CN110162637A - Information Atlas construction method, device and equipment - Google Patents
Information Atlas construction method, device and equipment Download PDFInfo
- Publication number
- CN110162637A CN110162637A CN201910114989.1A CN201910114989A CN110162637A CN 110162637 A CN110162637 A CN 110162637A CN 201910114989 A CN201910114989 A CN 201910114989A CN 110162637 A CN110162637 A CN 110162637A
- Authority
- CN
- China
- Prior art keywords
- information
- node
- attribute values
- subgraph
- map element
- 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.)
- Granted
Links
Abstract
A kind of Information Atlas construction method, device and equipment are disclosed, the Information Atlas construction method includes: to extract map element corresponding with the input information from multiclass subgraph spectrum according to input information；And the extracted map element of split is to generate Information Atlas corresponding with the input information；Wherein, at least part subgraph in the multiclass subgraph spectrum composes information node having the same, and extracting map element corresponding with the input information from multiclass subgraph spectrum according to input information includes: to determine the node attribute values of target information node according to input information；And map element corresponding with identified node attribute values is extracted from every class subgraph spectrum., can be effectively based on input information by map element corresponding with input information in extraction and split multiclass subgraph spectrum, the subgraph that be associated with multiple and different classifications is composed, and the information of multiple classifications associated with input information is returned.
Description
Technical field
This disclosure relates to map construction field, relate more specifically to a kind of Information Atlas construction method, Information Atlas building
Device and Information Atlas construct equipment.
Background technique
As artificial intelligence is in civilian and commercial kitchen area extensive use, map construction is in business big data, intelligence
Play the role of becoming more and more important Deng during, therefore map construction, especially Information Atlas building are also faced with higher
It is required that.The building for the general information map that user searches in recommendation business is concentrated in Information Atlas building at present and is based on artificial
Arrange obtained knowledge base, for example, the people cube of Microsoft, Baidu it is intimate, have become function in intelligent answer, machine translation etc.
In scene.
However, currently constructing obtained Information Atlas is the other Information Atlas of unitary class, and phase between each Information Atlas
It is mutually independent, unified association is not formed between the Information Atlas of multiple classifications.When inputting ad hoc inquiry information, return the result
The other information of mostly single or unitary class, and it is unable to get all information relevant to ad hoc inquiry information.
Therefore, it is necessary to one kind based on input information realization Information Atlas building, especially realization Financial Information map
When building, the Information Atlas structure of multiple classification informations associated with input information can be returned to effectively based on input information
Construction method.
Summary of the invention
In view of the above problems, present disclose provides a kind of Information Atlas construction method, Information Atlas construction device and information
Map construction equipment.The Information Atlas construction method provided using the disclosure can be based on input information realization Information Atlas structure
On the basis of building, effectively based on input information, it is associated with the subgraph spectrum of multiple and different classifications, is returned associated with input information
The information of multiple classifications.
According to the one side of the disclosure, a kind of Information Atlas construction method is proposed, comprising: according to input information from multiclass
Map element corresponding with the input information is extracted in subgraph spectrum；And the extracted map element of split is to generate and the input
The corresponding Information Atlas of information；Wherein, at least part subgraph in the multiclass subgraph spectrum composes information node having the same,
And extracting map element corresponding with the input information from multiclass subgraph spectrum according to input information includes: to be believed according to input
Breath, determines the node attribute values of target information node；And it is extracted and identified node attribute values pair from every class subgraph spectrum
The map element answered.
In some embodiments, every class subgraph spectrum has at least two information nodes, and every class subgraph spectrum includes multiple
Map element；For each information node, each map element has the node attribute values of the information node.
In some embodiments, the extracted map element of split is to generate Information Atlas packet corresponding with the input information
Include: using each identified node attribute values as linking point, link from the multiple subgraph spectrum in extract with the nodal community
It is worth corresponding map element.
In some embodiments, the multiclass subgraph spectrum includes concept subgraph spectrum, company's subgraph spectrum, stock subgraph spectrum, throws
Provide at least two classes during subgraph spectrum, product subgraph spectrum, event subgraph are composed, industry subgraph is composed.
In some embodiments, the target information node comprises at least one of the following: Business Name, organization, people
Name, film name, name of product, stock code, stock name, concept name.
In some embodiments, the multiclass subgraph spectrum includes at least event subgraph and composes, and the input information is thing
Part information, wherein the node attribute values that target information node is determined according to input information include: the event letter for being inputted
Breath determines the node attribute values of target information node corresponding with the event information based on event subgraph spectrum.
In some embodiments, map member corresponding with the input information is extracted from multiclass subgraph spectrum according to input information
It further include constructing the multiclass subgraph to compose before element, wherein every a kind of subgraph spectrum includes: to be arranged at least in building multiclass subgraph spectrum
Two information nodes, and the corresponding relationship between setting information node；The node category of the information node is extracted from external data
Property value；Extracted node attribute values are associated, to form map element；Obtained map element is denoised, is obtained
Subgraph spectrum.
In some embodiments, at least two information node includes Business Name information node, and for gained
To the denoising of map element include: each node attribute values for Business Name information node, based on the public affairs pre-established
Simple full name corresponding relationship is taken charge of, the multiple node attribute values for meeting company's letter full name corresponding relationship are associated.
In some embodiments, at least two information node includes name information node, and for obtained
The denoising of map element includes: that will be divided with multiple attribute values of the associated Business Name information node of same name information node
It cuts, and the node attribute values of name information node is marked using the node attribute values of each Business Name information node
Know, to eliminate name node ambiguity.
According to another aspect of the present disclosure, a kind of Information Atlas construction device is provided, comprising: map element extraction mould
Block is configured as extracting map element corresponding with the input information from multiclass subgraph spectrum according to input information；Information Atlas
Generation module is configured as the extracted map element of split to generate Information Atlas corresponding with the input information.
In some embodiments, every class subgraph spectrum has at least two information nodes, and every class subgraph spectrum includes multiple
Map element；For each information node, each map element has the node attribute values of the information node；Wherein, map member
Plain extraction module includes: target information node determining module, is configured as determining the section of target information node according to input information
Point attribute value；And map element respective modules, it is configured as extracting and identified node attribute values from every class subgraph spectrum
Corresponding map element.
In some embodiments, Information Atlas generation module includes: map element link module, is configured as with each institute
Determining node attribute values are linking point, link the map corresponding with the node attribute values extracted from the multiple subgraph spectrum
Element.
According to another aspect of the present disclosure, a kind of Information Atlas building equipment is provided, wherein the equipment includes processing
Device and memory, the memory include one group of instruction, and one group of instruction makes the information when being executed by the processor
Map construction equipment executes operation, and the operation includes: to be extracted and the input information from multiclass subgraph spectrum according to input information
Corresponding map element；And the extracted map element of split is to generate Information Atlas corresponding with the input information.Wherein,
At least part subgraph in the multiclass subgraph spectrum composes information node having the same, and sub from multiclass according to input information
It includes: to determine the node of target information node according to input information that map element corresponding with the input information is extracted in map
Attribute value；And map element corresponding with identified node attribute values is extracted from every class subgraph spectrum.
In some embodiments, every class subgraph spectrum has at least two information nodes, and every class subgraph spectrum includes multiple
Map element；For each information node, each map element has the node attribute values of the information node.It is mentioned using the disclosure
The Information Atlas construction method of confession, can be on the basis of based on input information realization Information Atlas building, effectively based on defeated
Enter information, is associated with the subgraph spectrum of multiple and different classifications, returns to the information of multiple classifications associated with input information.
