CN103631970B - The method and apparatus for excavating attribute and entity associated relation - Google Patents
The method and apparatus for excavating attribute and entity associated relation Download PDFInfo
- Publication number
- CN103631970B CN103631970B CN201310714291.6A CN201310714291A CN103631970B CN 103631970 B CN103631970 B CN 103631970B CN 201310714291 A CN201310714291 A CN 201310714291A CN 103631970 B CN103631970 B CN 103631970B
- Authority
- CN
- China
- Prior art keywords
- entity
- attribute
- fructification
- sample cluster
- user
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/288—Entity relationship models
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Fuzzy Systems (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Mathematical Physics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention proposes a kind of method and apparatus for excavating attribute and entity associated relation, and wherein this method includes：Obtain attribute to be associated；At least one is obtained from multiple entities according to attribute to be associated and plants fructification；And at least one associated entity for planting fructification is obtained, and attribute to be associated is associated with least one kind fructification, the associated entity of at least one kind fructification.The method of the embodiment of the present invention, can excavate multiple associated entities of attribute to be associated, similarly realize the user-specific attributes excavated corresponding to entity（Attribute i.e. to be associated）, so that provide more comprehensively, the finer, detailed service of more high-quality；Any domain entities and user-specific attributes can also be excavated（Attribute i.e. to be associated）Between incidence relation, do not limited, be widely used by application field.
Description
Technical field
The present invention relates to field of computer technology, more particularly to a kind of method and dress for excavating attribute and entity associated relation
Put.
Background technology
With the fast development of Internet technology, particularly wireless Internet technologies, information service becomes increasingly prevalent.
When information service provider provides information service, for example, search engine provides search service etc., it will usually excavate entity and attribute
Between incidence relation, and according between entity and attribute incidence relation provide information service.Specifically, can be by real world
In objective things be referred to as entity, such as concept, things or event.For example, movie and television play " I is special technical soldier ", Baidu are public
Department and Big Bang Theory are all the examples of entity.Meanwhile, each entity has attribute, the related letter of attribute reflection entity
Breath, for example, army's subject matter, corporate office place, modern universe theory are the corresponding attribute of above-mentioned entity respectively.
The current method for obtaining incidence relation between entity and attribute is mainly oriented from the structural data of website grabs
Entity attribute pair is taken, and according to entity attribute to setting up the incidence relation between entity and attribute.But, it is primarily present following ask
Topic, because the corresponding attribute of an entity is diversified, for one entity of correspondence, the attribute obtained from website is certain
On one side, the attribute possibly can not meet the demand of user well.Therefore prior art can not be excavated corresponding to entity
User-specific attributes, for example, can not excavate some film belongs to " Cock counteroffensives " attribute etc., similarly, can not also excavate
Go out film, the novel of the corresponding entity of attribute, such as " Cock counteroffensives " subject matter such as " Cock counteroffensives ", " curing system ", " the cruel heart ".
The content of the invention
It is contemplated that at least solving one of above-mentioned technical problem.
Therefore, first purpose of the present invention is to propose a kind of method for excavating attribute and entity associated relation.The party
Method can excavate multiple associated entities of attribute to be associated, similarly realize the user-specific attributes excavated corresponding to entity（I.e.
Attribute to be associated）, so that provide more comprehensively, the finer, detailed service of more high-quality.
Second object of the present invention is to propose a kind of excavation attribute and the device of entity associated relation.
To achieve these goals, the method for excavating attribute and entity associated relation of first aspect present invention embodiment,
Comprise the following steps：Obtain attribute to be associated；At least one seed is obtained from multiple entities according to the attribute to be associated real
Body；And obtain it is described at least one plant the associated entity of fructification, and will the attribute to be associated with it is described at least one plant
Fructification, at least one described associated entity for planting fructification are associated.
The method for excavating attribute and entity associated relation of the embodiment of the present invention, obtains seed real by attribute to be associated
Body, related associated entity is obtained further according to kind of fructification, thus, it is possible to multiple associated entities of attribute to be associated are excavated,
Similarly realize the user-specific attributes excavated corresponding to entity（Attribute i.e. to be associated）, so that provide more comprehensively, it is finer, more excellent
The detailed service of matter, for example, according to user-specific attributes to user's recommended entity；Any field can also be excavated according to this method
Entity and given attribute（Attribute i.e. to be associated）Between incidence relation, do not limited, be widely used by application field.
To achieve these goals, the excavation attribute and the device of entity associated relation of second aspect of the present invention embodiment,
Including：Attribute acquisition module to be associated, for obtaining attribute to be associated；Fructification acquisition module is planted, for waiting to close according to
It is attribute that at least one kind fructification is obtained from multiple entities；Associated entity acquisition module, for obtain it is described at least one
Plant the associated entity of fructification；And relating module, for by the attribute to be associated with it is described at least one plant fructification, institute
Stating at least one associated entity for planting fructification is associated.
The excavation attribute and the device of entity associated relation of the embodiment of the present invention, pass through attribute acquisition module to be associated and obtain
Attribute to be associated, then plants fructification acquisition module and obtains kind of a fructification according to attribute to be associated, associated entity obtains mould afterwards
Root tuber obtains the associated entity of kind of fructification according to kind of fructification, thus, it is possible to multiple associated entities of attribute to be associated are excavated,
Similarly realize the user-specific attributes excavated corresponding to entity（Attribute i.e. to be associated）, so that provide more comprehensively, it is finer, more excellent
The detailed service of matter, for example, according to user-specific attributes to user's recommended entity；Any field can also be excavated according to the device
Entity and user-specific attributes（Attribute i.e. to be associated）Between incidence relation, do not limited, be widely used by application field.
The additional aspect of the present invention and advantage will be set forth in part in the description, and will partly become from the following description
Obtain substantially, or recognized by the practice of the present invention.
Brief description of the drawings
Of the invention above-mentioned and/or additional aspect and advantage will become from the following description of the accompanying drawings of embodiments
Substantially and be readily appreciated that, wherein,
Fig. 1 is the flow chart of the method for excavation attribute according to an embodiment of the invention and entity associated relation；
Fig. 2 is the flow chart of the method for excavation attribute according to an embodiment of the invention and entity associated relation；
Fig. 3 is the flow chart of acquisition distributional difference value according to an embodiment of the invention；
Fig. 4 is the flow chart according to an embodiment of the invention for obtaining and obtaining associated entity；
Fig. 5 is the structural representation of excavation attribute according to an embodiment of the invention and the device of entity associated relation；
Fig. 6 is the structural representation of excavation attribute according to an embodiment of the invention and the device of entity associated relation.
