、朱志芸、张蝶依: "基于知识图谱词义消歧的文本聚类方法", 《华北理工大学学报(自然科学版)》 * |
| 沈思，王晓文，王璇，杨伏生: "《信息获取与利用研究》", 30 September 2016 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112100392A (en) * | 2020-07-28 | 2020-12-18 | 浙江大学 | Equipment fault tracing method based on reinforcement learning and knowledge graph |
| CN112100392B (en) * | 2020-07-28 | 2022-03-15 | 浙江大学 | Equipment fault tracing method based on reinforcement learning and knowledge graph |
| CN112733874A (en) * | 2020-10-23 | 2021-04-30 | 招商局重庆交通科研设计院有限公司 | Suspicious vehicle discrimination method based on knowledge graph reasoning |
| CN112612908A (en) * | 2021-01-05 | 2021-04-06 | 上海云扣科技发展有限公司 | Natural resource knowledge graph construction method and device, server and readable memory |
| CN112445919A (en) * | 2021-02-01 | 2021-03-05 | 深圳追一科技有限公司 | Knowledge graph construction method and device, server and computer readable storage medium |
| CN113505889A (en) * | 2021-07-23 | 2021-10-15 | 中国平安人寿保险股份有限公司 | Processing method and device of atlas knowledge base, computer equipment and storage medium |
| CN113505889B (en) * | 2021-07-23 | 2024-08-02 | 中国平安人寿保险股份有限公司 | Processing method and device of mapping knowledge base, computer equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110765275B (en) | Search method, search device, computer equipment and storage medium | |
| CN111291135A (en) | Knowledge graph construction method and device, server and computer readable storage medium | |
| CN109446343B (en) | Public safety knowledge graph construction method | |
| WO2019200752A1 (en) | Semantic understanding-based point of interest query method, device and computing apparatus | |
| KR101999152B1 (en) | English text formatting method based on convolution network | |
| CN109582799B (en) | Method and device for determining knowledge sample data set and electronic equipment | |
| CN110458324B (en) | Method and device for calculating risk probability and computer equipment | |
| AU2013329525C1 (en) | System and method for recursively traversing the internet and other sources to identify, gather, curate, adjudicate, and qualify business identity and related data | |
| EP3528180A1 (en) | Method, system and terminal for normalizingentities in a knowledge base, and computer readable storage medium | |
| CN112612908A (en) | Natural resource knowledge graph construction method and device, server and readable memory | |
| CN110990390B (en) | Data cooperative processing method, device, computer equipment and storage medium | |
| CN113254630B (en) | Domain knowledge map recommendation method for global comprehensive observation results | |
| CN117150050B (en) | Knowledge graph construction method and system based on large language model | |
| CN110851609A (en) | Representation learning method and device | |
| CN111221982A (en) | Information processing method, information processing device, computer-readable storage medium and computer equipment | |
| CN112559526A (en) | Data table export method and device, computer equipment and storage medium | |
| CN109933692A (en) | Establish the method and apparatus of mapping relations, the method and apparatus of information recommendation | |
| CN114860941A (en) | Industry data management method and system based on data brain | |
| WO2021047327A1 (en) | Method and apparatus for constructing target concept map, computer device, and storage medium | |
| CN112765183B (en) | Multi-source data fusion method and device, storage medium and electronic equipment | |
| CN108460116B (en) | Search method, search device, computer equipment, storage medium and search system | |
| CN115905705A (en) | Industrial algorithm model recommendation method based on industrial big data | |
| CN113849644A (en) | Text classification model configuration method and device, computer equipment and storage medium | |
| CN114780700A (en) | Intelligent question-answering method, device, equipment and medium based on machine reading understanding | |
| CN112685574B (en) | Method and device for determining hierarchical relationship of domain terms |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200616 |