Detailed description of the invention
It, below will be to required use in embodiment description in order to illustrate more clearly of the technical solution of the embodiment of the present disclosure
Attached drawing be briefly described, it should be apparent that, the accompanying drawings in the following description is only some embodiments of the present disclosure, for this
For the those of ordinary skill of field, without making creative work, it can also be obtained according to these attached drawings other
Attached drawing.The following drawings is not drawn by actual size equal proportion scaling deliberately, it is preferred that emphasis is shows the purport of the disclosure.
Fig. 1 shows the illustrative flow chart of the Information Atlas construction method 100 according to the embodiment of the present disclosure；
Fig. 2 shows the partial schematic diagrams that 200 are composed according to company's subgraph of the embodiment of the present disclosure；
Fig. 3 shows extracting and the input information from multiclass subgraph spectrum according to input information according to the embodiment of the present disclosure
The flow chart of the illustrative methods 300 of corresponding map element；
Fig. 4 shows extracting and the input information from multiclass subgraph spectrum according to input information according to the embodiment of the present disclosure
The schematic diagram of corresponding map element 400；
Fig. 5 is shown to be extracted from the multiple subgraph spectrum using each identified node attribute values as linking point link
The schematic diagram of map element 500 corresponding with the node attribute values；
Fig. 6, which is shown, determines that target corresponding with the event information is believed based on event subgraph spectrum according to the embodiment of the present disclosure
Cease the schematic diagram of the node attribute values 600 of node；
Fig. 7 shows the exemplary process diagram of the method 700 of the building subgraph spectrum according to the embodiment of the present disclosure；
The exemplary process diagram for the method 800 that the company name that Fig. 8 shows the embodiment of the present disclosure disambiguates；
The exemplary process diagram for the method 900 that the name that Fig. 9 shows the embodiment of the present disclosure disambiguates；
Figure 10 shows the illustrative block diagram of the Information Atlas construction device 110 according to the embodiment of the present disclosure；
Figure 11 shows the illustrative block diagram that equipment 950 is constructed according to the Information Atlas of the embodiment of the present disclosure.
Specific embodiment
The technical solution in the embodiment of the present disclosure is clearly and completely described below in conjunction with attached drawing, it is clear that
Ground, described embodiment are only the section Example of the disclosure, instead of all the embodiments.Implemented based on the disclosure
Example, every other embodiment obtained by those of ordinary skill in the art without making creative efforts also belong to
The range of disclosure protection.
As shown in the application and claims, unless context clearly prompts exceptional situation, " one ", "one", " one
The words such as kind " and/or "the" not refer in particular to odd number, may also comprise plural number.It is, in general, that term " includes " only prompts to wrap with "comprising"
Include clearly identify the step of and element, and these steps and element do not constitute one it is exclusive enumerate, method or apparatus
The step of may also including other or element.
Although the application is made that various references to the certain module in system according to an embodiment of the present application, however,
Any amount of disparate modules can be used and be operated on user terminal and/or server.The module is only illustrative
, and disparate modules can be used in the different aspect of the system and method.
Flow chart used herein is used to illustrate operation performed by system according to an embodiment of the present application.It should
Understand, before or operation below not necessarily accurately carry out in sequence.On the contrary, as needed, it can be according to inverted order
Or various steps are handled simultaneously.It is also possible to during other operations are added to these, or it is a certain from the removal of these processes
Step or number step operation.
Fig. 1 shows the exemplary process diagram of the Information Atlas construction method 100 according to the embodiment of the present disclosure.
Firstly, in step s101, extracting figure corresponding with the input information from multiclass subgraph spectrum according to input information
Spectral element.After extracting multiple map elements corresponding with input information, further, in step s 102, split is extracted
Map element to generate Information Atlas corresponding with the input information.
The input information can be the query information that user directly inputs, or be also possible to computer system in response to
User inputs information or controls information and the information voluntarily inquired.The embodiment of the present disclosure not source by input information and its input
The limitation of mode.For example, can be the query information that user inputs in Webpage search column, or it be also possible to the inquiry of user
The information that information generates after computer pre-processes.
The multiclass subgraph spectrum refers to is under the jurisdiction of different classes of multiple subgraphs spectrum respectively.Wherein, the multiclass subgraph spectrum
In at least part subgraph compose information node having the same.In some embodiments, the multiclass subgraph spectrum may include general
Subgraph spectrum, company's subgraph spectrum, stock subgraph spectrum, investment subgraph is read to compose, in product subgraph spectrum, event subgraph spectrum, industry subgraph spectrum
At least two classes.
According to the embodiment of the present disclosure, every one kind subgraph music score can such as have at least two information nodes, and every a kind of son
Map includes multiple map elements, and for each information node, each map element has the node attribute values of the information node.
In some embodiments, in each map element, for the same information node, there can be one or more
The node attribute values of the information node.Node attribute values can be particular content, or be also possible to " default " or " sky ".This Shen
It is not limited by the number of the node attribute values of information node and particular content in map element please.
In some embodiments, for the node attribute values can further progress partition of the level, such as can be classified as
First nodes attribute value and two-level node attribute value.By taking the map element of company's subgraph spectrum as an example, wherein may include nodal community
It is worth " Baidu " and node attribute values " iqiyi.com ", and " iqiyi.com " be the subsidiary of " Baidu ", based on the corresponding relationship, then " hundred
Degree " is first nodes attribute value, and " iqiyi.com " is two-level node attribute value.The application is not by the node attribute values in map element
Different stage and its mutual corresponding relationship limitation.
Fig. 2 shows the partial schematic diagrams that 200 are composed according to company's subgraph of the embodiment of the present disclosure.
It is composed referring to company's subgraph shown in Fig. 2, above- mentioned information node and map element is more specifically described.
The said firm's subgraph spectrum includes three information nodes, respectively Business Name information node KC, name information node KHAnd industry name
Claim information node KS, further, for example including map element E in the said firm's subgraph spectrumC1And map element EC2, wherein map is first
Plain EC1It further comprise the node attribute values of multiple information nodes.Referring further to Figure 2, map element EC1Including Business Name
The node attribute values " iqiyi.com " of information node, " Baidu ", the node attribute values " Li Yanhong " of name information node, " Lu Qi ",
The node attribute values " internet " of film name information node.Map element EC2Nodal community including Business Name information node
It is worth " big boundary ", the node attribute values " Wang Tao " of name information node, the node attribute values " unmanned plane " of film name information node.
Further, based on the corresponding relationship between node attribute values, under Business Name information node, " Baidu " is first nodes category
Property value, " iqiyi.com " be two-level node attribute value.
Fig. 3 shows extracting and the input information from multiclass subgraph spectrum according to input information according to the embodiment of the present disclosure
The flow chart of the illustrative methods 300 of corresponding map element.
Firstly, according to input information, determining the node attribute values of target information node in step S301.
According to the embodiment of the present disclosure, target information node may be, for example, multiple information nodes from multiclass subgraph spectrum, or
Person can be multiple information nodes from certain a kind of subgraph spectrum, or an information node being also possible in certain a kind of map.
The embodiment of the present disclosure is not limited by the type and number of identified target information node.Such as it can be by target information node only
It is set as Business Name, or target information node can be set to Business Name, name, product Business Name three.
In some embodiments, the target information node comprises at least one of the following: Business Name, organization, people
Name, film name, name of product, stock code, stock name, concept name.