Embodiment
Embodiments of the invention are described below in detail, the example of the embodiment is shown in the drawings, wherein from beginning to end
Same or similar label represents same or similar element or the element with same or like function.Below with reference to attached
The embodiment of figure description is exemplary, is only used for explaining the present invention, and is not considered as limiting the invention.On the contrary, this
All changes in the range of spirit and intension that the embodiment of invention includes falling into attached claims, modification and equivalent
Thing.
In the description of the invention, it is to be understood that term " first ", " second " etc. be only used for describe purpose, without
It is understood that to indicate or imply relative importance.In the description of the invention, it is necessary to which explanation, is provided unless otherwise clear and definite
And restriction, term " connected ", " connection " should be interpreted broadly, for example, it may be fixedly connected or be detachably connected,
Or be integrally connected；Can be mechanical connection or electrical connection；Can be joined directly together, intermediary can also be passed through
It is indirectly connected to.For the ordinary skill in the art, the tool of above-mentioned term in the present invention can be understood with concrete condition
Body implication.In addition, in the description of the invention, unless otherwise indicated, " multiple " are meant that two or more.
Any process described otherwise above or method description are construed as in flow chart or herein, represent to include
Module, fragment or the portion of the code of one or more executable instructions for the step of realizing specific logical function or process
Point, and the scope of the preferred embodiment of the present invention includes other realization, wherein can not be by shown or discussion suitable
Sequence, including according to involved function by it is basic simultaneously in the way of or in the opposite order, carry out perform function, this should be of the invention
Embodiment person of ordinary skill in the field understood.
In order to excavate entity and user-specific attributes in any field（Such as user-specific attributes）Between incidence relation,
So as to provide the user with information service more comprehensively, finer, the present invention proposes a kind of attribute and entity associated relation excavated
Method and apparatus.Below with reference to the accompanying drawings the method and apparatus for excavating attribute and entity associated relation of the embodiment of the present invention are described.
A kind of method for excavating attribute and entity associated relation, comprises the following steps：Obtain attribute to be associated；According to waiting to close
It is attribute that at least one kind fructification is obtained from multiple entities；And at least one associated entity for planting fructification is obtained, and
Attribute to be associated is associated with least one kind fructification, the associated entity of at least one kind fructification.
Fig. 1 is the flow chart of the method for excavation attribute according to an embodiment of the invention and entity associated relation.
As shown in figure 1, the method for excavating attribute and entity associated relation comprises the steps.
Step S101, obtains attribute to be associated.
In one embodiment of the invention, attribute to be associated is that a class describes the features such as user's impression, product performance
Attribute.Attribute to be associated can be with netspeak real-time update, for example, category to be associated can be obtained by carrying out analysis to multiple webpages
Property.For example, there can be the description users such as " Cock counteroffensives ", " evilness defeats justice ", " curing system ", " the cruel heart ", " exposing the wealth "
The attribute to be associated of impression；For product entity, there can be the description user experience such as " cost performance is high ", " durable "
Attribute to be associated.
Step S102, obtains at least one according to attribute to be associated from multiple entities and plants fructification.
Specifically, obtain after attribute to be associated, at least one seed is obtained from multiple entities according to attribute to be associated
Entity.Wherein, a kind fructification will be used as with the name of attribute relationship to be associated is close, the degree of correlation is high entity.If for example, waiting to close
Attribute is " cure and be ", then the kind fructification obtained can be the movie and television play entity of " curing system ", the novel reality of " healing system "
Other entities of body, the caricature entity of " curing system " or " curing system " etc..The degree of association of the process and user and entity, service
Using relevant with the degree of association of entity, it will be described in detail in subsequent embodiment.
Step S103, obtains at least one associated entity for planting fructification, and by attribute to be associated and at least one seed
Entity, at least one associated entity for planting fructification are associated.
Specifically, at least one is obtained from multiple entities to plant after fructification, then centered at least one kind fructification,
Obtain at least one and plant the higher associated entity of the fructification degree of correlation.Using obtained from multiple entities a kind fructification as
Example, if for example, the kind fructification obtained from multiple entities is the movie and television play seed entity A of " cure system ", then obtained
The associated entity of the movie and television play seed entity A of " curing system " must be somebody's turn to do, the associated entity of such as acquisition can be the small of " curing system "
Say entity B, the caricature entity C of " cure system ", the movie and television play seed F of other entities E of " curing system " or other " curing system "
With G etc..The process can expand the scope of entity, recall some associated entities.
More specifically, after obtaining the associated entity that at least one plants fructification, by attribute to be associated and at least one kind
Fructification, at least one associated entity for planting fructification are associated.For example, obtain associated entity " cure system " novel entity or
After the movie and television play entity of other " cure system ", by the movie and television play kind fructification of attribute to be associated " cure and be " and " cure and be ",
The associated entity of the movie and television play kind fructification of " curing system "（The novel entity for " curing system " or the movie and television play of other " curing system "
Entity）It is associated.
Wherein, associated operation can be to attribute to be associated, at least one kind fructification, at least one kind fructification
Associated entity is labelled or the corresponding relation set up between them etc..For example, " can cure by attribute to be associated and be "
Movie and television play kind fructification, the associated entity of the movie and television play kind fructification of " cure and be " with " cure and be "（That " cures system " is small
Say the movie and television play entity of entity or other " curing system "）The corresponding relation for sticking the label of " curing system " or setting up between them
Deng.
The method for excavating attribute and entity associated relation of the embodiment of the present invention, obtains seed real by attribute to be associated
Body, related associated entity is obtained further according to kind of fructification, thus, it is possible to multiple associated entities of attribute to be associated are excavated,
Similarly realize the user-specific attributes excavated corresponding to entity（Attribute i.e. to be associated）, so that provide more comprehensively, it is finer, more excellent
The detailed service of matter, for example, according to user-specific attributes to user's recommended entity；Any field can also be excavated according to this method
Entity and given attribute（Attribute i.e. to be associated）Between incidence relation, do not limited, be widely used by application field.
Fig. 2 is the flow chart of the method for excavation attribute in accordance with another embodiment of the present invention and entity associated relation.
In embodiments of the invention, kind of a fructification is obtained from multiple entities by the way of distributional difference.
Specifically, as shown in Fig. 2 the method for excavating attribute and entity associated relation comprises the steps.
Step S201, obtains attribute to be associated.
In one embodiment of the invention, attribute to be associated is that a class describes the features such as user's impression, product performance
Attribute.Attribute to be associated can be with netspeak real-time update, for example, category to be associated can be obtained by carrying out analysis to multiple webpages
Property.For example, there can be the description users such as " Cock counteroffensives ", " evilness defeats justice ", " curing system ", " the cruel heart ", " exposing the wealth "
The attribute to be associated of impression；For product entity, there can be the description user experience such as " cost performance is high ", " durable "
Attribute to be associated.
