[summary of the invention]
In view of this, the present invention provides a kind of method and apparatus of information recommendation, in order to facilitate user at intelligent end
End carries out cross-thread operation.
Specific technical solution is as follows:
A kind of method of information recommendation obtains the resource information recommendation local as information point deposit from each application server
Resources bank;
Related information point is determined for the information point in the recommendation resources bank, to establish distributed content network;
According to the distributed content network, the related information point for the information point paid close attention to intelligent terminal recommended user.
A preferred embodiment according to the present invention determines related information point packet for the information point in the recommendation resources bank
It includes:
Establish the label and tag types of current targeted information point;
Relevant information point of each label under corresponding label type is retrieved respectively in the recommendation resources bank;
The relevant information point for meeting preset requirement is determined as to the related information point of current targeted information point.
The relevant information point for meeting preset requirement is determined as current targeted letter by a preferred embodiment according to the present invention
Cease the related information point of point, comprising:
Determine that the expectation of each relevant information point shows score, the expectation shows score by information quality score and use
Family feature score determines;
Expectation is showed into score meet the relevant information point of preset requirement and is determined as the association letter of current targeted information point
Breath point.
A preferred embodiment according to the present invention, the label and tag types packet for establishing current targeted information point
It includes:
Choosing user is the label and tag types that current targeted information point determines;Alternatively,
Corresponding search result, chooses search when obtaining using the descriptor of current targeted information point as search key
As a result descriptor is as label, and the type of search result or the attribute of search result descriptor are as tag types.
A preferred embodiment according to the present invention, renewal time of the information quality score by relevant information point, information
At least one of pageview and label temperature determine.
A preferred embodiment according to the present invention, the user characteristics score by relevant information point and user property matching
Degree, relevant information point and the matching degree of user preferences resource type and the matching of relevant information point and user preferences label
At least one of degree determines.
The information point of a preferred embodiment according to the present invention, user's concern includes: that the user is browsing
Information point.
A preferred embodiment according to the present invention, this method further include: with card on the interface that the user is browsing
The construction module of sheet form shows the related information point of recommendation.
A preferred embodiment according to the present invention, this method further include: if selected in the related information point of recommendation
One information point then jumps to the corresponding application of selected information point and shows the content of selected information point.
The present invention also provides a kind of device of information recommendation, which includes:
Resource consolidation module, for obtaining the resource information recommendation money local as information point deposit from each application server
Source library;
Network establishes module, for determining related information point for the information point in the recommendation resources bank, to establish
Distributed content network;
Information recommendation module, the information for being paid close attention to intelligent terminal recommended user according to the distributed content network
The related information point of point.
A preferred embodiment according to the present invention, the network are established module and are specifically included:
Label setting up submodule, for establishing the label and tag types of current targeted information point;
Label search submodule, for retrieving the label respectively under corresponding label type in the recommendation resources bank
Relevant information point;
Information sifting unit, for the relevant information point for meeting preset requirement to be determined as to the pass of current targeted information point
Join information point.
A preferred embodiment according to the present invention, the information sifting unit specifically include:
Score determination unit, for determining that the expectation of each relevant information point shows score, the expectation shows score
It is determined by information quality score and user characteristics score;
Information sifting unit meets the relevant information point of preset requirement for expectation to be showed score and is determined as current institute's needle
To the related information point of information point.
A preferred embodiment according to the present invention, the label establish subelement, are specifically used for choosing user being current institute
The label and tag types determined for information point;Made alternatively, being specifically used for obtaining with the descriptor of current targeted information point
Corresponding search result when for search key, chooses the descriptor of search result as label, the type of search result or searches
The attribute of hitch fruit descriptor is as tag types.
A preferred embodiment according to the present invention, renewal time of the information quality score by relevant information point, information
At least one of pageview and label temperature determine.
A preferred embodiment according to the present invention, the user characteristics score by relevant information point and user property matching
Degree, relevant information point and the matching degree of user preferences resource type and the matching of relevant information point and user preferences label
At least one of degree determines.
The information point of a preferred embodiment according to the present invention, user's concern includes: that the user is browsing
Information point.
A preferred embodiment according to the present invention, the related information point that the information recommendation module is recommended is with card form
Construction module shows on the interface that the user is browsing.
As can be seen from the above technical solutions, through the invention, resource information is obtained as information from each application server
Point can be realized and recommend to the intersection that user carries out resource information, i.e., resource information recommended to the user is not limited to user pass
The resource information of note can also include the resource information from other application server, and on this basis, user can directly select
The resource information from multiple application servers is selected and browses, without entering back into corresponding application lookup by hand after returning to desktop
Want the resource information seen, user is facilitated to carry out cross-thread operation in intelligent terminal.
[specific embodiment]
To make the objectives, technical solutions, and advantages of the present invention clearer, right in the following with reference to the drawings and specific embodiments
The present invention is described in detail.
Realization of the invention can be based on system as shown in Figure 2, and resource consolidation server is this hair within the system
Bright core, there are interfaces between resource consolidation server and each application server, obtain resource information from each application server
(resource information may include the various resources in application server, for example, resource of the application server for recommendation) deposit
Local recommendation resources bank.The application provider for needing to carry out staggered form resource recommendation opens itself to resource consolidation server
Application server interface, the resource for being used to carry out staggered form resource recommendation is supplied to resource consolidation server.So-called friendship
V shape resource recommendation refers to realizing the resource recommendation between different application, such as recommends news category information in Video Applications.
In addition, the recommendation resources bank can be also used for carrying out desktop resource recommendation or screen locking resource recommendation etc..
Dock formula desktop layouts have existed relatively long, and people start slowly to be weary of this content structure, in light of this situation,
The construction module of card form is proposed, i.e., each data block is showed on the interface of intelligent terminal with the structure of card form, example
As shown in Figure 3.Usual a type of resource information is presented in the module of a card form, in card form
It may include the submodule of multiple card forms to show each resource information again in module.This structure carries out resource information
Good integration shows, and user is facilitated to browse and operation.If being applied to this application scenarios, above-mentioned resource consolidation clothes
Business device can be set in card form Resource Server, and so-called card form Resource Server is referred to for pushing to intelligent terminal
The resource information showed with the construction module of card form.Certainly, method and apparatus provided by the invention can also be not limited to this
Kind application scenarios.
Core of the invention thought essentially consists in, and is being browsed to user's recommendation with user by resource consolidation server
The associated resource information of resource information, the resource information of recommendation be not limited to corresponding to the resource information that is browsing
Type, associated resource information is probably exactly the correlated resources information that user wants to see here, and user can be direct
Select and browse the resource information of recommendation.Realize that the method and apparatus that above-mentioned resource information is recommended are real in resource consolidation server end
It is existing, for the convenience of description, in embodiments of the present invention, each resource information is known as information point, wherein information point can pass through money
The descriptor of source information, the summary of resource information or the keyword extracted from resource information etc. indicate.Involved in the present invention
" information point " can jump to the corresponding application of the information point by click " information point " with corresponding association, i.e. user,
And the content that can further load the information point in this application, that is, jump to the corresponding application of the information point and presentation should
The corresponding content of information point.
Fig. 4 is the method flow diagram of information recommendation provided in an embodiment of the present invention, and as shown in Figure 4, main includes two
Stage: it establishes in step 401 in the stage and step 402 of distributed content network and is closed according to distributed content network recommendation user
The stage of the related information point of the information point of note.
Wherein distributed content network is the network being made of the information point there are incidence relation, these information points derive from
Recommend resources bank, that is to say, that after determining the related information point for recommending the information point in resources bank, recommend the information in resources bank
Point just constitutes distributed content network.The distributed content network, which can be, to be pre-established, and is also possible to browse in user
It is established in real time when the resource information of certain application.It is carried out first below referring to establishment process of the Fig. 5 to distributed content network detailed
Thin description.
As shown in Figure 5, resource consolidation server for recommend resources bank in information point execute respectively following steps come
Determine the related information point of information point, wherein current targeted information point is known as current information point:
Step 501: establishing the label and tag types of current information point.
In this step, for current information point establish label can be can express current information point theme or with work as
The relevant keyword of preceding information point, by taking " 4 trailer of transformer " as an example, label and tag types can include:
[video] transformer 1
[video] transformer 2
[video] transformer 3
[director] Michael shellfish
[performer] mark Wall Burger
[performer] Li Bingbing
[caricature] man in a responsible position
[caricature] hornet
[original sound tape] Battle Cry
[automobile] Chevrolet Corvette
[automobile] passes auspicious GS5
[automobile] Ferrari 458SPECIALEE
Wherein the content in [] is tag types, and the content after [] is label.
Label and tag types establish mode and can use but be not limited to following two:
First way: establishing the mode of label by hand, i.e., manually stamps for current information point related to the information point
Label and tag types.This mode is easier, but compares labor intensive.
The second way: corresponding search result when obtaining using the descriptor of current information point as search key, choosing
Take the descriptor of search result as label, the type of search result or the attribute of search result descriptor are as tag types.
For example, the descriptor of " 4 trailer of transformer " is " transformer 4 ", scanned for as search key, selection is searched
Several search result, extracts the descriptor and label of these search results, it is assumed that in search result before coming in hitch fruit
Several search result before coming are as follows: watch the video web-pages of transformer 1, caricature webpage, the original sound tape of man in a responsible position
" BattleCry's " listens song webpage etc. online, then can therefrom extract descriptor " transformer 1 ", " man in a responsible position ",
" Battle Cry " etc. is used as label, and corresponding tag types are respectively " video ", " caricature ", " original sound tape " etc..
Of course, it is possible to be used simultaneously using any one of both the above mode or both the above mode.
Step 502: retrieving relevant information point of each label under corresponding label type respectively in recommending resources bank.
Information point that each label is hit under corresponding label type is determined in retrieval as relevant information point, hit is tactful
It can be using calculating between label and information point by the way of similarity, wherein similarity meets the information point of preset requirement for selection,
Such as similarity is greater than before preset threshold or similarity come several.Each label is when hitting information point herein, preferably
Keep the tag types of the label consistent with the type of the information point of hit.For example, label " transformer 1 ", " Michael
The relevant information point difference that shellfish ", " Battle Cry " and " Chevrolet Corvette " are obtained is as shown in table 1:
Table 1
Step 503: determining that the expectation of each relevant information point shows score, which shows score and obtained by information quality
Divide and user characteristics score determines.
The expectation of relevant information point, which shows score and embodies user, it is expected that the relevant information point is associated the possibility showed
Property.Wherein information quality score embodies the quality condition of the relevant information point itself, can be clear by information update time, information
At least one of the amount of looking at, label temperature etc. determine.Preferably, information quality score can be by information update time score, letter
Breath pageview score, label temperature score etc. obtain after being weighted.Wherein information update time score can be by correlation
The renewal time of information point determines that renewal time is closer, and score is higher, otherwise information update time more long score is lower.Information is clear
It lookes to measure and point can determine that pageview is higher, and score is higher by the pageview of relevant information point, otherwise the lower score of pageview is more
It is low.Label temperature score can the volumes of searches of the label as corresponding to relevant information point determine that the volumes of searches of corresponding label is higher
Score is higher, otherwise the lower score of volumes of searches of corresponding label is lower.
User characteristics score embodies the matching degree of the relevant information point and user, can be by relevant information point and user
Matching degree, relevant information point and the user preferences mark of the matching degree of attribute, relevant information point and user preferences resource type
At least one of matching degree of label etc. determines.Preferably, user characteristics score can be by these above-mentioned matching degree scores
It is obtained after being weighted.Wherein, the matching degree of relevant information point and user property can be by the user couple of the attribute
The browsing frequency or duration of the relevant information point determine, it is assumed for example that the age of certain user is 25 years old, and gender is female, then 20 years old
Women to 30 years old this age level just can determine the relevant information point to the browsing frequency or duration of the relevant information point
With the matching degree of user property, user property here is not limited to age and gender certainly, can also be such as region, learns
It goes through, occupation etc..The relevant information point that the matching degree of relevant information point and user preferences resource type was browsed by user
The frequency or duration of the resource type determine that the frequency is more or duration is longer, then matching degree is higher.Relevant information point with
The matching degree of user preferences label browsed the browsing frequency of label corresponding to the relevant information point as user or duration determines,
The frequency is more or duration is longer, then matching degree is higher.
Then after the information quality score of relevant information point and user characteristics score being weighted, phase is obtained
The expectation for closing information point shows score.
Step 504: expectation being showed into score meet the relevant information point of preset requirement and be determined as the association of current information point
Information point.
After the step 503, the multiple relevant information points for determining current information point, in step 504 further from phase
It closes in information point and filters out related information point.Such as expectation can be showed into the relevant information that score is greater than or equal to preset threshold
Point is determined as the related information point of current information point, or will it is expected that the relevant information point for showing preset quantity before score comes is true
It is set to the related information point of current information point.
In above-mentioned steps 503 and step 504, score is showed by the expectation of determining relevant information point to select current information
The related information point of point, which are only a kind of implementation for providing of the present invention, can be with based on inventive concept of the invention
The related information point of current information point is selected from the relevant information point that step 502 retrieves by various other ways, for example,
It can choose the related information point of all relevant information points that step 502 retrieves as current information point, can also set pair
The requirement of related information point determines relevant information point by the sets requirement, such as sets the upper limit of the number of relevant information point, or
Person sets tag types limitation etc., will not enumerate herein.
After above-mentioned steps, the related information point of current information point, i.e., the information network of one information point are determined.And
There are related information points for each information point in recommendation resources bank, and which forms a network structure, that is, distributed content nets
Network.It should be noted that the stage for establishing distributed content network can be lasting progress, with resource in recommendation resources bank
The update of information, above-mentioned establishment process re-execute, and distributed content network also updates therewith.
Resource information in this way on intelligent terminal can no longer be top-down Directory tree structure, but a distribution
Formula network structure, such as shown in fig. 6, the node in Fig. 6 indicates information point, arrow, which is directed toward, indicates that incidence relation, i.e. arrow refer to
To related information point, it should be noted that Fig. 6 has only intercepted a segment of distributed content network.In above-mentioned distribution
Content network can be through distributed content network, can also be related to user, for example, the distributed content network can with
The preference at family is related, the corresponding distributed content network of a user.
For intelligent terminal user establish distributed content network after, when user is browsing some information point, hold
Row step 402 recommends the pass of the information point currently browsed in distributed content network on the interface of the information point currently browsed
Join information point.Wherein, step 402 actively executes after can browsing some information point detecting user, can also be touched by user
It sends out and executes, for example, user clicks correlation function button, or stopping in the information point is more than setting duration etc..
Since the related information point of information point in distributed content network is not limited to the information point of same type, this is just real
Staggered form resource recommendation is showed.Such as user is browsing the positive of transformer 4, can recommend to be associated on same interface
Information point " 4 trailer of transformer ", " 4 film ticket of transformer ", " Li Bingbing become 4 moulding ", " Chevrolet Corvette " and
" passing auspicious GS5 " etc..If user wants to check the relevant information of Chevrolet, " Chevrolet Corvette " can choose, at this time
The corresponding application of relevant information point, such as browser are just automatic jumped to, which directly shows the correlation of user's selection
The content of information point.After user need not just return to desktop, the relevant information that browser finds Chevrolet by hand is reentered.
In addition, can also show relevant information point with the construction module of card form when showing recommendation resource, such as
Same type of relevant information point is presented on the construction module of a card form, it again can in the module of same card form
Show relevant information point with the submodule comprising multiple card forms.User, can when selecting relevant information point to browse
Be not particularly limited herein using the selection mode of such as construction module of click, sliding or the corresponding card form of long-pressing.
By the process of above-mentioned Fig. 4 and Fig. 5, resource consolidation server is described in detail in the present invention to user recommended user
The process of the associated resource information point of the resource information browsed.In the present invention, by resource consolidation server to user
Resource information associated with the resource information that user is browsing that the resource information of recommendation is not limited to, can be any user
The associated resource information of the resource information of concern, for example, user concern resource information may include: it is customized oneself
Interested resource information, resource information of application etc. from user's concern can actively or be triggered based on user to user
Recommended user concern information point and related information point, for example, by recommendation show desktop or locking screen interface or specifically
Software interface.Detailed recommended flowsheet and process described in above-mentioned Fig. 4 and Fig. 5 are almost the same, are not described in detail herein.
It is method flow provided in an embodiment of the present invention above, corresponding device is as shown in Figure 7, which is set to
The resource consolidation server end stated may include: that resource consolidation module 00, network establish module 10 and information recommendation module 20.
Resource consolidation module 00 is responsible for obtaining the recommendation for recommending resource local as information point deposit from each application server
Resources bank.Here there are interface between resource consolidation server and each application server, need to carry out staggered form resource recommendation
Application provider opens the interface of the application server of itself to resource consolidation server, will be used to carry out staggered form resource and push away
The resource recommended is supplied to resource consolidation server.
Network is established module 10 and is responsible for for recommending the information point in resources bank to determine related information point, to establish distribution
Formula content network.Wherein distributed content network is the network being made of the information point there are incidence relation, these information points are come
Derived from recommendation resources bank, that is to say, that after determining the related information point for recommending each information point in resources bank, recommend in resources bank
Information point just constitute distributed content network.
Specifically, network, which establishes module 10, can specifically include label setting up submodule 11,12 and of label search submodule
Information sifting submodule 13.
Wherein, label setting up submodule 11 is responsible for establishing the label and tag types of current targeted information point.Here it builds
Vertical label is can to express current targeted information point theme or keyword relevant to currently targeted information point.Label
Establishing subelement 11 can be at least one of specifically to use but be not limited in a manner of following two:
First way: choosing user is the label and tag types that current targeted information point determines.
The second way: corresponding search knot when obtaining using the descriptor of current targeted information point as search key
Fruit chooses the descriptor of search result as label, and the type of search result or the attribute of search result descriptor are as label
Type.
Then label search submodule 12 distinguishes correlation of the Checking label under corresponding label type in recommending resources bank
Information point.Information point that each label is hit under corresponding label type is determined in retrieval as relevant information point, hit is tactful
It can be using calculating between label and information point by the way of similarity, wherein similarity meets the information point of preset requirement for selection,
Such as similarity is greater than before preset threshold or similarity come several.Each label is when hitting information point herein, preferably
Keep the tag types of the label consistent with the type of the information point of hit.
Information sifting submodule 13 is responsible for for the relevant information point for meeting preset requirement being determined as current targeted information point
Related information point.Specifically, which may include: score determination unit 131 and information sifting unit
132。
Score determination unit 131 be responsible for determine each relevant information point expectation show score, it is expected that show score by
Information quality score and user characteristics score determine.The expectation of relevant information point show score embody user it is expected the correlation believe
Breath point is associated a possibility that showing.Wherein information quality score embodies the quality condition of the relevant information point itself, can
To be determined by least one of information update time, information browse amount, label temperature etc..Preferably, information quality score can
To be measured point by information update time score, information browse, label temperature score etc. obtains after being weighted.Wherein information
Renewal time score can determine that renewal time is closer, and score is higher by the renewal time of relevant information point, otherwise information update
Time is more long, and score is lower.Information browse is measured and point can be determined by the pageview of relevant information point, and the higher score of pageview is more
Height, on the contrary the lower score of pageview is lower.Label temperature score can the volumes of searches of the label as corresponding to relevant information point determine,
The volumes of searches of corresponding label is higher, and score is higher, otherwise the lower score of volumes of searches of corresponding label is lower.
User characteristics score embodies the matching degree of the relevant information point and user, can be by relevant information point and user
Matching degree, relevant information point and the user preferences mark of the matching degree of attribute, relevant information point and user preferences resource type
At least one of matching degree of label etc. determines.It preferably, can be by these above-mentioned matching degrees with user characteristics score
Score obtains after being weighted.Wherein, the matching degree of relevant information point and user property can be by the use of the attribute
Family determines the browsing frequency or duration of the relevant information point, it is assumed for example that the age of certain user is 25 years old, and gender is female, then
20 years old women to 30 years old this age level just can determine that the correlation is believed to the browsing frequency or duration of the relevant information point
Matching degree of the breath point with user property, user property here is not limited to age and gender certainly, can also be such as
Domain, educational background, occupation etc..The related letter that relevant information point was browsed to the matching degree of user preferences resource type by user
The frequency or duration of the breath point resource type determine that the frequency is more or duration is longer, then matching degree is higher.Relevant information
Put the browsing frequency or duration for browsing label corresponding to the relevant information point as user with the matching degree of user preferences label
It determines, the frequency is more or duration is longer, then matching degree is higher.
Then after the information quality score of relevant information point and user characteristics score being weighted, phase is obtained
The expectation for closing information point shows score.
Information sifting unit 132, which is responsible for expectation showing score, to be met the relevant information point of preset requirement and is determined as current institute
For the related information point of information point.Information sifting unit 132 is actually further to filter out association from relevant information point
Information point.Such as expectation can be showed into score and be determined as current information point more than or equal to the relevant information point of preset threshold
Related information point, or will it is expected that the relevant information point for showing preset quantity before score comes is determined as the association of current information point
Information point.
Network establishes the related information point that module 10 determines that out current information point, i.e., the Information Network of one information point in this way
Network.And recommend in resources bank each information point there are related information points, it is distribution which forms a network structure
Content network.It should be noted that network establishes module 10 and establishes stage of distributed content network and can be lasting progress,
With the update for recommending resource information in resources bank, above-mentioned establishment process is re-executed, and distributed content network also updates therewith.
Information recommendation module 20 can be according to distributed content network, to the letter of intelligent terminal recommended user concern later
Cease the related information point of point.Since the related information point of information point in distributed content network is not limited to the information of same type
Point, which achieves staggered form resource recommendations.Preferably, the related information point that information recommendation module is recommended is with the knot of card form
Structure module shows on the interface that user is browsing.User, can be using such as when selecting relevant information point to browse
The selection mode of the construction module of click, sliding or the corresponding card form of long-pressing, is not particularly limited herein.
The information displaying that the present invention recommends can apply whole in the intelligence such as tablet computer, smart phone, smart television
On end.
In several embodiments provided by the present invention, it should be understood that disclosed device and method, it can be by other
Mode realize.For example, the apparatus embodiments described above are merely exemplary, for example, the division of the unit, only
For a kind of logical function partition, there may be another division manner in actual implementation.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of hardware adds SFU software functional unit.
The above-mentioned integrated unit being realized in the form of SFU software functional unit can store and computer-readable deposit at one
In storage media.Above-mentioned SFU software functional unit is stored in a storage medium, including some instructions are used so that a computer
It is each that equipment (can be personal computer, server or the network equipment etc.) or processor (processor) execute the present invention
The part steps of embodiment the method.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (Read-
Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic or disk etc. it is various
It can store the medium of program code.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Within mind and principle, any modification, equivalent substitution, improvement and etc. done be should be included within the scope of the present invention.