According to the embodiment of the present disclosure, the node attribute values of the target information node can be determined based on preset strategy, such as
When all or part of the content for inputting information is the node attribute values of target information node, then target information can be directly obtained
The attribute value of node；When the content in input information does not include the node attribute values of target information node or only includes part
Target information node node attribute values when, input information can be pre-processed, will treated data as mesh
Mark the node attribute values of information node.The embodiment of the present disclosure is not limited by the mode for the node attribute values for determining destination node.
Such as setting name is target information node, then node category of " Ma Yun " that can be inputted user directly as target information node
Property value；" Baidu takes Ali by the hand and opens up shared bicycle market " that can also be inputted to user, was waited based on Entity recognition, Relation extraction
Journey, extraction obtain node attribute values " Baidu ", " Ali " of Business Name information node, the node category of name of product information node
Property value " shared bicycle ", and corresponding name information is further obtained based on " Baidu ", " Ali " in company's subgraph spectrum
The node attribute values " Li Yanhong " of node, " Ma Yun ".
In addition, for the same target information node, the section of the target information node based on determined by input information
Point attribute value can be one or more, and the embodiment of the present disclosure is not by of the node attribute values of identified target information node
Several limitations.
According to input information, after the node attribute values for determining target information node, further, in step s 302,
Map element corresponding with identified node attribute values is extracted from every class subgraph spectrum.
In some embodiments, map corresponding with the identified node attribute values member extracted from every class subgraph spectrum
Element can be multiple, such as when there are name duplication of name, for the node attribute values " Ma Yun " of target information node, in company's subgraph
There may be multiple corresponding map elements in spectrum, it is under the jurisdiction of the nodal community with different Business Names respectively
Under value.Under some embodiments, in certain class subgraphs spectrum, it will be unable to extract the node attribute values pair with target information node
The map element answered, such as when the node attribute values of target information node are " quick worker App ", it may be simultaneously in stock subgraph spectrum
It does not include the map element of the node attribute values corresponding to the target information node.The embodiment of the present disclosure is not by extracted map
The number of element and its limitation in source.
Fig. 4 shows extracting and the input information from multiclass subgraph spectrum according to input information according to the embodiment of the present disclosure
The schematic diagram of corresponding map element.
Referring to Fig. 4, the above process can be described more specifically.Such as preset target information node is name and company name
Claim, then when the input information of user is " Lu Qi ", is composed by company's subgraph, search map element corresponding to " Lu Qi ", obtain
To map element EC1；Further, it can get map element EC1In corresponding to node attribute values " Lu Qi " Business Name believe
Cease node KCNode attribute values.It, can be by Business Name information node K according to the embodiment of the present disclosureCFirst nodes attribute value
As its node attribute values, the node attribute values of the Business Name information node corresponding to the input information " Lu Qi " can be obtained
" Baidu ".
Obtain the node attribute values " Lu Qi " of name information node and the node attribute values " hundred of Business Name information node
After degree ", further, extracted and identified node attribute values pair in each subgraph spectrum of the multiple subgraphs spectrum constructed
The map element answered.Such as there is currently company's subgraph spectrum, product subgraph spectrum and stock subgraph spectrum, then it can be believed based on Business Name
The node attribute values " Baidu " of node and the node attribute values " Lu Qi " of name information node are ceased, corresponding company's is extracted
Map element EC1, product subgraph spectral element EP1, stock subgraph spectral element EG1。
After extraction obtains multiple map elements, in step s 102, the extracted map element of split is defeated with this to generate
Enter the corresponding Information Atlas of information can include: using each identified node attribute values as linking point, link from the multiple son
The map element corresponding with the node attribute values extracted in map.
Fig. 5 is shown to be extracted from the multiple subgraph spectrum using each identified node attribute values as linking point link
The schematic diagram of map element corresponding with the node attribute values.
Referring to Fig. 4 and Fig. 5, the map split process can be more specifically described.Based on process shown in Fig. 4, mentioning
It, further, as shown in Figure 5, can be by the node attribute values of Business Name information node after obtaining multiple map elements
The node attribute values " Lu Qi " of " Baidu " and name information node are used as linking point, split company subgraph spectral element EC1, product
Map element EP1, stock subgraph spectral element EG1, generating has Business Name information node, name information node, stock information section
Point, product information node, film name information node Information Atlas.
By extracting the node attribute values of target information node, being effectively associated with multiple and different classes based on input information
Map element in other subgraph spectrum allows to generate the Information Atlas of the target information node with multiple classifications, thus
The information of multiple classifications associated with information is inputted can be returned.
In some embodiments, the multiclass subgraph spectrum includes at least event subgraph and composes, and the input information is thing
Part information, wherein the node attribute values that target information node is determined according to input information include: the event letter for being inputted
Breath determines the node attribute values of target information node corresponding with the event information based on event subgraph spectrum.
For example, the input information is specific event title or the description for a certain event, such as it can be only
Abbreviation including dependent event, or also may include the relevant personage of event, company, theme or concept name.The disclosure is not
It is limited by the particular content of the event information inputted.It can for example input " Wei Zexi event ", or can also input
" Ma Huateng takes Yang Zhenyu by the hand and initiates Science Explorations prize ".
It for example, can have at least two category information nodes in the event subgraph spectrum, such as can be entity class information section
Point and theme class information node.For example including event title, Business Name, name, stock etc., theme in entity class information node
Category information node is for example including subject name, concept name and semantic label etc..
According to the embodiment of the present disclosure, event subgraph spectrum includes multiple map elements, for each information node, each map
Element has the node attribute values of the information node.
In some embodiments, in the map element of event subgraph spectrum, for the same information node, can have one
The node attribute values of a or multiple information nodes.Wherein, which can be particular content, or be also possible to
" default " or " sky ".The application is not by the number of the node attribute values of information node in the map element of event subgraph spectrum and specifically
The limitation of content.
In some embodiments, for the node attribute values can further progress partition of the level, such as can be classified as
First nodes attribute value and two-level node attribute value.Such as the map element of event subgraph spectrum may include corresponding to subject
The node attribute values " mobile communication " and node attribute values " 5G communication " for claiming information node, wherein " 5G communication " is " mobile communication "
Sub-topics concept, be based on the corresponding relationship, then " mobile communication " be first nodes attribute value, " 5G communication " be two-level node category
Property value.The application is not limited by the different stage of the node attribute values in map element and its mutual corresponding relationship.
Fig. 6 shows the node category that target information node corresponding with the event information is determined based on event subgraph spectrum
The schematic diagram of property value.
Referring to Fig. 6, it is specifically described and target information section corresponding with the event information is determined based on event subgraph spectrum
The process of point.For the event information of " the Baidu's discussion bar Wei Zexi event " of user's input, as target information node, then
Can be in event subgraph spectrum, lookup and map element corresponding to the event " Baidu's discussion bar Wei Zexi event " obtain map
Element EV1；Further, dependent event " Baidu's termination and the conjunction of Putian hospital, system for obtaining the event can be composed by event subgraph
Make ", and obtaining its relevant company as " Baidu is online ", relevant theme is " medical tangle ".It is based on the event as a result,
Subgraph spectrum has determined the node attribute values of relevant to the event Business Name information node and subject information node.
In the node attribute values and subject information node that Business Name information node relevant to the incoming event has been determined
Node attribute values after, can continue the node attribute values and/or subject information of identified Business Name information node
Node attribute values of the node attribute values of node as target information node further search for corresponding figure in company's subgraph spectrum
Mark element.
In some embodiments, map member corresponding with the input information is extracted from multiclass subgraph spectrum according to input information
It further include the process for constructing the multiclass subgraph spectrum before element.
Fig. 7 shows the exemplary process diagram of the method 700 of the building subgraph spectrum according to the embodiment of the present disclosure.
According to the embodiment of the present disclosure, construct every a kind of subgraph spectrum in multiclass subgraph spectrum include: firstly, in step s 701,
The default corresponding relationship that at least two information nodes are set, and are arranged between the information node.Further, in step S702,
The node attribute values of the information node are extracted from external data.After the extraction for completing node attribute values, in step S703,
Extracted node attribute values are associated, to form map element.Further, in step S704, for obtained
The denoising of map element obtains subgraph spectrum.
According to the embodiment of the present disclosure, in map element, for the same information node, there can be one or more be somebody's turn to do
The node attribute values of information node.Node attribute values can be particular content, or be also possible to " default " or " sky ".The application
It is not limited by the number of the node attribute values of information node and particular content in map element.
In some embodiments, for the node attribute values can further progress partition of the level, such as can be classified as
First nodes attribute value and two-level node attribute value.By taking the map element of company's subgraph spectrum as an example, wherein may include nodal community
It is worth " Baidu " and node attribute values " iqiyi.com ", wherein " iqiyi.com " is the subsidiary of " Baidu ", is based on the corresponding relationship, then
" Baidu " is first nodes attribute value, and " iqiyi.com " is two-level node attribute value.The application is not by the node category in map element
The limitation of the different stage and its mutual corresponding relationship of property value.
According to the embodiment of the present disclosure, for the denoising of obtained map element for example can by calculate node different degree come
It realizes, by calculating different degree filtering noise node attribute values, saves the memory space of map, accelerate inquiry velocity；Another party
Face, when fuzzy query map, while when hitting multiple both candidate nodes, the preferential node for selecting importance high；Or it can lead to
Source number and each source itself the confidence level COMPREHENSIVE CALCULATING node that calculate node confidence level is crossed to realize, for example, by using node
Confidence level；Or can also be realized using calculated relationship confidence level, i.e., using the source number of relationship, source itself
Confidence level, relationship both sides node belief, the confidence level of egress and the didactic calculated relationship of ingress co-occurrence number.This public affairs
It opens and is not limited by map element denoising mode.
In the following, illustrating the process of building subgraph spectrum for constructing company's subgraph spectrum.Firstly, in building company
When map, as shown in step S701, such as Business Name, name, industry can choose as presupposed information node, and further
Its preset relation is set are as follows: for the attribute value of each company's information node, the respective attributes value of name information node is to supply
In the individual of the said firm, the respective attributes value of trade information node is the industry where the said firm for duty.Further, for company
Each different attribute value in name information node, can also be arranged the subordinate relation that it includes main company and subsidiary.
Further, in step S702, the node attribute values of the information node are extracted from external data.For example,
It can be by naming Entity recognition process to extract the nodal community of the information nodes such as name, Business Name from non-structured text
Value.
Further, in step S703, it can be based on Relation extraction process, obtained between extracted node attribute values
Relationship, the process can for example be obtained by the structuring list data on website, or can also be based on Entity recognition, according to
Predefined relationship classification, classifies for the relationship between node attribute values.Further, by extracted multiple nodes
The node attribute values for meeting corresponding relationship in attribute value are associated to form map element.Such as by the node of company's information node
Attribute value " Baidu ", with the personal name information node respective nodes attribute value " Li Yanhong " for taking service in the said firm, the said firm
The node respective attributes value " internet " of the trade information node of place industry is interrelated, obtains a map element.
Wherein the associated step of above-mentioned node attribute values, which can for example be used, connects it to realize by line segment, or
It can place it in corresponding list, the disclosure is not limited by the interrelated used form of node attribute values.Into one
Step ground can also be indicated mutual between node attribute values when being realized the association of node attribute values using line segment on line segment
Relationship.
After generating multiple map elements, in step S704, obtained map element is denoised, obtains subgraph spectrum.
The denoising process for example can calculate node different degree, node belief, relationship confidence level first, its three is added later
Weight average, to screen out the noisy unstable data of tool.
In some embodiments, the denoising of map element further includes that the node attribute values of information node are disambiguated, into one
Step may include that company name disambiguates and name disambiguates.
Company name is disambiguated, by each node attribute values for Business Name information node, based on building in advance
Vertical company's letter full name corresponding relationship, the multiple node attribute values for meeting company's letter full name corresponding relationship are associated.
The exemplary process diagram for the method 800 that the company name that Fig. 8 shows the embodiment of the present disclosure disambiguates.
Referring to Fig. 8, firstly, in step S801, for each nodal community in subgraph Pu Zhong company information node
Value extracts its company's full name and its corresponding multiple companies referred to as from external data.Further, in step S802,
Referred to as based on acquired company's full name and corresponding multiple companies, the full name and abbreviation corresponding relationship of each company are obtained, it is raw
Cheng company letter full name subgraph spectrum；Finally, in step S803, it is corresponding based on simple full name obtained in company's letter full name subgraph spectrum
The multiple node attribute values for meeting wherein full name and abbreviation corresponding relationship are associated by relationship.
In some embodiments, after company's letter full name subgraph spectrum is obtained in step S802, step S803 can not also be used
It realizes that Business Name is entirely referred to as associated, but is composed based on the said firm's letter full name subgraph, by each company's full name therein and right
The input terminal for the multiple abbreviations input neural network algorithm answered, generates company's letter full name relational model by neural network algorithm.
Thereafter, the full name that designated company is inputted by the input terminal in company's letter full name relational model, can be in company's letter full name relationship mould
The output end of type obtains its multiple corresponding abbreviation, also obtains the simple full name relationship of designated company accordingly.Further, based on finger
Determine the simple full name relationship of company, the multiple node attribute values for meeting wherein full name and abbreviation corresponding relationship can be associated.
Name is disambiguated, by by multiple attributes with the associated Business Name information node of same name information node
Value is split, and using the node attribute values of each Business Name information node to the node attribute values of name information node
It is identified, so as to eliminate name node ambiguity.
The exemplary process diagram for the method that the name that Fig. 9 shows the embodiment of the present disclosure disambiguates.
Referring to Fig. 9, firstly, in step S901, for the node attribute values of the name information node in map element, with
Point divides the map element centered on the node attribute values, obtains multiple map daughter elements independent of each other.Further
It, for each daughter element in the multiple map daughter element independent of each other, extracts and believes with the name in step S902 in ground
Cease the multiple Business Name information nodes and its node attribute values that the node attribute values of node are directly linked.Finally, in step
In S903, for each Business Name information node, using the node attribute values of the said firm's information node come to name information section
The node attribute values of point are identified.
Wherein, in the point centered on the node attribute values of name information node, when being divided to map element, such as can
Thus to obtain multiple map independent of each other by the attribute value for removing the name information node in current map element
Element.
In some embodiments, it is extracted in each map daughter element in multiple map daughter elements independent of each other
The multiple Business Name information nodes and its node attribute values being directly linked with the node attribute values of the name information node can wrap
It includes: when there are multiple Business Name information sections being directly linked with the name information node attribute value in a map daughter element
When the node attribute values of point, the node attribute values of multiple Business Name information nodes and the node of the name information node can be calculated
The relationship confidence level of attribute value；And extract the nodal community wherein with the Business Name information node of maximum relationship confidence level
Value.
By the way that name is carried out corresponding association with Business Name, the method to identify name can avoid passing through name letter
When ceasing node as target information node, due to the case where there are personage's duplications of name, and lead to wrong close occur when subgraph spectrum association
Connection, or introduce the subgraph spectral element of excessive noise and redundancy.
Figure 10 shows the exemplary block diagram of the Information Atlas construction device 110 according to the embodiment of the present disclosure.
Information Atlas construction device 110 as shown in Figure 10 includes: that map element extraction module 120 and Information Atlas generate
Module 130.
Wherein, the map element extraction module 120 be configured as according to input information from multiclass subgraph spectrum in extract with
The corresponding map element of the input information.
The Information Atlas generation module 130 is configured as the extracted map element of split to generate and the input information
Corresponding Information Atlas.
The input information can be the query information that user directly inputs, or be also possible to computer system in response to
User inputs information or controls information and the information voluntarily inquired.The embodiment of the present disclosure not source by input information and its input
The limitation of mode.For example, can be the query information that user inputs in Webpage search column, or it be also possible to the inquiry of user
The information that information generates after computer pre-processes.
The multiclass subgraph spectrum refers to is under the jurisdiction of different classes of multiple subgraphs spectrum respectively.Wherein, the multiclass subgraph spectrum
In at least part subgraph compose information node having the same.In some embodiments, the multiclass subgraph spectrum may include general
Subgraph spectrum, company's subgraph spectrum, stock subgraph spectrum, investment subgraph is read to compose, in product subgraph spectrum, event subgraph spectrum, industry subgraph spectrum
At least two classes.
In some embodiments, in map element, for the same information node, there can be one or more letters
Cease the node attribute values of node.Node attribute values can be particular content, or be also possible to " default " or " sky ".The application is not
It is limited by the number of the node attribute values of information node and particular content in map element.
In some embodiments, for the node attribute values can further progress partition of the level, such as can be classified as
First nodes attribute value and two-level node attribute value.By taking the map element of company's subgraph spectrum as an example, wherein may include nodal community
It is worth " Baidu " and node attribute values " iqiyi.com ", wherein " iqiyi.com " is the subsidiary of " Baidu ", is based on the corresponding relationship, then
" Baidu " is first nodes attribute value, and " iqiyi.com " is two-level node attribute value.The application is not by the node category in map element
The limitation of the different stage and its mutual corresponding relationship of property value.
Fig. 2 shows the partial schematic diagrams that 200 are composed according to company's subgraph of the embodiment of the present disclosure.
It is composed referring to company's subgraph shown in Fig. 2, above- mentioned information node and map element is more specifically described.
The said firm's subgraph spectrum includes three information nodes, respectively Business Name information node KC, name information node KHAnd industry name
Claim information node KS.Further, for example including map element E in the said firm's subgraph spectrumC1And map element EC2, wherein map is first
Plain EC1It further comprise the node attribute values of multiple information nodes.Referring further to Figure 2, map element EC1Including Business Name
The node attribute values " iqiyi.com " of information node, " Baidu ", the node attribute values " Li Yanhong " of name information node, " Lu Qi ",
The node attribute values " internet " of film name information node.Map element EC2Nodal community including Business Name information node
It is worth " big boundary ", the node attribute values " Wang Tao " of name information node, the node attribute values " unmanned plane " of film name information node.
Further, based on the corresponding relationship between node attribute values, node attribute values " Baidu " are first nodes attribute value, and " love is odd
Skill " is two-level node attribute value.
Wherein, in map element extraction module 120, process as shown in Figure 3 can be executed, according to input information from more
Map element corresponding with the input information is extracted in class subgraph spectrum.It is further can include: target information node determining module
121 and map element respective modules 122.
The target information node determining module 121 is configured as executing the operation as shown in step S301 in Fig. 3, according to
Information is inputted, determines the node attribute values of target information node.
According to the embodiment of the present disclosure, target information node may be, for example, multiple information nodes from multiclass subgraph spectrum, or
Person can be multiple information nodes from certain a kind of subgraph spectrum, or an information node being also possible in certain a kind of map.
The embodiment of the present disclosure is not limited by the type and number of identified target information node.Such as it can be by target information node only
It is set as Business Name, or target information node can be set to Business Name, name, product Business Name three.
In some embodiments, the target information node comprises at least one of the following: Business Name, organization, people
Name, film name, name of product, stock code, stock name, concept name.
Wherein, the node attribute values of the target information node can be determined based on preset strategy, such as when input information
When all or part of the content is the node attribute values of target information node, then the attribute of target information node can be directly obtained
Value；When input information in content do not include target information node node attribute values or only include part target information
The node attribute values of node can pre-process input information, using treated data as target information node
Node attribute values.The embodiment of the present disclosure is not limited by the mode for the node attribute values for determining destination node.Such as setting name
For target information node, then node attribute values of " Ma Yun " that user can be inputted directly as target information node；It can also be right
" Baidu takes Ali by the hand and opens up shared bicycle market " of user's input, is obtained based on processes, extractions such as Entity recognition, Relation extractions
The node attribute values of the node attribute values " Baidu " of Business Name information node, " Ali ", name of product information node are " shared single
Vehicle ", and the node category of corresponding name information node is further obtained based on " Baidu ", " Ali " in company's subgraph spectrum
Property value " Li Yanhong ", " Ma Yun ".
In addition, for the same target information node, the section of the target information node based on determined by input information
Point attribute value can be one or more, and the embodiment of the present disclosure is not by of the node attribute values of identified target information node
Several limitations.
The map element respective modules 122 are configured as executing the operation as shown in step S302 in Fig. 3, from every class
Map element corresponding with identified node attribute values is extracted in map.
In some embodiments, map corresponding with the identified node attribute values member extracted from every class subgraph spectrum
Element can be multiple, such as when there are name duplication of name, for the node attribute values " Ma Yun " of target information node, in company's subgraph
There may be multiple corresponding map elements in spectrum, it is under the jurisdiction of the nodal community with different Business Names respectively
Under value.Under some embodiments, in certain class subgraphs spectrum, it will be unable to extract the node attribute values pair with target information node
The map element answered, such as when the node attribute values of target information node are " quick worker App ", it may be simultaneously in stock subgraph spectrum
It does not include the map element of the node attribute values corresponding to the target information node.The embodiment of the present disclosure is not by extracted map
The number of element and its limitation in source.
In some embodiments, further, the Information Atlas generation module 130 may include map element link module
131, in the map element link module 131, process as shown in Figure 5 can be executed, with each identified nodal community
Value is linking point, links the map element corresponding with the node attribute values extracted from the multiple subgraph spectrum.
Referring to Fig. 4 and Fig. 5, the map split process can be more specifically described.Based on process shown in Fig. 4, mentioning
It, further, as shown in Figure 5, can be by the node attribute values of Business Name information node after obtaining multiple map elements
The node attribute values " Lu Qi " of " Baidu " and name information node are used as linking point, split company subgraph spectral element EC1, product
Map element EP1, stock subgraph spectral element EG1, generating has Business Name information node, name information node, stock information section
The Information Atlas of point, name of product information node.
In some embodiments, the multiclass subgraph spectrum includes at least event subgraph and composes, and the input information is thing
Part information.Wherein, target information node determining module 121 may include event information respective modules 121', corresponding in event information
In module 121', it is corresponding with the event information that determination can be composed based on the event subgraph for the event information inputted
The node attribute values of target information node.
For example, the input information is specific event title or the description for a certain event, such as it can be only
Abbreviation including dependent event, or also may include the relevant personage of event, company, theme or concept name.The disclosure is not
It is limited by the particular content of the event information inputted.It can for example input " Wei Zexi event ", or can also input
" Ma Huateng takes Yang Zhenyu by the hand and initiates Science Explorations prize ".
It for example, can have at least two category information nodes in the event subgraph spectrum, such as can be entity class information section
Point and theme class information node.For example including event title, Business Name, name, stock etc., theme in entity class information node
Category information node is for example including subject name, concept name and semantic label etc..
According to the embodiment of the present disclosure, event subgraph spectrum includes multiple map elements, for each information node, each map
Element has the node attribute values of the information node.
In some embodiments, in the map element of event subgraph spectrum, for the same information node, can have one
The node attribute values of a or multiple information nodes.Wherein, which can be particular content, or be also possible to
" default " or " sky ".The application is not by the number of the node attribute values of information node in the map element of event subgraph spectrum and specifically
The limitation of content.
In some embodiments, for the node attribute values can further progress partition of the level, such as can be classified as
First nodes attribute value and two-level node attribute value.Such as the map element of event subgraph spectrum may include corresponding to subject
The node attribute values " mobile communication " and node attribute values " 5G communication " for claiming information node, wherein " 5G communication " is " mobile communication "
Sub-topics concept, be based on the corresponding relationship, then " mobile communication " be first nodes attribute value, " 5G communication " be two-level node category
Property value.The application is not limited by the different stage of the node attribute values in map element and its mutual corresponding relationship.
Figure 11 shows the illustrative block diagram that equipment 950 is constructed according to the Information Atlas of the embodiment of the present disclosure.
Information Atlas building equipment 950 as shown in figure 11 can be implemented as one or more dedicated or general computers
System module or component, such as PC, laptop, tablet computer, mobile phone, personal digital assistant (personal
Digital assistance, PDA) and any intelligent and portable equipment.Wherein, Information Atlas building equipment 950 may include to
A few processor 960 and memory 970.
Wherein, at least one described processor is for executing program instructions.The memory 970 is set in Information Atlas building
In standby 950 can program storage unit in different forms and data storage element exist, such as hard disk, read-only memory
(ROM), random access memory (RAM), it can be used to during storage processor processing and/or execution information map construction
Possible program instruction performed by the various data files and processor used.Although being not shown, hum pattern
Spectrum building equipment 950 can also include an input output assembly, support Information Atlas building equipment 950 and other assemblies (such as
Image capture device 980) between input/output data stream.Information Atlas construct equipment 950 can also by communication port from
Network sends and receives information and data.
In some embodiments, one group of instruction that the memory 970 is stored by the processor 960 execute when,
The Information Atlas building equipment 950 is set to execute operation, the operation includes: to be extracted from multiclass subgraph spectrum according to input information
Map element corresponding with the input information；And the extracted map element of split is to generate letter corresponding with the input information
Cease map.
The input information can be the query information that user directly inputs, or be also possible to computer system in response to
User inputs information or controls information and the information voluntarily inquired.The embodiment of the present disclosure not source by input information and its input
The limitation of mode.For example, can be the query information that user inputs in Webpage search column, or it be also possible to the inquiry of user
The information that information generates after computer pre-processes.
The multiclass subgraph spectrum refers to is under the jurisdiction of different classes of multiple subgraphs spectrum respectively.Wherein, the multiclass subgraph spectrum
In at least part subgraph compose information node having the same.In some embodiments, the multiclass subgraph spectrum may include general
Subgraph spectrum, company's subgraph spectrum, stock subgraph spectrum, investment subgraph is read to compose, in product subgraph spectrum, event subgraph spectrum, industry subgraph spectrum
At least two classes.
According to the embodiment of the present disclosure, every one kind subgraph music score can such as have at least two information nodes, and every a kind of son
Map includes multiple map elements, and for each information node, each map element has the node attribute values of the information node.
In some embodiments, in map element, for the same information node, there can be one or more letters
Cease the node attribute values of node.Node attribute values can be particular content, or be also possible to " default " or " sky ".The application is not
It is limited by the number of the node attribute values of information node and particular content in map element.
In some embodiments, for the node attribute values can further progress partition of the level, such as can be classified as
First nodes attribute value and two-level node attribute value.By taking the map element of company's subgraph spectrum as an example, wherein may include nodal community
It is worth " Baidu " and node attribute values " iqiyi.com ", wherein " iqiyi.com " is the subsidiary of " Baidu ", is based on the corresponding relationship, then
" Baidu " is first nodes attribute value, and " iqiyi.com " is two-level node attribute value.The application is not by the node category in map element
The limitation of the different stage and its mutual corresponding relationship of property value.
In some embodiments, map member corresponding with the input information is extracted from multiclass subgraph spectrum according to input information
Element includes: to determine the node attribute values of target information node according to input information；And it is extracted from every class subgraph spectrum with institute really
The corresponding map element of fixed node attribute values.
According to the embodiment of the present disclosure, target information node may be, for example, multiple information nodes from multiclass subgraph spectrum, or
Person can be multiple information nodes from certain a kind of subgraph spectrum, or an information node being also possible in certain a kind of map.
The embodiment of the present disclosure is not limited by the type and number of identified target information node.Such as it can be by target information node only
It is set as Business Name, or target information node can be set to Business Name, name, product Business Name three.
In some embodiments, the target information node comprises at least one of the following: Business Name, organization, people
Name, film name, name of product, stock code, stock name, concept name.
In addition, for the same target information node, the section of the target information node based on determined by input information
Point attribute value can be one or more, and the embodiment of the present disclosure is not by of the node attribute values of identified target information node
Several limitations.
In some embodiments, map corresponding with the identified node attribute values member extracted from every class subgraph spectrum
Element can be multiple, such as when there are name duplication of name, for the node attribute values " Ma Yun " of target information node, in company's subgraph
There may be multiple corresponding map elements in spectrum, it is under the jurisdiction of the nodal community with different Business Names respectively
Under value.Under some embodiments, in certain class subgraphs spectrum, it will be unable to extract the node attribute values pair with target information node
The map element answered, such as when the node attribute values of target information node are " quick worker App ", it may be simultaneously in stock subgraph spectrum
It does not include the map element of the node attribute values corresponding to the target information node.The embodiment of the present disclosure is not by extracted map
The number of element and its limitation in source.
In some embodiments, the multiclass subgraph spectrum includes at least event subgraph and composes, and the input information is thing
Part information, wherein the node attribute values that target information node is determined according to input information include: the event letter for being inputted
Breath determines the node attribute values of target information node corresponding with the event information based on event subgraph spectrum.
For example, the input information is specific event title or the description for a certain event, such as it can be only
Abbreviation including dependent event, or also may include the relevant personage of event, company, theme or concept name.The disclosure is not
It is limited by the particular content of the event information inputted.It can for example input " Wei Zexi event ", or can also input
" Ma Huateng takes Yang Zhenyu by the hand and initiates Science Explorations prize ".
It for example, can have at least two category information nodes in the event subgraph spectrum, such as can be entity class information section
Point and theme class information node.For example including event title, Business Name, name, stock etc., theme in entity class information node
Category information node is for example including subject name, concept name and semantic label etc..
According to the embodiment of the present disclosure, event subgraph spectrum includes multiple map elements, for each information node, each map
Element has the node attribute values of the information node.
In some embodiments, in the map element of event subgraph spectrum, for the same information node, can have one
The node attribute values of a or multiple information nodes.Wherein, which can be particular content, or be also possible to
" default " or " sky ".The application is not by the number of the node attribute values of information node in the map element of event subgraph spectrum and specifically
The limitation of content.
In some embodiments, for the node attribute values can further progress partition of the level, such as can be classified as
First nodes attribute value and two-level node attribute value.Such as the map element of event subgraph spectrum may include corresponding to subject
The node attribute values " mobile communication " and node attribute values " 5G communication " for claiming information node, wherein " 5G communication " is " mobile communication "
Sub-topics concept, be based on the corresponding relationship, then " mobile communication " be first nodes attribute value, " 5G communication " be two-level node category
Property value.The application is not limited by the different stage of the node attribute values in map element and its mutual corresponding relationship.
In some embodiments, Information Atlas building equipment 950 can receive outside Information Atlas building equipment 950
The input information that the input equipment in portion is inputted executes above-described Information Atlas construction method, reality based on the input information
The function of existing above-described Information Atlas construction device.
In some embodiments, map member corresponding with the input information is extracted from multiclass subgraph spectrum according to input information
It further include the process for constructing the multiclass subgraph spectrum before element.
Although processor 960, memory 970 are rendered as individual module in Figure 10, those skilled in the art can be managed
Solution, above equipment module may be implemented as individual hardware device, can also be integrated into one or more hardware devices.Only
It can be realized the principle of disclosure description, the specific implementation of different hardware devices should not be used as limitation disclosure protection
The factor of range.
The application has used particular words to describe embodiments herein.Such as " first/second embodiment ", " one implements
Example ", and/or " some embodiments " mean a certain feature relevant at least one embodiment of the application, structure or feature.Cause
This, it should be highlighted that and it is noted that " embodiment " or " an implementation referred to twice or repeatedly in this specification in different location
Example " or " alternate embodiment " are not necessarily meant to refer to the same embodiment.In addition, in one or more embodiments of the application
Certain features, structure or feature can carry out combination appropriate.
In addition, it will be understood by those skilled in the art that the various aspects of the application can be by several with patentability
Type or situation are illustrated and described, the combination or right including any new and useful process, machine, product or substance
Their any new and useful improvement.Correspondingly, the various aspects of the application can completely by hardware execute, can be complete
It is executed, can also be executed by combination of hardware by software (including firmware, resident software, microcode etc.).Hardware above is soft
Part is referred to alternatively as " data block ", " module ", " engine ", " unit ", " component " or " system ".In addition, the various aspects of the application
The computer product being located in one or more computer-readable mediums may be shown as, which includes computer-readable program
Coding.
Unless otherwise defined, all terms (including technical and scientific term) used herein have leads with belonging to the present invention
The identical meanings that the those of ordinary skill in domain is commonly understood by.It is also understood that those of definition term such as in usual dictionary
The meaning consistent with their meanings in the context of the relevant technologies should be interpreted as having, without application idealization or
The meaning of extremely formalization explains, unless being clearly defined herein.
The above is the description of the invention, and is not considered as limitation ot it.Notwithstanding of the invention several
Exemplary embodiment, but those skilled in the art will readily appreciate that, before without departing substantially from teaching and advantage of the invention
Many modifications can be carried out to exemplary embodiment by putting.Therefore, all such modifications are intended to be included in claims institute
In the scope of the invention of restriction.It should be appreciated that being the description of the invention above, and it should not be considered limited to disclosed spy
Determine embodiment, and the model in the appended claims is intended to encompass to the modification of the disclosed embodiments and other embodiments
In enclosing.The present invention is limited by claims and its equivalent.
Claims (14)
1. a kind of Information Atlas construction method, comprising:
Map element corresponding with the input information is extracted from multiclass subgraph spectrum according to input information；And
The extracted map element of split is to generate Information Atlas corresponding with the input information；
Wherein, at least part subgraph in the multiclass subgraph spectrum composes information node having the same, and is believed according to input
Breath extracts map element corresponding with the input information from multiclass subgraph spectrum
According to input information, the node attribute values of target information node are determined；And
Map element corresponding with identified node attribute values is extracted from every class subgraph spectrum.
2. Information Atlas construction method as described in claim 1, wherein every class subgraph spectrum has at least two information nodes,
And every class subgraph spectrum includes multiple map elements；For each information node, each map element has the information node
Node attribute values.
3. Information Atlas construction method as claimed in claim 2, wherein the extracted map element of split is defeated with this to generate
Entering the corresponding Information Atlas of information includes:
Using each identified node attribute values as linking point, link from the multiple subgraph spectrum in extract with the nodal community
It is worth corresponding map element.
4. Information Atlas construction method as described in claim 1, wherein the multiclass subgraph spectrum includes concept subgraph spectrum, public affairs
Take charge of at least two classes during subgraph is composed, stock subgraph is composed, investment subgraph spectrum, product subgraph spectrum, event subgraph are composed, industry subgraph is composed.
5. Information Atlas construction method as claimed in claim 2, wherein the target information node includes following at least one
Kind: Business Name, organization, name, film name, name of product, stock code, stock name, concept name.
6. Information Atlas construction method as claimed in claim 2, wherein the multiclass subgraph spectrum includes at least event subgraph
Spectrum, and the input information is event information,
Wherein, determine that the node attribute values of target information node include: according to input information
For the event information inputted, target information node corresponding with the event information is determined based on event subgraph spectrum
Node attribute values.
7. Information Atlas construction method as described in claim 1, wherein according to input information from multiclass subgraph spectrum in extract with
It further include constructing the multiclass subgraph spectrum before the corresponding map element of the input information, wherein every in building multiclass subgraph spectrum
A kind of subgraph is composed
At least two information nodes, and the corresponding relationship between setting information node are set；
The node attribute values of the information node are extracted from external data；
Extracted node attribute values are associated, to form map element；
Obtained map element is denoised, subgraph spectrum is obtained.
8. map construction method as claimed in claim 7, wherein at least two information node includes Business Name information
Node, and include: for the denoising of obtained map element
For each node attribute values of Business Name information node, based on the company's letter full name corresponding relationship pre-established,
The multiple node attribute values for meeting company's letter full name corresponding relationship are associated.
9. map construction method as claimed in claim 7, wherein at least two information node includes name information section
Point, and include: for the denoising of obtained map element
It will be split with multiple attribute values of the associated Business Name information node of same name information node, and using often
The node attribute values of a Business Name information node are identified the node attribute values of name information node, to eliminate name
Node ambiguity.
10. a kind of Information Atlas construction device, comprising:
Map element extraction module is configured as being extracted from multiclass subgraph spectrum according to input information corresponding with the input information
Map element；
Information Atlas generation module is configured as the extracted map element of split to generate information corresponding with the input information
Map.
11. Information Atlas construction device as claimed in claim 10, wherein every class subgraph spectrum has at least two information sections
Point, and every class subgraph spectrum includes multiple map elements；For each information node, each map element has the information node
Node attribute values；
Wherein, map element extraction module includes:
Target information node determining module is configured as determining the node attribute values of target information node according to input information；With
And
Map element respective modules are configured as extracting map corresponding with identified node attribute values from every class subgraph spectrum
Element.
12. Information Atlas construction device as claimed in claim 10, wherein Information Atlas generation module includes:
Map element link module is configured as linking using each identified node attribute values as linking point from the multiple
The map element corresponding with the node attribute values extracted in subgraph spectrum.
13. a kind of Information Atlas constructs equipment, wherein the equipment includes processor and memory, the memory includes one group
Instruction, one group of instruction make the Information Atlas building equipment execute operation, the operation when being executed by the processor
Include:
Map element corresponding with the input information is extracted from multiclass subgraph spectrum according to input information；And
The extracted map element of split is to generate Information Atlas corresponding with the input information；
Wherein, at least part subgraph in the multiclass subgraph spectrum composes information node having the same, and is believed according to input
Breath extracts map element corresponding with the input information from multiclass subgraph spectrum
According to input information, the node attribute values of target information node are determined；And
Map element corresponding with identified node attribute values is extracted from every class subgraph spectrum.
14. Information Atlas as claimed in claim 13 constructs equipment, wherein every class subgraph spectrum has at least two information sections
Point, and every class subgraph spectrum includes multiple map elements；For each information node, each map element has the information node
Node attribute values.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910114989.1A CN110162637B (en) | 2019-02-14 | 2019-02-14 | Information map construction method, device and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910114989.1A CN110162637B (en) | 2019-02-14 | 2019-02-14 | Information map construction method, device and equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110162637A true CN110162637A (en) | 2019-08-23 |
| CN110162637B CN110162637B (en) | 2023-06-20 |
Family
ID=67644878
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910114989.1A Active CN110162637B (en) | 2019-02-14 | 2019-02-14 | Information map construction method, device and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110162637B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110781317A (en) * | 2019-10-29 | 2020-02-11 | 北京明略软件系统有限公司 | Method and device for constructing event map and electronic equipment |
| CN111930906A (en) * | 2020-07-29 | 2020-11-13 | 北京北大软件工程股份有限公司 | Knowledge graph question-answering method and device based on semantic block |
| CN112612899A (en) * | 2020-11-24 | 2021-04-06 | 中国传媒大学 | Knowledge graph construction method and device, storage medium and electronic equipment |
| CN113708980A (en) * | 2021-10-27 | 2021-11-26 | 中国光大银行股份有限公司 | Topological graph generation method and device |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9026524B1 (en) * | 2013-01-10 | 2015-05-05 | Relationship Science LLC | Completing queries using transitive closures on a social graph |
| CN106777331A (en) * | 2017-01-11 | 2017-05-31 | 北京航空航天大学 | Knowledge mapping generation method and device |
| US20180060326A1 (en) * | 2016-08-26 | 2018-03-01 | Facebook, Inc. | Classifying Search Queries on Online Social Networks |
| CN108509420A (en) * | 2018-03-29 | 2018-09-07 | 赵维平 | Gu spectrum and ancient culture knowledge mapping natural language processing method |
| CN108763555A (en) * | 2018-06-01 | 2018-11-06 | 北京奇虎科技有限公司 | Representation data acquisition methods and device based on demand word |
| CN109033223A (en) * | 2018-06-29 | 2018-12-18 | 北京百度网讯科技有限公司 | For method, apparatus, equipment and computer readable storage medium across type session |
| CN109145122A (en) * | 2018-08-02 | 2019-01-04 | 北京仿真中心 | A kind of product know-how map construction and querying method and system |
-
2019
- 2019-02-14 CN CN201910114989.1A patent/CN110162637B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9026524B1 (en) * | 2013-01-10 | 2015-05-05 | Relationship Science LLC | Completing queries using transitive closures on a social graph |
| US20180060326A1 (en) * | 2016-08-26 | 2018-03-01 | Facebook, Inc. | Classifying Search Queries on Online Social Networks |
| CN106777331A (en) * | 2017-01-11 | 2017-05-31 | 北京航空航天大学 | Knowledge mapping generation method and device |
| CN108509420A (en) * | 2018-03-29 | 2018-09-07 | 赵维平 | Gu spectrum and ancient culture knowledge mapping natural language processing method |
| CN108763555A (en) * | 2018-06-01 | 2018-11-06 | 北京奇虎科技有限公司 | Representation data acquisition methods and device based on demand word |
| CN109033223A (en) * | 2018-06-29 | 2018-12-18 | 北京百度网讯科技有限公司 | For method, apparatus, equipment and computer readable storage medium across type session |
| CN109145122A (en) * | 2018-08-02 | 2019-01-04 | 北京仿真中心 | A kind of product know-how map construction and querying method and system |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110781317A (en) * | 2019-10-29 | 2020-02-11 | 北京明略软件系统有限公司 | Method and device for constructing event map and electronic equipment |
| CN110781317B (en) * | 2019-10-29 | 2022-03-01 | 北京明略软件系统有限公司 | Method and device for constructing event map and electronic equipment |
| CN111930906A (en) * | 2020-07-29 | 2020-11-13 | 北京北大软件工程股份有限公司 | Knowledge graph question-answering method and device based on semantic block |
| CN112612899A (en) * | 2020-11-24 | 2021-04-06 | 中国传媒大学 | Knowledge graph construction method and device, storage medium and electronic equipment |
| CN113708980A (en) * | 2021-10-27 | 2021-11-26 | 中国光大银行股份有限公司 | Topological graph generation method and device |
| CN113708980B (en) * | 2021-10-27 | 2022-04-19 | 中国光大银行股份有限公司 | Topological graph generation method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110162637B (en) | 2023-06-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11327978B2 (en) | Content authoring | |
| CN112131366B (en) | Method, device and storage medium for training text classification model and text classification | |
| US10437868B2 (en) | Providing images for search queries | |
| CN110209897B (en) | Intelligent dialogue method, device, storage medium and equipment | |
| Feagin et al. | Rethinking racial formation theory: A systemic racism critique | |
| CN110162637A (en) | Information Atlas construction method, device and equipment | |
| US20170161619A1 (en) | Concept-Based Navigation | |
| US20170262783A1 (en) | Team Formation | |
| US20210406473A1 (en) | System and method for building chatbot providing intelligent conversational service | |
| CN113535974B (en) | Diagnostic recommendation method and related device, electronic equipment and storage medium | |
| CN111597314A (en) | Reasoning question-answering method, device and equipment | |
| US11650979B2 (en) | Assigning a new entigen to a word group | |
| CN108287875B (en) | Character co-occurrence relation determining method, expert recommending method, device and equipment | |
| CN111259154B (en) | Data processing method and device, computer equipment and storage medium | |
| CN111783903B (en) | Text processing method, text model processing method and device and computer equipment | |
| JP2022006173A (en) | Knowledge pre-training model training method, device and electronic equipment | |
| CN112287085B (en) | Semantic matching method, system, equipment and storage medium | |
| CN113761220A (en) | Information acquisition method, device, equipment and storage medium | |
| JP2023002690A (en) | Semantics recognition method, apparatus, electronic device, and storage medium | |
| CN113407850A (en) | Method and device for determining and acquiring virtual image and electronic equipment | |
| CN113656587A (en) | Text classification method and device, electronic equipment and storage medium | |
| US20200210643A1 (en) | Generating a query response utilizing a knowledge database | |
| EP4327197A1 (en) | Task execution based on real-world text detection for assistant systems | |
| US20200175064A1 (en) | Image processing utilizing an entigen construct | |
| CN114611529B (en) | Intention recognition method and device, electronic equipment and storage medium |
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 |