Step S202, multiple entities are obtained from default entity storehouse.
Specifically, be stored with multiple realities in the entity storehouse for presetting entity storehouse to obtain from network in advance, default entity storehouse
Body, wherein, default entity storehouse can be stored in the server or in miscellaneous equipment.Default entity storehouse can also be divided
Class, different application services can have different default entity storehouses.
Step S203, obtains the association user sample cluster with attribute to be associated from overall user sample cluster.
Specifically, according to attribute to be associated, the association user with attribute to be associated is obtained from overall user sample cluster
Sample cluster.For example, if attribute to be associated is " exposing the wealth ", overall user sample cluster is the user of 10,000,000 viewing movie and television plays,
So obtain the user of 1,000,000 viewing " exposing the wealth " movie and television plays in overall user sample cluster, the i.e. pass with attribute to be associated
It is combined family sample cluster.
Step S204, obtains multiple distributional difference values of multiple entities in association user sample cluster respectively.
Specifically, same entity is in overall user sample cluster and in the association user sample cluster with attribute to be associated
Distribution be different.The size of distributional difference value can embody entity and the height of the degree of correlation of attribute to be associated with corresponding
It is low, it is easy to subsequently screen entity according to distributional difference value.Specifically the acquisition of distributional difference value will be in subsequent embodiment
In describe in detail.
Multiple entities are screened and plant fructification to obtain at least one by step S205 according to multiple distributional difference values.
Specifically, obtain after multiple distributional difference values of multiple entities in association user sample cluster, according to multiple distributions
Difference value is screened to multiple entities plants fructification to obtain at least one.Wherein, it is according to distributional difference value to plant fructification
The entity higher with attributes correlation to be associated screened from multiple entities.
Step S206, obtains at least one associated entity for planting fructification, and by attribute to be associated and at least one seed
Entity, at least one associated entity for planting fructification are associated.
Specifically, at least one is obtained from multiple entities to plant after fructification, then centered at least one kind fructification,
Obtain at least one and plant the higher associated entity of the fructification degree of correlation.Using obtained from multiple entities a kind fructification as
Example, if for example, the kind fructification obtained from multiple entities is the movie and television play seed entity A of " cure system ", then obtained
The associated entity of the movie and television play seed entity A of " curing system " must be somebody's turn to do, the associated entity of such as acquisition can be the small of " curing system "
Say entity B, the caricature entity C of " cure system ", the movie and television play seed F of other entities E of " curing system " or other " curing system "
With G etc..The process can expand the scope of entity, recall some associated entities.
More specifically, after obtaining the associated entity that at least one plants fructification, by attribute to be associated and at least one kind
Fructification, at least one associated entity for planting fructification are associated.For example, obtain associated entity " cure system " novel entity or
After the movie and television play entity of other " cure system ", by the movie and television play kind fructification of attribute to be associated " cure and be " and " cure and be ",
The associated entity of the movie and television play kind fructification of " curing system "（The novel entity for " curing system " or the movie and television play of other " curing system "
Entity）It is associated.
Wherein, associated operation can be to attribute to be associated, at least one kind fructification, at least one kind fructification
Associated entity is labelled or the corresponding relation set up between them etc..For example, " can cure by attribute to be associated and be "
Movie and television play kind fructification, the associated entity of the movie and television play kind fructification of " cure and be " with " cure and be "（That " cures system " is small
Say the movie and television play entity of entity or other " curing system "）The corresponding relation for sticking the label of " curing system " or setting up between them
Deng.
The method for excavating attribute and entity associated relation of the embodiment of the present invention, using distributional difference value from multiple entities
Kind of a fructification is obtained, distributional difference value truly reflects the distribution of kind of fructification, kind fructification and the attribute to be associated of acquisition
The degree of correlation is higher, more accurate, so as to further lift the quality of information service.
Fig. 3 is the flow chart of acquisition distributional difference value according to an embodiment of the invention.In the implementation of the present invention
In example, as shown in figure 3, step S204 is specifically included：
S2041, obtains multiple first points of the multiple users related to multiple entities in overall user sample cluster respectively
Cloth proportion.
For example, overall user sample cluster is the user of 10,000,000 viewing movie and television plays, wherein the user for having 500,000 have viewed
Movie and television play entity M, then distribution proportion of the viewing movie and television play entity M user in overall user sample cluster for 500,000 divided by
10000000, i.e., the first distribution proportion is 5%.Similarly, the multiple users related to multiple entities are obtained successively in overall user sample
Multiple first distribution proportions in this group.
S2042, obtains second distribution ratio of the multiple users related to multiple entities in association user sample cluster respectively
Weight.
For example, attribute to be associated is " exposing the wealth ", association user sample cluster is the use of 1,000,000 viewing " exposing the wealth " movie and television plays
Family, wherein, 300,000 users have viewed movie and television play entity M, then viewing movie and television play entity M user is in association user sample cluster
In distribution proportion be 300,000 divided by 1,000,000, i.e., second distribution proportion be 30%.Similarly, obtain related to multiple entities successively
Multiple users in association user sample cluster it is multiple second distribution proportions.
S2043, distributional difference value is obtained according to the second distribution proportion and the first distribution proportion.
Specifically, according to the second of acquisition the distribution proportion and the first distribution proportion, with the second distribution proportion divided by first point
Cloth proportion is to obtain distributional difference value.
For example, overall user sample cluster is the user of 10,000,000 viewing movie and television plays, wherein the user for having 500,000 have viewed
Movie and television play entity M, then the first distribution proportion is 5%；If attribute to be associated is " exposing the wealth ", association user sample cluster is 1,000,000
The user of individual viewing " exposing the wealth " movie and television play, wherein, 300,000 users have viewed movie and television play entity M, then second, which is distributed proportion, is
30%, then with 30% divided by 5%, that is, it is 6 to obtain distributional difference value.Wherein distributional difference value is bigger, illustrates movie and television play entity M with treating
The degree of correlation of relating attribute " exposing the wealth " is higher.
Thus, the distributional difference value obtained according to the first distribution proportion and the second distribution proportion can more embody the degree of association, point
Cloth difference value is more accurate.
In one embodiment of the invention, in step S205, overall user sample cluster is multiple, is corresponded to respectively multiple
Network english teaching, then the corresponding distributional difference value of each entity is multiple, and multiple entities are entered according to multiple distributional difference values
Row screens to obtain at least one described kind fructification（That is step S205）Also include：According to default distributional difference value screening rule
The multiple entity is screened；Or, distributional difference value grader is created, and according to distributional difference value grader to multiple realities
Body is screened, furthermore it is also possible to use other methods.
Specifically, known below with entity in association user sample cluster, Baidu's mhkc, Baidu, the distribution in Baidu's session
The method for illustrating to screen multiple entities according to default distributional difference value screening rule exemplified by difference.What this method was used
Screening rule is as follows：
（1）Output entity is known in association user sample cluster, Baidu's mhkc, Baidu, the distributional difference value in Baidu's session
Larger entity, distinguishes presentation-entity with Suser, Stieba, Siknow, Ssession and is pasted in association user sample cluster, Baidu
, Baidu know, the distributional difference value in Baidu's session, such as：Export Suser>10、Stieba>50、Siknow>50 or
Ssession>30 entity；
（2）Export at least one in Stieba, Siknow, Ssession and be more than entities of the 3 and Suser also greater than 3；
（3）Export the entity of Stieba, Siknow, Ssession all greater than 3；
（4）Export at least one in Stieba, Siknow, Ssession be more than 3, one be more than 8 entity.
Grader can also be set up according to above-mentioned screening rule, it is for instance possible to use prior art sets up grader
Method set up classification, the foundation of grader can improve efficiency.The foundation of grader can use prior art, herein no longer
Repeat.
It is above-mentioned that at least one method accuracy rate height for planting fructification is screened in multiple entities according to distributional difference value, still
Entity in the screening rule of setting below threshold value can not be called back, and be the association that this subsequently also needs to obtain kind of fructification
Entity.
Fig. 4 is the flow chart according to an embodiment of the invention for obtaining and obtaining associated entity.In the reality of the present invention
Apply in example, specifically included as shown in figure 4, obtaining at least one associated entity for planting fructification in step S206：
S2061, obtains at least one and plants fructification to first between user's sample cluster with attribute to be associated respectively
Incidence relation.
Specifically, for example, can be described by matrix kind of fructification to attribute to be associated user's sample cluster it
Between the first incidence relation, such as matrix A.
S2062, obtains the associated entity group of user's sample cluster with attribute to be associated, and obtains with attribute to be associated
User's sample cluster to associated entity group between the second incidence relation.
Specifically, the associated entity group of user's sample cluster with attribute to be associated is obtained, if for example, with to be associated
User's sample cluster of attribute for the movie and television play entity of viewing " cure system " user, then obtain " curing system " movie and television play entity,
Other entities of " curing system " novel entity, " curing system " caricature entity or " curing system ", as with attribute to be associated
The associated entity group of user's sample cluster.
More specifically, user's sample cluster with attribute to be associated can be described by matrix between associated entity group
The second incidence relation, such as matrix B.
S2063, obtains at least one kind fructification real to association respectively according to the first incidence relation and the second incidence relation
The 3rd incidence relation of body group.
Specifically, at least one can be for example obtained according to matrix A and matrix B and plants fructification to the of associated entity group
Three incidence relations, can be described with Matrix C., can be with for example, Matrix C can be got by simple matrix multiple
It is weighted after handling and is multiplied again.
S2064, is screened to obtain at least according to the 3rd incidence relation to each associated entity in associated entity group
The associated entity of one kind fructification.
For example, the 3rd incidence relation can be identified with Matrix C, each element in Matrix C is the entity seed to association
Degree of correlation information between entity, path phase of kind of the fructification to the path of each associated entity can be obtained according to the matrix
Like degree pathsim features, the associated entity of kind of fructification is obtained according to this feature.In addition, pathsim features can also be found
With entity peer objects, the influence of popular entity is reduced.Wherein, the calculation formula of Pathsim features is as follows：
Wherein, aiFor i-th of entity, ajFor j-th of entity, pcR(ai,aj) it is the i-th row, the element value of jth row in Matrix C
（That is entity aiWith entity ajBetween the degree of correlation）, pcR(ai,ai) it is the i-th row, the element value of the i-th row in Matrix C（That is entity ai
The degree of correlation of itself）, pcR-1(aj,ai) be Matrix C inverse matrix C-1The element value of middle jth row, the i-th row, pcR-1(aj,aj) be
The inverse matrix C of Matrix C-1The element value of middle jth row, jth row.
Filter the entity associated out of above-mentioned acquisition.Specifically, it can be filtered out with given threshold in the entity associated out
Doubtful incoherent entity, wherein, threshold value can be the multiple of distributional difference value of the seed entity on association user sample cluster,
Such as 2 times, 3 times or other multiples.
Thus, the 3rd incidence relation of acquisition has more directly reacted the associated entity for planting fructification, makes the association of acquisition
Entity is more accurate.
In order to realize above-described embodiment, the present invention also proposes a kind of excavation attribute and the device of entity associated relation.
A kind of device of excavation attribute and entity associated relation, including：Attribute acquisition module to be associated, waits to close for obtaining
It is attribute；Fructification acquisition module is planted, fructification is planted for obtaining at least one from multiple entities according to attribute to be associated；Close
Join entity acquisition module, for obtaining at least one associated entity for planting fructification；And relating module, for by category to be associated
Property with least one kind fructification, at least one plant fructification associated entity it is associated.
Fig. 5 is the structural representation of excavation attribute according to an embodiment of the invention and the device of entity associated relation.
As shown in figure 5, excavating the device of attribute and entity associated relation includes：Attribute acquisition module 100 to be associated, seed
Entity acquisition module 200, associated entity acquisition module 300 and relating module 400.
Wherein, attribute acquisition module 100 to be associated is used to obtain attribute to be associated.
Specifically, attribute to be associated is the attribute that a class describes the features such as user's impression, product performance.Attribute to be associated can
With with netspeak real-time update, for example, attribute to be associated can be obtained by carrying out analysis to multiple webpages.For example, Ke Yiyou
The attribute to be associated of description user's impression such as " Cock counteroffensives ", " evilness defeats justice ", " curing system ", " the cruel heart ", " exposing the wealth "；It is right
In product entity, there can be the attribute to be associated of the description user experience such as " cost performance is high ", " durable ".
Planting fructification acquisition module 200 is used to obtain at least one seed reality from multiple entities according to attribute to be associated
Body.
Specifically, obtain after attribute to be associated, at least one seed is obtained from multiple entities according to attribute to be associated
Entity.Wherein, a kind fructification will be used as with the name of attribute relationship to be associated is close, the degree of correlation is high entity.If for example, waiting to close
Attribute is " cure and be ", then the kind fructification obtained can be the movie and television play entity of " curing system ", the novel reality of " healing system "
Other entities of body, the caricature entity of " curing system " or " curing system " etc..The degree of association of the process and user and entity, service
Using relevant with the degree of association of entity, it will be described in detail in subsequent embodiment.
Associated entity acquisition module 300 is used to obtain at least one associated entity for planting fructification.
Specifically, at least one is obtained from multiple entities to plant after fructification, then centered at least one kind fructification,
Obtain at least one and plant the higher associated entity of the fructification degree of correlation.Using obtained from multiple entities a kind fructification as
Example, if for example, the kind fructification obtained from multiple entities is the movie and television play seed entity A of " cure system ", then obtained
The associated entity of the movie and television play seed entity A of " curing system " must be somebody's turn to do, the associated entity of such as acquisition can be the small of " curing system "
Say entity B, the caricature entity C of " cure system ", the movie and television play seed F of other entities E of " curing system " or other " curing system "
With G etc..The process can expand the scope of entity, recall some associated entities.
Relating module 400 is used to associate attribute to be associated and at least one kind fructification, at least one kind fructification
Entity is associated.
Specifically, after the associated entity for obtaining at least one kind fructification, by attribute to be associated and at least one seed
Entity, at least one associated entity for planting fructification are associated.
For example, after obtaining the novel entity of associated entity " curing system " or the movie and television play entity of other " curing system ", will
The movie and television play kind fructification of attribute to be associated " curing system " and " cure and be ", the movie and television play kind fructification of " curing system " associate reality
Body（The novel entity for " curing system " or the movie and television play entity of other " curing system "）It is associated.
Wherein, associated operation can be to attribute to be associated, at least one kind fructification, at least one kind fructification
Associated entity is labelled or the corresponding relation set up between them etc..For example, " can cure by attribute to be associated and be "
Movie and television play kind fructification, the associated entity of the movie and television play kind fructification of " cure and be " with " cure and be "（That " cures system " is small
Say the movie and television play entity of entity or other " curing system "）The corresponding relation for sticking the label of " curing system " or setting up between them
Deng.
The excavation attribute and the device of entity associated relation of the embodiment of the present invention, pass through attribute acquisition module to be associated and obtain
Attribute to be associated, then plants fructification acquisition module and obtains kind of a fructification according to attribute to be associated, associated entity obtains mould afterwards
Root tuber obtains the associated entity of kind of fructification according to kind of fructification, thus, it is possible to multiple associated entities of attribute to be associated are excavated,
Similarly realize the user-specific attributes excavated corresponding to entity（Attribute i.e. to be associated）, so that provide more comprehensively, it is finer, more excellent
The detailed service of matter, for example, according to user-specific attributes to user's recommended entity；Any field can also be excavated according to the device
Entity and user-specific attributes（Attribute i.e. to be associated）Between incidence relation, do not limited, be widely used by application field.
Fig. 6 is the structural representation of excavation attribute according to an embodiment of the invention and the device of entity associated relation.
As shown in fig. 6, excavating the device of attribute and entity associated relation includes：Attribute acquisition module 100 to be associated, seed
Entity acquisition module 200, entity acquiring unit 210, association user sample cluster acquiring unit 220, distributional difference value acquiring unit
230th, screening unit 240, associated entity acquisition module 300, the first incidence relation acquiring unit 310, the second incidence relation are obtained
Unit 320, the 3rd incidence relation acquiring unit 330, screening unit 340 and relating module 400.Wherein, plant fructification and obtain mould
Block 200 includes entity acquiring unit 210, association user sample cluster acquiring unit 220, distributional difference value acquiring unit 230, screening
Unit 240；Associated entity acquisition module 300 includes the first incidence relation acquiring unit 310, the second incidence relation acquiring unit
320th, the 3rd incidence relation acquiring unit 330, screening unit 340.
In one embodiment of the invention, the first incidence relation acquiring unit 310, the second incidence relation acquiring unit
320th, the 3rd incidence relation acquiring unit 330, screening unit 340 are optional.
Specifically, attribute acquisition module 100 to be associated is used to obtain attribute to be associated.
In one embodiment of the invention, attribute to be associated is that a class describes the features such as user's impression, product performance
Attribute.Attribute to be associated can be with netspeak real-time update, for example, category to be associated can be obtained by carrying out analysis to multiple webpages
Property.For example, there can be the description users such as " Cock counteroffensives ", " evilness defeats justice ", " curing system ", " the cruel heart ", " exposing the wealth "
The attribute to be associated of impression；For product entity, there can be the description user experience such as " cost performance is high ", " durable "
Attribute to be associated.
Entity acquiring unit 210 is used to obtain multiple entities from default entity storehouse.
Specifically, be stored with multiple realities in the entity storehouse for presetting entity storehouse to obtain from network in advance, default entity storehouse
Body, wherein, default entity storehouse can be stored in the server or in miscellaneous equipment.Default entity storehouse can also be divided
Class, different application services can have different default entity storehouses.
Association user sample cluster acquiring unit 220 is used to from overall user sample cluster obtain the pass with attribute to be associated
It is combined family sample cluster.
Specifically, according to attribute to be associated, the association user with attribute to be associated is obtained from overall user sample cluster
Sample cluster.For example, if attribute to be associated is " exposing the wealth ", overall user sample cluster is the user of 10,000,000 viewing movie and television plays,
So obtain the user of 1,000,000 viewing " exposing the wealth " movie and television plays in overall user sample cluster, the i.e. pass with attribute to be associated
It is combined family sample cluster.
Distributional difference value acquiring unit 230 is used to obtain multiple distributions of multiple entities in association user sample cluster respectively
Difference value.
Specifically, same entity is in overall user sample cluster and in the association user sample cluster with attribute to be associated
Distribution be different.The size of distributional difference value can embody entity and the height of the degree of correlation of attribute to be associated with corresponding
It is low, it is easy to subsequently screen entity according to distributional difference value.Specifically the acquisition of distributional difference value will be in subsequent embodiment
In describe in detail.
In one embodiment of the invention, distributional difference value acquiring unit 230 also particularly useful for：Respectively obtain with it is multiple
Multiple first distribution proportions of the related multiple users of entity in overall user sample cluster, and obtain and multiple entity phases respectively
Second distribution proportion of the multiple users closed in association user sample cluster, and compared according to the second distribution proportion and the first distribution
Recapture and take distributional difference value.
Wherein, the acquisition of the first distribution proportion is exemplified below, for example, overall user sample cluster is 10,000,000 viewings
The user of movie and television play, wherein the user for having 500,000 have viewed movie and television play entity M, then viewing movie and television play entity M user is total
Distribution proportion in body user's sample cluster is 500,000 divided by 10,000,000, i.e., the first distribution proportion is 5%.Similarly, successively obtain with
Multiple first distribution proportions of the related multiple users of multiple entities in overall user sample cluster.
The acquisition of the second distribution proportion is exemplified below, for example, attribute to be associated is " exposing the wealth ", association user sample cluster
For 1,000,000 viewing " exposing the wealth " movie and television play users, wherein, 300,000 users have viewed movie and television play entity M, then viewing video display
Distribution proportion of the acute entity M user in association user sample cluster is 300,000 divided by 1,000,000, i.e., the second distribution proportion is 30%.
Similarly, multiple second distribution proportions of the multiple users related to multiple entities in association user sample cluster are obtained successively.
According to the second of acquisition the distribution proportion and the first distribution proportion, it is with the second distribution proportion divided by the first distribution proportion
Obtain distributional difference value.For example, overall user sample cluster is the user of 10,000,000 viewing movie and television plays, wherein there is 500,000 user
It has viewed movie and television play entity M, then the first distribution proportion is 5%；If attribute to be associated is " exposing the wealth ", association user sample cluster
For 1,000,000 viewing " exposing the wealth " movie and television play users, wherein, 300,000 users have viewed movie and television play entity M, then second distribution
Proportion is 30%, then with 30% divided by 5%, that is, it is 6 to obtain distributional difference value.Wherein distributional difference value is bigger, illustrates movie and television play entity
M and the degree of correlation of attribute to be associated " exposing the wealth " are higher.
Thus, the distributional difference value obtained according to the first distribution proportion and the second distribution proportion can more embody the degree of association, point
Cloth difference value is more accurate.
Screening unit 240 is used to multiple entities are screened according to multiple distributional difference values to obtain at least one seed
Entity.
Specifically, obtain after multiple distributional difference values of multiple entities in association user sample cluster, according to multiple distributions
Difference value is screened to multiple entities plants fructification to obtain at least one.Wherein, it is according to distributional difference value to plant fructification
The entity higher with attributes correlation to be associated screened from multiple entities.
In addition, overall user sample cluster is multiple, multiple network english teachings are corresponded to respectively, then corresponding point of each entity
Cloth difference value is multiple, and screening unit 240 is screened according to multiple distributional difference values to multiple entities also to be included：According to default
Distributional difference value screening rule is screened to multiple entities；Or, distributional difference value grader is created, and according to distributional difference value
Grader is screened to multiple entities, furthermore it is also possible to use other methods.
Specifically, known below with entity in association user sample cluster, Baidu's mhkc, Baidu, the distribution in Baidu's session
The method for illustrating to screen multiple entities according to default distributional difference value screening rule exemplified by difference.What this method was used
Screening rule is as follows：
（1）Output entity is known in association user sample cluster, Baidu's mhkc, Baidu, the distributional difference value in Baidu's session
Larger entity, distinguishes presentation-entity with Suser, Stieba, Siknow, Ssession and is pasted in association user sample cluster, Baidu
, Baidu know, the distributional difference value in Baidu's session, such as：Export Suser>10、Stieba>50、Siknow>50 or
Ssession>30 entity；
（2）Export at least one in Stieba, Siknow, Ssession and be more than entities of the 3 and Suser also greater than 3；
（3）Export the entity of Stieba, Siknow, Ssession all greater than 3；
（4）Export at least one in Stieba, Siknow, Ssession be more than 3, one be more than 8 entity.
Grader can also be set up according to above-mentioned screening rule, it is for instance possible to use prior art sets up grader
Method set up classification, the foundation of grader can improve efficiency.The foundation of grader can use prior art, herein no longer
Repeat.
It is above-mentioned that at least one method accuracy rate height for planting fructification is screened in multiple entities according to distributional difference value, still
Entity in the screening rule of setting below threshold value can not be called back, and be the association that this subsequently also needs to obtain kind of fructification
Entity.
The first incidence relation acquiring unit 310 is used to obtaining at least one respectively and plants fructification to having attribute to be associated
The first incidence relation between user's sample cluster.
Specifically, for example, can be described by matrix kind of fructification to attribute to be associated user's sample cluster it
Between the first incidence relation, such as matrix A.
Second incidence relation acquiring unit 320 is used for the associated entity for obtaining user's sample cluster with attribute to be associated
Group, and user's sample cluster with attribute to be associated is obtained to the second incidence relation between associated entity group.
Specifically, the associated entity group of user's sample cluster with attribute to be associated is obtained, if for example, with to be associated
User's sample cluster of attribute for the movie and television play entity of viewing " cure system " user, then obtain " curing system " movie and television play entity,
Other entities of " curing system " novel entity, " curing system " caricature entity or " curing system ", as with attribute to be associated
The associated entity group of user's sample cluster.
More specifically, user's sample cluster with attribute to be associated can be described by matrix between associated entity group
The second incidence relation, such as matrix B.
3rd incidence relation acquiring unit 330 is used to be obtained respectively at least according to the first incidence relation and the second incidence relation
Threeth incidence relation of one kind fructification to associated entity group.
Specifically, at least one can be for example obtained according to matrix A and matrix B and plants fructification to the of associated entity group
Three incidence relations, can be described with Matrix C., can be with for example, Matrix C can be got by simple matrix multiple
It is weighted after handling and is multiplied again.
Screening unit 340 is used to sieve each associated entity in associated entity group according to the 3rd incidence relation
Select to obtain the associated entity that at least one plants fructification.
For example, the 3rd incidence relation can be identified with Matrix C, each element in Matrix C is the entity seed to association
Degree of correlation information between entity, path phase of kind of the fructification to the path of each associated entity can be obtained according to the matrix
Like degree pathsim features, the associated entity of kind of fructification is obtained according to this feature.In addition, pathsim features can also be found
With entity peer objects, the influence of popular entity is reduced.Wherein, the calculation formula of Pathsim features is as follows：
Wherein, aiFor i-th of entity, ajFor j-th of entity, pcR(ai,aj) it is the i-th row, the element value of jth row in Matrix C
（That is entity aiWith entity ajBetween the degree of correlation）, pcR(ai,ai) it is the i-th row, the element value of the i-th row in Matrix C（That is entity ai
The degree of correlation of itself）, pcR-1(aj,ai) be Matrix C inverse matrix C-1The element value of middle jth row, the i-th row, pcR-1(aj,aj) be
The inverse matrix C of Matrix C-1The element value of middle jth row, jth row.
Filter the entity associated out of above-mentioned acquisition.Specifically, it can be filtered out with given threshold in the entity associated out
Doubtful incoherent entity, wherein, threshold value can be the multiple of distributional difference value of the seed entity on association user sample cluster,
Such as 2 times, 3 times or other multiples.
Thus, the 3rd incidence relation of acquisition has more directly reacted the associated entity for planting fructification, makes the association of acquisition
Entity is more accurate.
Relating module 400 is used to associate attribute to be associated and at least one kind fructification, at least one kind fructification
Entity is associated.
Specifically, after the associated entity for obtaining at least one kind fructification, by attribute to be associated and at least one seed
Entity, at least one associated entity for planting fructification are associated.
For example, after obtaining the novel entity of associated entity " curing system " or the movie and television play entity of other " curing system ", will
The movie and television play kind fructification of attribute to be associated " curing system " and " cure and be ", the movie and television play kind fructification of " curing system " associate reality
Body（The novel entity for " curing system " or the movie and television play entity of other " curing system "）It is associated.
Wherein, associated operation can be to attribute to be associated, at least one kind fructification, at least one kind fructification
Associated entity is labelled or the corresponding relation set up between them etc..For example, " can cure by attribute to be associated and be "
Movie and television play kind fructification, the associated entity of the movie and television play kind fructification of " cure and be " with " cure and be "（That " cures system " is small
Say the movie and television play entity of entity or other " curing system "）The corresponding relation for sticking the label of " curing system " or setting up between them
Deng.Thus, the 3rd incidence relation of acquisition has more directly reacted the associated entity for planting fructification, makes the associated entity of acquisition more
Plus it is accurate.
The excavation attribute and the device of entity associated relation of the embodiment of the present invention, distributional difference value acquiring unit is according to first
The distributional difference value that distribution proportion and the second distribution proportion are obtained can more embody the degree of association, and distributional difference value is more accurate；3rd
The 3rd incidence relation that incidence relation acquiring unit is obtained according to the first incidence relation and the second incidence relation more directly reacts
The associated entity of fructification is planted, makes the associated entity of acquisition more accurate；Thus, it is possible to excavate more accurately to be associated
Multiple associated entities of attribute, similarly realize the user-specific attributes excavated corresponding to entity（Attribute i.e. to be associated）, so as to carry
For more comprehensively, the detailed service of finer, more high-quality, for example, according to user-specific attributes to user's recommended entity；According to the dress
Any domain entities and user-specific attributes can also be excavated by putting（Attribute i.e. to be associated）Between incidence relation, not by application lead
The limitation in domain, is widely used.
It should be appreciated that each several part of the present invention can be realized with hardware, software, firmware or combinations thereof.Above-mentioned
In embodiment, the software that multiple steps or method can in memory and by suitable instruction execution system be performed with storage
Or firmware is realized.If, and in another embodiment, can be with well known in the art for example, realized with hardware
Any one of row technology or their combination are realized：With the logic gates for realizing logic function to data-signal
Discrete logic, the application specific integrated circuit with suitable combinational logic gate circuit, programmable gate array（PGA）, scene
Programmable gate array（FPGA）Deng.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
The description of example " or " some examples " etc. means to combine specific features, structure, material or the spy that the embodiment or example are described
Point is contained at least one embodiment of the present invention or example.In this manual, to the schematic representation of above-mentioned term not
Necessarily refer to identical embodiment or example.Moreover, specific features, structure, material or the feature of description can be any
One or more embodiments or example in combine in an appropriate manner.
Although an embodiment of the present invention has been shown and described, it will be understood by those skilled in the art that：Not
In the case of departing from the principle and objective of the present invention a variety of change, modification, replacement and modification can be carried out to these embodiments, this
The scope of invention is limited by claim and its equivalent.
Claims (6)
1. a kind of method for excavating attribute and entity associated relation, it is characterised in that comprise the following steps：
Obtain attribute to be associated；
At least one is obtained from multiple entities according to the attribute to be associated and plants fructification, wherein, it is described according to category to be associated
Property obtained from multiple entities at least one plant fructification specifically include：The multiple entity is obtained from default entity storehouse；From
The association user sample cluster with the attribute to be associated is obtained in overall user sample cluster；The multiple entity is obtained respectively to exist
Multiple distributional difference values in the association user sample cluster, wherein, the multiple use related to the multiple entity are obtained respectively
Multiple first distribution proportions of the family in the overall user sample cluster；The multiple use related to the multiple entity are obtained respectively
Second distribution proportion of the family in the association user sample cluster；According to the described second distribution proportion and the first distribution proportion
Obtain the distributional difference value；And the multiple entity is screened according to the multiple distributional difference value with described in obtaining
At least one plants fructification；And
The associated entity of at least one kind fructification is obtained, and the attribute to be associated and at least one described seed is real
Body, at least one described associated entity for planting fructification are associated.
2. according to the method described in claim 1, it is characterised in that the overall user sample cluster is multiple, is corresponded to respectively many
Individual network english teaching, then the corresponding distributional difference value of each entity be it is multiple, it is described according to multiple distributional difference values pair
The multiple entity is screened also to be included with obtaining at least one described kind fructification：
The multiple entity is screened according to default distributional difference value screening rule；Or,
Distributional difference value grader is created, and the multiple entity is screened according to the distributional difference value grader.
3. method according to any one of claim 1 to 2, it is characterised in that at least one kind fructification of the acquisition
Associated entity specifically include：
At least one described kind fructification is obtained respectively to close to first between user's sample cluster with the attribute to be associated
Connection relation；
The associated entity group of user's sample cluster with the attribute to be associated is obtained, and there is the category to be associated described in acquisition
Property user's sample cluster to the associated entity group between the second incidence relation；
At least one plants fructification described according to first incidence relation and second incidence relation are obtained respectively
The 3rd incidence relation of associated entity group；And
Each associated entity in associated entity group is screened to obtain according to the 3rd incidence relation
State at least one associated entity for planting fructification.
4. a kind of excavate attribute and the device of entity associated relation, it is characterised in that including：
Attribute acquisition module to be associated, for obtaining attribute to be associated；
Fructification acquisition module is planted, fructification is planted for obtaining at least one from multiple entities according to the attribute to be associated,
Wherein, described kind of fructification acquisition module includes：Entity acquiring unit, for obtaining the multiple reality from default entity storehouse
Body；Association user sample cluster acquiring unit, for obtaining the association with the attribute to be associated from overall user sample cluster
User's sample cluster；Distributional difference value acquiring unit, for obtaining the multiple entity respectively in the association user sample cluster
Multiple distributional difference values, wherein, the distributional difference value acquiring unit also particularly useful for：Obtain and the multiple entity respectively
Multiple first distribution proportions of the related multiple users in the overall user sample cluster, and obtain and the multiple reality respectively
Second distribution proportion of the multiple users that body phase is closed in the association user sample cluster, and according to the described second distribution proportion
The distributional difference value is obtained with the described first distribution proportion；And screening unit, for according to the multiple distributional difference value
The multiple entity is screened to obtain at least one described kind fructification；
Associated entity acquisition module, the associated entity for obtaining at least one kind fructification；And
Relating module, for by the attribute to be associated with it is described at least one plant fructification, it is described at least one plant fructification
Associated entity be associated.
5. device according to claim 4, it is characterised in that the overall user sample cluster is multiple, is corresponded to respectively many
Individual network english teaching, then the corresponding distributional difference value of each entity be it is multiple, it is described according to multiple distributional difference values pair
The multiple entity, which is screened, also to be included：
The multiple entity is screened according to default distributional difference value screening rule；Or,
Distributional difference value grader is created, and the multiple entity is screened according to the distributional difference value grader.
6. the device according to any one of claim 4 to 5, it is characterised in that the associated entity acquisition module includes：
First incidence relation acquiring unit, for obtaining at least one described kind fructification respectively to the attribute to be associated
User's sample cluster between the first incidence relation；
Second incidence relation acquiring unit, the associated entity group for obtaining user's sample cluster with the attribute to be associated,
And user's sample cluster with the attribute to be associated is obtained to the second incidence relation between associated entity group；
3rd incidence relation acquiring unit, for obtaining institute respectively according to first incidence relation and second incidence relation
State at least one and plant fructification to the 3rd incidence relation of associated entity group；And
Screening unit, for being carried out according to the 3rd incidence relation to each associated entity in associated entity group
Screen to obtain the associated entity of at least one kind fructification.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310714291.6A CN103631970B (en) | 2013-12-20 | 2013-12-20 | The method and apparatus for excavating attribute and entity associated relation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310714291.6A CN103631970B (en) | 2013-12-20 | 2013-12-20 | The method and apparatus for excavating attribute and entity associated relation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103631970A CN103631970A (en) | 2014-03-12 |
| CN103631970B true CN103631970B (en) | 2017-08-18 |
Family
ID=50213011
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310714291.6A Active CN103631970B (en) | 2013-12-20 | 2013-12-20 | The method and apparatus for excavating attribute and entity associated relation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103631970B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105224642B (en) * | 2015-09-25 | 2019-03-12 | 百度在线网络技术（北京）有限公司 | The abstracting method and device of entity tag |
| CN105760491B (en) * | 2016-02-18 | 2020-02-07 | 中国科学院信息工程研究所 | Data modeling method and device based on equipment function |
| CN107402933A (en) * | 2016-05-20 | 2017-11-28 | 富士通株式会社 | Entity polyphone disambiguation method and entity polyphone disambiguation equipment |
| CN107544992A (en) * | 2016-06-27 | 2018-01-05 | 阿里巴巴集团控股有限公司 | The method and apparatus of data analysis |
| CN108304493B (en) * | 2018-01-10 | 2020-06-12 | 深圳市腾讯计算机系统有限公司 | Hypernym mining method and device based on knowledge graph |
| CN108334632B (en) * | 2018-02-26 | 2021-03-23 | 深圳市腾讯计算机系统有限公司 | Entity recommendation method and device, computer equipment and computer-readable storage medium |
| CN110188148B (en) * | 2019-05-23 | 2021-02-02 | 北京建筑大学 | Entity identification method and device facing multimode heterogeneous characteristics |
| CN111047453A (en) * | 2019-12-04 | 2020-04-21 | 兰州交通大学 | Detection method and device for decomposing large-scale social network community based on high-order tensor |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101308493A (en) * | 2007-05-18 | 2008-11-19 | 亿览在线网络技术（北京）有限公司 | Entity relation exhibition method and system |
| CN102063433A (en) * | 2009-11-16 | 2011-05-18 | 华为技术有限公司 | Method and device for recommending related items |
| CN102915335A (en) * | 2012-09-17 | 2013-02-06 | 北京大学 | Information associating method based on user operation record and resource content |
| CN103425748A (en) * | 2013-07-19 | 2013-12-04 | 百度在线网络技术（北京）有限公司 | Method and device for mining document resource recommended words |
-
2013
- 2013-12-20 CN CN201310714291.6A patent/CN103631970B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101308493A (en) * | 2007-05-18 | 2008-11-19 | 亿览在线网络技术（北京）有限公司 | Entity relation exhibition method and system |
| CN102063433A (en) * | 2009-11-16 | 2011-05-18 | 华为技术有限公司 | Method and device for recommending related items |
| CN102915335A (en) * | 2012-09-17 | 2013-02-06 | 北京大学 | Information associating method based on user operation record and resource content |
| CN103425748A (en) * | 2013-07-19 | 2013-12-04 | 百度在线网络技术（北京）有限公司 | Method and device for mining document resource recommended words |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103631970A (en) | 2014-03-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103631970B (en) | The method and apparatus for excavating attribute and entity associated relation | |
| Luo et al. | Latticenet: Towards lightweight image super-resolution with lattice block | |
| CN104602042B (en) | Label setting method based on user behavior | |
| US9495282B2 (en) | Method and systems for a dashboard testing framework in an online demand service environment | |
| WO2008057178A2 (en) | Collecting votes in a decision model | |
| CN106055617A (en) | Data pushing method and device | |
| CN106527892A (en) | Screen capture method and system for electronic device | |
| Chies-Santos et al. | An optical/NIR survey of globular clusters in early-type galaxies-III. On the colour bimodality of globular cluster systems | |
| CN103914545B (en) | Search shows method and device | |
| CN110083739A (en) | For using the system and method apart from relevance hashing to media database addressing | |
| CN105260414B (en) | User behavior similarity calculation method and device | |
| CN103425661B (en) | A kind of website data is analyzed method and analyzes system | |
| CN110516815A (en) | The characteristic processing method, apparatus and electronic equipment of artificial intelligence recommended models | |
| CN110032597A (en) | The visible processing method and device of application program operation behavior | |
| CN106296498A (en) | Data processing method and device | |
| CN106886535A (en) | A kind of data pick-up method and apparatus for being adapted to multiple data sources | |
| CN108574669A (en) | User behavior tree constructing method and device | |
| CN108133078A (en) | Virtual model based on 720 distant view photographs is led the way quality management-control method | |
| CN108830376A (en) | For the multivalence value network depth intensified learning method of the environment of time-sensitive | |
| Šavrič et al. | A new pseudocylindrical equal-area projection for adaptive composite map projections | |
| CN106169961A (en) | The network parameter processing method and processing device of neutral net based on artificial intelligence | |
| CN109785279A (en) | A kind of image co-registration method for reconstructing based on deep learning | |
| AU2015309696A1 (en) | Analysing medical data | |
| Jones | Giving birth to next generation repositories | |
| JP2018133077A (en) | Method for recommending patent search keyword |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |