CN110110155B - Character knowledge graph attribute acquisition method and device based on meta-social relationship circle - Google Patents

Character knowledge graph attribute acquisition method and device based on meta-social relationship circle Download PDF

Info

Publication number
CN110110155B
CN110110155B CN201910266511.0A CN201910266511A CN110110155B CN 110110155 B CN110110155 B CN 110110155B CN 201910266511 A CN201910266511 A CN 201910266511A CN 110110155 B CN110110155 B CN 110110155B
Authority
CN
China
Prior art keywords
attribute
entity
knowledge graph
meta
social relationship
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201910266511.0A
Other languages
Chinese (zh)
Other versions
CN110110155A (en
Inventor
尹美娟
刘晓楠
邱庆云
郑燕
王奕博
高文龙
王灿
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Information Engineering University of PLA Strategic Support Force
Original Assignee
Information Engineering University of PLA Strategic Support Force
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Information Engineering University of PLA Strategic Support Force filed Critical Information Engineering University of PLA Strategic Support Force
Priority to CN201910266511.0A priority Critical patent/CN110110155B/en
Publication of CN110110155A publication Critical patent/CN110110155A/en
Application granted granted Critical
Publication of CN110110155B publication Critical patent/CN110110155B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/901—Indexing; Data structures therefor; Storage structures
    • G06F16/9024—Graphs; Linked lists
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/02—Knowledge representation; Symbolic representation
    • G06N5/022—Knowledge engineering; Knowledge acquisition
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/01—Social networking

Abstract

The invention belongs to the technical field of data mining, and particularly relates to a character knowledge graph attribute acquisition method and device based on a meta-social relationship ring, wherein the method comprises the following steps: traversing users in a social network, and constructing a figure knowledge graph, wherein the figure knowledge graph comprises a figure entity set, a semantic edge set among the figure entities, a relation set corresponding to the semantic edges and a mapping function from the semantic edges to the relation set, and each figure entity is set with a known attribute; and (4) presuming unknown attributes of the users based on the meta-social relationship circle of the users and iteratively updating attribute values in the figure knowledge graph until each user attribute value in the social network is not updated. The method avoids the problems of attribute value swing and the like, is limited by the conditions of the meta-social relationship circle, has high convergence speed in the iterative process, limits the consistency of the specific social relationship and the corresponding reasonable attribute type, has more accurate inferred unknown attribute, and has important guiding significance for the figure attribute inference based on the figure relationship.

Description

Character knowledge graph attribute acquisition method and device based on meta-social relationship circle
Technical Field
The invention belongs to the technical field of data mining, and particularly relates to a character knowledge graph attribute acquisition method and device based on a meta-social relationship ring.
Background
The method for constructing the figure knowledge graph has very wide application prospect in the fields of civil market, safety and the like. The knowledge graph is a good knowledge management and organization tool by fusing information from various data sources and describing the information in a normalized ontology. However, the data source range of the character attribute information and the limitation of the information extraction technology cause information loss to a certain extent in the preliminarily constructed character knowledge graph, inference learning is an important method for perfect construction of the knowledge graph, and the character knowledge graph inference is to complement the relation and the attribute of the preliminarily constructed character knowledge graph loss by using the semantic characteristics of the knowledge graph, so that a complete character knowledge graph is constructed. How to deduce the missing person attribute information becomes a key problem for constructing a complete knowledge graph.
At present, the entity attribute reasoning method based on the knowledge graph is less, while the attribute reasoning method for social network users is more researched, and the method mainly comprises a classifier-based reasoning method, a majority iterative voting-based reasoning method and a community mining-based reasoning method. The classifier-based method is an early and widely-used method, the classifier is used for classifying users according to the known attributes of the users, and the unknown attributes of the users are presumed according to the attribute information of other users in the classes. The reality of the majority voting-based method is that social relationships tend to be established between users in the social network who have the same attribute, so that most neighbors of a user in the social network are considered to have a certain attribute, and the user also has the attribute. The method determines the attribute of the current unknown node according to the label with the most attributes in the neighbor node in each iteration, and the phenomenon that the unknown node swings between two or more attributes in the iteration process can occur, so that the problems of slow convergence and low accuracy of the guessed result in the attribute guessing process are caused. The attribute presumption method based on community mining considers that in a social network, due to the fact that certain attributes are shared among users, the phenomenon that the attributes of node parts in a community tend to be consistent occurs, the social network attribute presumption problem is converted into a community structure optimization problem, optimal community division is solved, and the missing attributes of the users are presumed according to the known attributes in the community where the users are located. According to the method, communities are divided only according to the attribute information of the nodes, the relationship among users is not considered, and the condition that the association of the nodes in the communities is not tight can occur, so that the presumed result is inaccurate. Although the relationship among users is considered in the voting algorithm, the result is inaccurate because the types of social relationships with different neighbor nodes are not considered, and a certain attribute is presumed according to the social relationships of different types.
Disclosure of Invention
Therefore, the invention provides a figure knowledge graph attribute acquisition method and device based on meta-social relationship circles, solves the problems of slow convergence of an attribute presumption iteration process, low accuracy of a presumption result and the like in the current figure knowledge graph, is used for presuming attributes of figures in the figure knowledge graph to construct a complete figure knowledge graph, and has strong practicability and operability.
According to the design scheme provided by the invention, the character knowledge graph attribute acquisition method based on the meta-social relationship ring comprises the following contents:
A) traversing users in a social network, and constructing a figure knowledge graph, wherein the figure knowledge graph comprises a figure entity set, a semantic edge set among the figure entities, a relation set corresponding to the semantic edges and a mapping function from the semantic edges to the relation set, and each figure entity is set with a known attribute;
B) and (4) presuming unknown attributes of the users based on the meta-social relationship circle of the users and iteratively updating attribute values in the figure knowledge graph until each user attribute value in the social network is not updated.
In the above description, in a), the person knowledge map is represented by KG ═ V, E, R, F, where V represents a set of person entities, E represents a set of semantic edges between the person entities, R represents a set of relationships corresponding to the semantic edges, and F represents a mapping function from the semantic edges to R.
In the step B), the meta-social relationship circle of the user is obtained according to the mapping function between the person entities in the person knowledge graph.
Preferably, according to the person knowledge graph, for the person entity, all the neighbor nodes of the person entity are traversed, and the meta social relationship circle set related to the person entity is determined.
In the step B), based on the attribute characteristics corresponding to the social relationship shared among the character entities with the same social relationship, the unknown attribute of the character entity in the character knowledge graph is presumed by using a majority iterative voting mechanism.
Preferably, in the iteration process, the number of votes obtained for the new attribute value each time the attribute value of the character entity is updated is recorded, after each round of voting is finished, the attribute value with the largest number of votes obtained is taken as the candidate attribute value for the attribute type, and when the number of votes obtained for the candidate attribute value is larger than the number of votes obtained for the attribute value in the previous round of updating, the attribute value of the character entity is updated.
Preferably, in the attribute value voting, the meta social relationship circle of the character entity is traversed, the attribute block data of the neighbor node in the meta social relationship circle is acquired, and the attribute value of the corresponding attribute type is voted according to whether the attribute type and the attribute value of the neighbor node attribute block data are the same.
Preferably, in the updating of the attribute value by the majority iterative voting mechanism, a character entity updating flag bit is set, the attribute type of the current character entity vote obtaining record is traversed for each character entity in the meta social relationship circle, the current character entity attribute value is updated according to the number of votes obtained by the attribute type candidate attribute value, and the character entity updating flag bit is recorded until all the attribute types of the current character entity are traversed.
Furthermore, the invention also provides a character knowledge graph attribute acquisition device based on the meta-social relationship ring, which comprises: a building module and an updating module, wherein,
the system comprises a construction module, a database module and a database module, wherein the construction module is used for traversing users in the social network and constructing a figure knowledge graph, the figure knowledge graph comprises a figure entity set, a semantic edge set between figure entities, a relation set corresponding to the semantic edge and a mapping function from the semantic edge to the relation set, and each figure entity is set with a known attribute;
and the updating module is used for presuming unknown attributes of the users based on the meta-social relationship circles of the users and iteratively updating the attribute values in the figure knowledge graph until each user attribute value in the social network is not updated.
The invention has the beneficial effects that:
the invention is based on the characteristic that the character entities with the same social relationship often share the attribute corresponding to the social relationship, and uses the majority iteration voting thought to guess the unknown attribute of the character entities in the character knowledge map, in the iteration process, the label is updated only when the number of votes obtained by the new label is higher than the number of votes obtained by the previous round of labels, so the problem of attribute value swing can not occur, and the invention is limited by the condition of the meta-social relationship circle, the convergence speed of the iteration process is high, the defined meta-social relationship circle can not only ensure the close relationship of the users participating in the voting, but also limit the consistency of the specific social relationship and the corresponding reawayable attribute type, so the estimated unknown attribute is more accurate, so as to reflect the social relationship of the character entities more comprehensively and accurately, and has important guiding significance for the character attribute reasonment based on the character relationship.
Description of the drawings:
FIG. 1 is a flowchart of person knowledge graph attribute updating in an embodiment;
FIG. 2 is a schematic diagram of attribute reasoning in an embodiment;
FIG. 3 is a schematic diagram of a person knowledge graph attribute obtaining device in an embodiment;
FIG. 4 is an example of a person knowledge graph in an embodiment.
The specific implementation mode is as follows:
in order to make the objects, technical solutions and advantages of the present invention clearer and more obvious, the present invention is further described in detail below with reference to the accompanying drawings and technical solutions. The technical terms involved in the examples are as follows:
the Knowledge map (also called scientific Knowledge map) is a Knowledge domain visualization or Knowledge domain mapping map in the book intelligence world, and is a series of different graphs for displaying the relationship between the Knowledge development process and the structure, describing Knowledge resources and carriers thereof by using a visualization technology, and mining, analyzing, constructing, drawing and displaying Knowledge and the mutual relation among the Knowledge resources and the carriers. The theory and method of applying mathematics, graphics, information visualization technology, information science and other disciplines are combined with the method of metrology citation analysis, co-occurrence analysis and the like, and the visual map is utilized to vividly show the core structure, development history, frontier field and overall knowledge framework of the disciplines to achieve the modern theory of multi-discipline fusion, thereby providing practical and valuable reference for discipline research. In order to solve the problems of slow convergence of an attribute presumption iteration process, low accuracy of a presumed result and the like in the current figure knowledge graph, the embodiment of the invention provides a figure knowledge graph attribute acquisition method based on a meta-social relationship ring, which is shown in figure 1 and comprises the following steps:
traversing users in a social network, and constructing a figure knowledge graph, wherein the figure knowledge graph comprises a figure entity set, a semantic edge set among the figure entities, a relation set corresponding to the semantic edges and a mapping function from the semantic edges to the relation set, and each figure entity is set with a known attribute;
and (4) presuming unknown attributes of the users based on the meta-social relationship circle of the users and iteratively updating attribute values in the figure knowledge graph until each user attribute value in the social network is not updated.
The figure knowledge graph reasoning utilizes the semantic characteristics of the knowledge graph to carry out incompletion on the preliminarily established figure knowledge graph missing relation and attributes, so that a complete figure knowledge graph is constructed, and knowledge management and organization are carried out by fusing various data source information. Further, in the embodiment of the present invention, the person knowledge map may be defined as KG ═ { V, E, R, F }, where V ═ u ═ F }, where V ═ u may be defined as1,u2,…uNRepresents a set of people entities, E { (u) } { (u)i,uj) I, j 1,2 … N represents a set of semantic edges between people entities, R1,r2,……rZRepresents a corresponding relation set of the semantic edges, F represents a mapping function of the semantic edges to R,
Figure BDA0002017010890000041
for each human entity uiE.g. for V
Figure BDA0002017010890000042
Description uiIn the context of a known attribute,
Figure BDA0002017010890000043
Figure BDA0002017010890000051
further, the meta-social relationship circle of the user is obtained according to the mapping function between the person entities in the person knowledge graph. Let ui，uj，ukIs three human entities of human knowledge map KG ═ { V, E, R, F }, if ui，uj，ukSatisfy the requirement of
Figure BDA0002017010890000052
Figure BDA0002017010890000053
Then call ui，uj，ukForm a meta social relationship ring, denoted as (u)i，uj，uk)。
The attribute voting rule based on the meta social relationship circle is as follows: let ui，uj，ukIs three person entities of person knowledge graph KG ═ { V, E, R, F }, if person entity ui、uj、ukSatisfies F (u)i，uj)∩F(ui，uk)∩F(uj，uk)＝riThen the three form a social relationship riThereby sharing a social relationship riCorresponding attribute type and attribute value. At this time, if two of the people entities are in the attribute type ajHave the same attribute value vjThen a isjWill become a social relationship riCorresponding potential attribute type vjReferred to as attribute values shared with each other, so that a third persona also tends to be in attribute type ajThereon possess vjAn attribute value.
And after each round of voting is finished, regarding a certain attribute type, taking the attribute value with the largest number of votes as a candidate attribute value. And updating the attribute value according to the voting result of the character entity. The number of votes obtained for the new attribute value is recorded each time the attribute value of the character entity is updated, and the attribute value of the character entity is updated only when the number of votes obtained for the candidate attribute value is larger than the number of votes obtained for the attribute value when the attribute value is updated last time.
Referring to fig. 2, the meta social relationship circle mining is performed according to the character knowledge graph, voting is performed based on the original social relationship circle, so as to update the attribute type attribute value, and the character knowledge graph attribute reasoning is completed, and the specific steps can be designed as follows: step 1, traversing each user in the social network and recording as ui；
Step 2, based on user uiThe meta social relationship circle votes for the unknown attribute and updates the attribute value, which can be specifically designed as follows:
step 2.1, mining the meta-social relationship circle, wherein the mining method comprises the following contents:
step 2.1.1 determination of human entity uiA set of related meta-social relationship circles. For the character entity uiE.g. V, go through uiAll neighbor nodes, if uiWith neighbor node uj,ukSatisfy the requirement of
Figure BDA0002017010890000054
Then (u)i,uj,uk) Composition uiThe meta social relationship circle. u. ofiIs represented as a set of meta-social relationship circles
Figure BDA0002017010890000055
Figure BDA0002017010890000056
Step 2.2, voting reasoning is carried out on the attributes based on the meta-social relationship circle;
step 2.2.1 traverse uiThe meta social relationship circle of (u)i,uj,uk)
Step 2.2.2 traversal of u in Meta social relationship circlejAttribute block data of (2), note
Figure BDA0002017010890000057
Step 2.2.3 go through u in the meta social relationship circlekAttribute block data of (2), note
Figure BDA0002017010890000058
If u isjIs/are as follows
Figure BDA0002017010890000059
And ukIs/are as follows
Figure BDA00020170108900000510
The attribute types are the same and corresponding to the same attribute values, step 2.2.5 is carried out, otherwise, the step is continued until all the attributes are traversed
Figure BDA00020170108900000511
Step 2.2.4 if ujIs/are as follows
Figure BDA00020170108900000512
And ukIs/are as follows
Figure BDA00020170108900000513
The attribute types are the same, corresponding to the same attribute values, and uiAttribute data in
Figure BDA00020170108900000514
Attribute type does not contain vjGo to step 2.2.5, otherwise uiRecorded for obtaining tickets
Figure BDA00020170108900000515
V of typejThe value vote is set to 1.
Step 2.2.5uiThe ticket record contains vjGo to step 2.2.6, otherwise uiRecorded for obtaining tickets
Figure BDA00020170108900000516
V of typejThe value vote is set to 1.
Step 2.2.6uiRecorded for obtaining tickets
Figure BDA0002017010890000061
V of typejThe value is added to 1.
Step 2.2.7 repeat steps 2.2.3-2.2.6 until ukAnd finishing taking the attribute block.
Step 2.2.8 repeat steps 2.2.2-2.2.7 until ujAnd finishing taking the attribute block.
Step 2.2.9 repeating step 2.2.2-2.2.8, taking out uiAll meta social relationship circles.
And 2.3, updating the attributes, wherein the attributes comprise the following contents:
step 2.3.1 human entity uiThe update flag bit for this round is initialized to 0.
Step 2.3.2 ergodic personage uiThe attribute type of the ticket obtaining record is marked as aj。
Step 2.3.3 if the iteration is the first iteration, step 2.3.3.1-2.3.3.3
Step 2.3.3.1 human entity uiAt ajSetting the number of votes obtained for the latest attribute on the attribute type to 0, and if the number of votes obtained for the candidate attribute is greater than 0, setting the attribute type ajIs given to uiAnd go to 2.3.3.2-2.3.3.3, otherwise, the next round of voting is performed.
Step 2.3.3.2 updates the persona uiAt ajThe number of votes for the latest attribute.
Step 2.3.3.3 human entity uiThe update flag bit for this round is set to 1.
Step 2.3.4 if not the first iteration, go to steps 2.3.4.1-2.3.4.3
Step 2.3.4.1 if attribute type ajCandidate attribute value greater than number of votes of person entity uiAt ajThe latest attribute vote number, the attribute type ajIs given to uiAnd go to steps 2.3.4.2-2.3.4.3, otherwise, the next round of voting is performed.
Step 2.3.4.2 persona uiThe update flag bit of this round is set to 1
Step 2.3.4.3 updates the persona uiIn attribute type ajNumber of votes to be paid for the latest attribute
Step 2.3.5 repeat steps 2.3.2-2.3.4 until all u's are traversediAll attributes
Step 2.3.6 repeats steps 2.3.1-2.3.5 until all of the persona entities u are traversedi。
And 3, iteration, namely repeating the steps until the attribute value of each user in the social network of a certain iteration is not updated.
Based on the above method, an embodiment of the present invention further provides a device for obtaining attributes of a person knowledge graph based on a meta-social relationship ring, as shown in fig. 3, including: a building module and an updating module, wherein,
the system comprises a construction module, a database module and a database module, wherein the construction module is used for traversing users in the social network and constructing a figure knowledge graph, the figure knowledge graph comprises a figure entity set, a semantic edge set between figure entities, a relation set corresponding to the semantic edge and a mapping function from the semantic edge to the relation set, and each figure entity is set with a known attribute;
and the updating module is used for presuming unknown attributes of the users based on the meta-social relationship circles of the users and iteratively updating the attribute values in the figure knowledge graph until each user attribute value in the social network is not updated.
Referring to fig. 4, an exemplary simple people knowledge graph is shown, wherein the three, four and five of the people entities have respective attribute types. In order to verify the effectiveness and stability of the figure knowledge graph attribute acquisition scheme in the embodiment of the invention, further explanation is made through a specific data experiment:
the data set adopted in the experiment is derived from a social graph used for mining a social relationship circle by Julian McAuley, and totally comprises 10 social graphs, wherein attributes of the human entities mainly comprise birthday, class, conservation, degree, school, year, producer, hydrometow, location, employee and the like, and the relationships among the human entities comprise colleague relationships, alumni relationships, hometown relationships and the like. The effectiveness of the verification method for the three attributes of the person entity, namely the location, school of graduation and hometown, is selected in the experiment and represents the location, school of graduation and hometown of the person entity respectively, and the three attributes are main attributes of the person entity for forming a social circle. The social graph contains 4177 personal belonging entities in total, involving 463 school attribute values, 132 location attribute values, and 90 methodown attribute values.
In the experiment, the character entity attributes in a certain proportion in the data set are randomly marked to serve as a training set, the rest character entity attributes serve as a test set, the character entity attributes in the test set are reasoned by the method, and the effectiveness of the method is evaluated according to the accuracy of the result. The accuracy is defined as the ratio of the presumed correct character entity to all the character entities to be tested in all the social graphs; and regarding the attributes with multi-value condition, if the inferred attribute value is consistent with one of the attributes, the inference is correct.
Two sets of attribute reasoning experiments were performed: 1) in order to fully verify the effectiveness and stability of the method, attributes of the human entities in a certain proportion are marked, 5 groups of test sets are generated by each social graph through random marking, and the average value and the variance of the accuracy of the inference results of the 5 groups of test sets are calculated; 2) and marking test sets with different proportions, and verifying the reasoning accuracy of the method under different marking proportions. The experimental parameter selection and experimental result analysis are as follows:
and in the first group of experiments, the marking proportion is set to be 20%, each social graph generates 5 groups of test sets, and the variance and the average value of the experiment accuracy of the 5 groups of test sets of the social graphs are counted. The results are shown in table 1:
TABLE 1 Attribute inference accuracy and variance
Attribute type school location hometown
Mean value of accuracy 87％ 76％ 81％
Variance (variance) 2E-03 4E-04 3E-04
As can be seen from Table 1, the method can achieve better accuracy and stability.
In the second set of experiments, test sets with label ratios of 0.2, 0.4, 0.6, and 0.8 were constructed, and the results of the test methods based on the estimation accuracy under different label ratios are shown in table 2 below:
TABLE 2 method accuracy at different mark ratios
Mark ratio 0.2 0.4 0.6 0.8
school 87％ 90％ 92％ 94％
location 76％ 82％ 87％ 92％
hometown 81％ 92％ 93％ 96％
As can be seen from the experimental results in table 2, the higher the labeling ratio, the higher the accuracy of the method, that is, the more complete the preliminarily constructed person knowledge graph, the more correctly the missing attribute of the person entity can be inferred.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.
The elements of the various examples and method steps described in connection with the embodiments disclosed herein may be embodied in electronic hardware, computer software, or combinations of both, and the components and steps of the examples have been described in a functional generic sense in the foregoing description for clarity of hardware and software interchangeability. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
Those skilled in the art will appreciate that all or part of the steps of the above methods may be implemented by instructing the relevant hardware through a program, which may be stored in a computer-readable storage medium, such as: read-only memory, magnetic or optical disk, and the like. Alternatively, all or part of the steps of the foregoing embodiments may also be implemented by using one or more integrated circuits, and accordingly, each module/unit in the foregoing embodiments may be implemented in the form of hardware, and may also be implemented in the form of a software functional module. The present invention is not limited to any specific form of combination of hardware and software.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (8)

1. A character knowledge graph attribute obtaining method based on a meta-social relationship circle is characterized by comprising the following contents:
A) traversing users in a social network, and constructing a figure knowledge graph, wherein the figure knowledge graph comprises a figure entity set, a semantic edge set among the figure entities, a relation set corresponding to the semantic edges and a mapping function from the semantic edges to the relation set, and each figure entity is set with a known attribute;
B) presume the unknown attribute of users and iterate and upgrade the attribute value in the figure knowledge map on the basis of the meta social relationship circle of users, until every user attribute value in the social network is not upgraded again;
B) in the method, based on the attribute characteristics corresponding to the social relationship shared among the character entities with the same social relationship, the unknown attribute of the character entity in the character knowledge graph is presumed by using a majority iterative voting mechanism;
and in the iteration process, recording the number of votes obtained by the new attribute value when the attribute value of the character entity is updated every time, taking the attribute value with the largest number of votes obtained as a candidate attribute value for the attribute type after the voting of each round is finished, and updating the attribute value of the character entity when the number of votes obtained by the candidate attribute value is larger than the number of votes obtained by the attribute value in the previous round of updating.
2. The method of claim 1, wherein in a), the person knowledge graph is represented as KG { V, E, R, F }, where V represents a set of person entities, E represents a set of semantic edges between the person entities, R represents a set of relationships corresponding to the semantic edges, and F represents a mapping function from the semantic edges to R.
3. The method for obtaining attributes of a person knowledge graph based on meta-social relationship circles as claimed in claim 1, wherein in B), the meta-social relationship circles of the users are obtained according to a mapping function between person entities in the person knowledge graph.
4. The method of claim 3, wherein the set of meta-social relationship circles associated with the human entity is determined by traversing all neighboring nodes of the human entity with respect to the human entity according to the human knowledge graph.
5. The method of claim 1, wherein in attribute value voting, a meta social relationship circle of a character entity is traversed, attribute block data of neighboring nodes in the meta social relationship circle is obtained, and attribute values of corresponding attribute types are voted according to whether attribute types and attribute values of the neighboring node attribute block data are the same.
6. The method for obtaining attributes of a person knowledge graph based on a meta-social relationship circle as claimed in claim 1 or 5, wherein in the updating of the attribute values by the majority iterative voting mechanism, an updating flag bit of a person entity is set, the attribute type of the current person entity vote-obtaining record is traversed for each person entity in the meta-social relationship circle, the attribute value of the current person entity is updated according to the number of votes obtained from the candidate attribute values of the attribute type, and the updating flag bit of the person entity is recorded until all the attribute types of the current person entity are traversed.
7. The method for obtaining attributes of a person knowledge graph based on a meta-social relationship ring according to claim 1 or 5, wherein in the updating of the attribute values by a majority iteration voting mechanism, a person entity updating flag is set and initialized, the attribute type of the current person entity vote record is traversed for each person entity in the meta-social relationship ring, if the first iteration is performed, the vote count of the latest attribute of the person entity on the attribute type is set to 0, and if the vote count of the candidate attribute is greater than 0, the candidate attribute value is given to the person entity, the vote count of the latest attribute of the person entity on the attribute type is updated, and a person entity updating flag is set to 1; if the iteration is not the first iteration, when the number of the candidate attribute values of the attribute types is larger than the number of votes obtained by the character entity in the latest attribute of the attribute types, the candidate attribute values of the attribute types are given to the character entity, a character entity updating zone bit 1 is set, and the number of votes obtained by the character entity in the latest attribute of the attribute types is updated until all attributes of the current character entity are traversed.
8. A character knowledge graph attribute acquisition device based on meta-social relationship circles, which is realized based on the method of claim 1 and comprises: a building module and an updating module, wherein,
the system comprises a construction module, a database module and a database module, wherein the construction module is used for traversing users in the social network and constructing a figure knowledge graph, the figure knowledge graph comprises a figure entity set, a semantic edge set between figure entities, a relation set corresponding to the semantic edge and a mapping function from the semantic edge to the relation set, and each figure entity is set with a known attribute;
and the updating module is used for presuming unknown attributes of the users based on the meta-social relationship circles of the users and iteratively updating the attribute values in the figure knowledge graph until each user attribute value in the social network is not updated.
CN201910266511.0A 2019-04-03 2019-04-03 Character knowledge graph attribute acquisition method and device based on meta-social relationship circle Active CN110110155B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910266511.0A CN110110155B (en) 2019-04-03 2019-04-03 Character knowledge graph attribute acquisition method and device based on meta-social relationship circle

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910266511.0A CN110110155B (en) 2019-04-03 2019-04-03 Character knowledge graph attribute acquisition method and device based on meta-social relationship circle

Publications (2)

Publication Number Publication Date
CN110110155A CN110110155A (en) 2019-08-09
CN110110155B true CN110110155B (en) 2021-01-26

Family

ID=67485140

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910266511.0A Active CN110110155B (en) 2019-04-03 2019-04-03 Character knowledge graph attribute acquisition method and device based on meta-social relationship circle

Country Status (1)

Country Link
CN (1) CN110110155B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110598124B (en) * 2019-08-14 2023-05-26 平安科技（深圳）有限公司 Numerical attribute mining method, device, computer equipment and storage medium
CN111026816B (en) * 2019-12-09 2022-09-02 重庆锐云科技有限公司 High-net-value customer group identification method and device based on knowledge graph and storage medium
CN111667840A (en) * 2020-06-11 2020-09-15 刘娴 Robot knowledge graph node updating method based on voiceprint recognition

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107016072A (en) * 2017-03-23 2017-08-04 成都市公安科学技术研究所 Knowledge-based inference system and method based on social networks knowledge mapping
CN107943874A (en) * 2017-11-13 2018-04-20 平安科技（深圳）有限公司 Knowledge mapping processing method, device, computer equipment and storage medium

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10311050B2 (en) * 2017-01-23 2019-06-04 International Business Machines Corporation Crowdsourced discovery of paths in a knowledge graph

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107016072A (en) * 2017-03-23 2017-08-04 成都市公安科学技术研究所 Knowledge-based inference system and method based on social networks knowledge mapping
CN107943874A (en) * 2017-11-13 2018-04-20 平安科技（深圳）有限公司 Knowledge mapping processing method, device, computer equipment and storage medium

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
一种社会网络用户特征身份识别方法;胡开先等;《计算机研究与发展》;20161130;2636-2637页 *

Also Published As

Publication number Publication date
CN110110155A (en) 2019-08-09

Similar Documents

Publication Publication Date Title
WO2017211051A1 (en) Mining method and server for social network account of target subject, and storage medium
CN110110155B (en) Character knowledge graph attribute acquisition method and device based on meta-social relationship circle
Irving et al. The stable marriage problem with master preference lists
CN110597870A (en) Enterprise relation mining method
CN107423820B (en) Knowledge graph representation learning method combined with entity hierarchy categories
CN110727852A (en) Method, device and terminal for pushing recruitment recommendation service
CN106789338B (en) Method for discovering key people in dynamic large-scale social network
Hasan et al. Graphettes: Constant-time determination of graphlet and orbit identity including (possibly disconnected) graphlets up to size 8
CN104572915B (en) One kind is based on the enhanced customer incident relatedness computation method of content environment
CN104615740B (en) A kind of volunteer's geography information credit worthiness computational methods
CN109885797B (en) Relational network construction method based on multi-identity space mapping
CN116415199B (en) Business data outlier analysis method based on audit intermediate table
CN105138636A (en) Graph construction method and device for entity relationship
DE112020002372T5 (en) ESTIMATION OF A REALIZATION OF FEATURE VECTORS
CN110489660A (en) A kind of user's economic situation portrait method of social media public data
Rudas et al. Pathways between social science and computational social science: theories, methods, and interpretations
US8548988B2 (en) System and method for comparing universities based on their university model graphs
CN111737466B (en) Method for quantizing interactive information of deep neural network
US7904431B1 (en) Method and system for automated request modelling
Bohner Decision-support systems for sustainable urban planning
CN107402984B (en) A kind of classification method and device based on theme
Kotseruba et al. Industry and Academic Research in Computer Vision
Lee et al. Lifted representation of relational causal models revisited: Implications for reasoning and structure learning
Belohlavek et al. Adding background knowledge to formal concept analysis via attribute dependency formulas
Wang et al. Research on the Youth Group's Expectations for the Future Development of self-Media while in the Digital Economy

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant