CN107562845A - Method for pushing, system and electronic equipment - Google Patents
Method for pushing, system and electronic equipment Download PDFInfo
- Publication number
- CN107562845A CN107562845A CN201710747528.9A CN201710747528A CN107562845A CN 107562845 A CN107562845 A CN 107562845A CN 201710747528 A CN201710747528 A CN 201710747528A CN 107562845 A CN107562845 A CN 107562845A
- Authority
- CN
- China
- Prior art keywords
- user
- information
- module
- neutral net
- pushing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
Present disclose provides a kind of method for pushing, for the humping section object from multiple objects, methods described includes, obtain the image information of the multiple object, pass through neutral net, the characteristics of image of each object is extracted from the image information of the multiple object, and based on described image feature, pushes the partial objects in the multiple object.
Description
Technical field
This disclosure relates to Internet technical field, more particularly, to a kind of method for pushing, system and electronic equipment.
Background technology
With the continuous development of information technology, also increasing fast, user need to spend largely at user option object
Time can just find the object of oneself needs.This process for browsing a large amount of irrelevant informations can undoubtedly reduce the experience of user.Push away
The system of recommending the object interested to user recommended user, can be widely used in many fields according to the characteristics of user.
During present inventive concept is realized, inventor has found that at least there are the following problems in the prior art：It is existing
Method for pushing is normally based on the data such as user browsing behavior, under many circumstances, when some new contents occur, due to lacking
These behavioral datas, therefore effectively new content can not be recommended, that is, the problem of cold start-up difficulty be present.
The content of the invention
In view of this, present disclose provides a kind of method for pushing, system and electronic equipment for cold start-up problem.
An aspect of this disclosure provides a kind of method for pushing, described for the humping section object from multiple objects
Method includes, and obtains the image information of the multiple object, by neutral net, is carried from the image information of the multiple object
The characteristics of image of each object is taken, and based on described image feature, pushes the partial objects in the multiple object.
In accordance with an embodiment of the present disclosure, it is described to be based on described image feature, push the partial objects in the multiple object
Including according to user's history behavior, it is determined that the image information of at least one object related to the user, passes through nerve net
Network, determine that each object in the multiple object is similar to the first of at least one object related to the user
Degree, and based on first similarity, push the partial objects in the multiple object.
In accordance with an embodiment of the present disclosure, it is described to be based on described image feature, push the partial objects in the multiple object
Including according to user's current behavior, determining the image information of existing object, by neutral net, determining in the multiple object
Each object and the existing object the second similarity, and based on second similarity, push the multiple object
In partial objects.
In accordance with an embodiment of the present disclosure, the partial objects in the multiple object of the push include, from the multiple right
As middle determination partial objects are as first set, the attribute information of each object in the first set, the row of user
And/or, the attribute information of user, by neutral net, the recommended priority of each object, root are obtained for information
According to the recommended priority, part or all of object is determined from the first set as second set, and described in push
Object in second set.
In accordance with an embodiment of the present disclosure, the attribute information of each object in the first set, user
Behavioural information and/or, the attribute information of user, by neutral net, obtain the recommended priority bag of each object
Include, train the neutral net using training set, the training set includes user behavior data, and the recommended priority is user
The probability that behavior occurs.
In accordance with an embodiment of the present disclosure, the neutral net includes deep neural network part and logistic regression part, institute
State the attribute information of each object in the first set, the behavioural information of user and/or, the attribute of user
Information, by neutral net, obtaining the recommended priority of each object includes, according to customer attribute information, user behavior
Information and object properties information, by the deep neural network part, the first result is obtained, according to user behavior information,
And object properties information, by the logistic regression part, obtain the second result, and processing first result and described
Second result, obtain the recommended priority of each object.
Another aspect of the present disclosure provides a kind of supplying system, and the system includes, image collection module, for obtaining
The image information of the multiple object, characteristic extracting module, for by neutral net, from the image information of the multiple object
The characteristics of image of the middle each object of extraction, and pushing module, for based on described image feature, pushing in the multiple object
Partial objects.
In accordance with an embodiment of the present disclosure, the pushing module includes, the first determination sub-module, for according to user's history row
For it is determined that the image information of at least one object related to the user, the first comparison sub-module, for passing through nerve net
Network, determine that each object in the multiple object is similar to the first of at least one object related to the user
Degree, and the first push submodule, for based on first similarity, pushing the partial objects in the multiple object.
In accordance with an embodiment of the present disclosure, the pushing module includes, the second determination sub-module, for according to user's current line
To determine the image information of existing object, the second comparison sub-module, for by neutral net, determining in the multiple object
Each object and the existing object the second similarity, and the second push submodule, for similar based on described second
Degree, pushes the partial objects in the multiple object.
In accordance with an embodiment of the present disclosure, the pushing module includes, the 3rd determination sub-module, for from the multiple object
Middle determination partial objects are as first set, priority determination sub-module, for each object in the first set
Attribute information, user behavioural information and/or, the attribute information of user, by neutral net, obtain described each
The recommended priority of object, the 4th determination sub-module, for according to the recommended priority, the determining section from the first set
Divide or whole objects are as second set, and the 3rd push submodule, for pushing the object in the second set.
In accordance with an embodiment of the present disclosure, the priority determination sub-module includes, training unit, for using training training
Practice the neutral net, the training set includes user behavior data, and the recommended priority is the probability that user behavior occurs.
In accordance with an embodiment of the present disclosure, the neutral net includes deep neural network part and logistic regression part, institute
Stating priority determination sub-module includes, first processing units, for according to customer attribute information, user behavior information and right
As attribute information, by the deep neural network part, the first result, second processing unit, for according to user's row are obtained
For information, and object properties information, by the logistic regression part, the second result, and the 3rd processing unit are obtained, is used
In handling first result and second result, the recommended priority of each object is obtained.
Another aspect of the present disclosure provides a kind of electronic equipment, including, one or more processors, and storage dress
Put, for storing one or more programs, wherein, when one or more of programs are by one or more of computing devices
When so that one or more of computing device any one methods as described above.
Another aspect of the present disclosure provides a kind of computer-readable medium, is stored thereon with executable instruction, the finger
Order makes computing device any one method as described above when being executed by processor.
Another aspect of the present disclosure provides a kind of computer program, and the computer program includes the executable finger of computer
Order, the instruction are used to realize method as described above when executed.
In accordance with an embodiment of the present disclosure, can solve the problems, such as cold start-up during pushing at least in part, pass through nerve
Network, the characteristics of image of learning object, it can preferably push emerging object.
Brief description of the drawings
By the description to the embodiment of the present disclosure referring to the drawings, the above-mentioned and other purposes of the disclosure, feature and
Advantage will be apparent from, in the accompanying drawings：
Fig. 1 diagrammatically illustrates can apply showing for method for pushing, system and electronic equipment according to the embodiment of the present disclosure
Example sexual system framework；
Fig. 2 diagrammatically illustrates the flow chart of the method for pushing according to the embodiment of the present disclosure；
Fig. 3 A are diagrammatically illustrated according to the embodiment of the present disclosure based on described image feature, are pushed in the multiple object
Partial objects flow chart；
Fig. 3 B diagrammatically illustrate the multiple right based on described image feature, push according to another embodiment of the disclosure
The flow chart of partial objects as in；
Fig. 4 diagrammatically illustrates the flow of the partial objects in the multiple object of push according to the embodiment of the present disclosure
Figure；
Fig. 5 diagrammatically illustrates to be believed according to the attribute of each object in the first set of the embodiment of the present disclosure
Breath, user behavioural information and/or, the attribute information of user, by neutral net, obtain pushing away for each object
Recommend the flow chart of priority；
Fig. 6 diagrammatically illustrates the block diagram of the supplying system according to the embodiment of the present disclosure；
Fig. 7 A diagrammatically illustrate the block diagram of the pushing module according to the embodiment of the present disclosure；
Fig. 7 B diagrammatically illustrate the block diagram of the pushing module according to another embodiment of the disclosure；
Fig. 8 diagrammatically illustrates the block diagram of the pushing module according to another embodiment of the disclosure；
Fig. 9 diagrammatically illustrates the block diagram of the priority determination sub-module according to the embodiment of the present disclosure；
Figure 10 diagrammatically illustrates the block diagram of the priority determination sub-module according to another embodiment of the disclosure；
Figure 11 diagrammatically illustrates the department of computer science for being adapted for carrying out method for pushing and/or system according to the embodiment of the present disclosure
The block diagram of system.
Embodiment
Hereinafter, it will be described with reference to the accompanying drawings embodiment of the disclosure.However, it should be understood that these descriptions are simply exemplary
, and it is not intended to limit the scope of the present disclosure.In addition, in the following description, the description to known features and technology is eliminated, with
Avoid unnecessarily obscuring the concept of the disclosure.
Term as used herein is not intended to limit the disclosure just for the sake of description specific embodiment.Used here as
Word " one ", " one (kind) " and "the" etc. should also include " multiple ", the meaning of " a variety of ", unless context clearly refers in addition
Go out.In addition, term " comprising " as used herein, "comprising" etc. indicate the presence of the feature, step, operation and/or part,
But it is not excluded that in the presence of or other one or more features of addition, step, operation or parts.
All terms (including technology and scientific terminology) as used herein have what those skilled in the art were generally understood
Implication, unless otherwise defined.It should be noted that term used herein should be interpreted that with consistent with the context of this specification
Implication, without should by idealization or it is excessively mechanical in a manner of explain.
, in general should be according to this using in the case of being similar to that " in A, B and C etc. at least one " is such and stating
Art personnel are generally understood that the implication of the statement to make an explanation (for example, " having system at least one in A, B and C "
Should include but is not limited to individually with A, individually with B, individually with C, with A and B, with A and C, with B and C, and/or
System with A, B, C etc.).Using in the case of being similar to that " in A, B or C etc. at least one " is such and stating, it is general come
Say be generally understood that the implication of the statement to make an explanation (for example, " having in A, B or C at least according to those skilled in the art
The system of one " should include but is not limited to individually with A, individually with B, individually with C, with A and B, with A and C, with
B and C, and/or system etc. with A, B, C).It should also be understood by those skilled in the art that substantially arbitrarily represent two or more
The adversative conjunction and/or phrase of optional project, either in specification, claims or accompanying drawing, shall be construed as
Give including one of these projects, the possibility of these projects either one or two projects.For example, " A or B " should for phrase
It is understood to include " A " or " B " or " A and B " possibility.
Fig. 1 diagrammatically illustrates can apply showing for method for pushing, system and electronic equipment according to the embodiment of the present disclosure
Example sexual system framework 100.
As shown in figure 1, terminal device 101,102,103, network can be included according to the system architecture 100 of the embodiment
104 and server 105.Network 104 is to the offer communication link between terminal device 101,102,103 and server 105
Medium.Network 104 can include various connection types, such as wired, wireless communication link or fiber optic cables etc..
User can be interacted with using terminal equipment 101,102,103 by network 104 with server 105, to receive or send out
Send message etc..Various telecommunication customer end applications, such as the application of shopping class, net can be installed on terminal device 101,102,103
(merely illustrative) such as the application of page browsing device, searching class application, JICQ, mailbox client, social platform softwares.
Terminal device 101,102,103 can have a display screen and a various electronic equipments that supported web page browses, bag
Include but be not limited to smart mobile phone, tablet personal computer, pocket computer on knee and desktop computer etc..
Server 105 can be to provide the server of various services, such as utilize terminal device 101,102,103 to user
The website browsed provides the back-stage management server (merely illustrative) supported.Back-stage management server can be to the use that receives
The data such as family request analyze etc. processing, and by result (such as according to user's acquisition request or the webpage of generation, believe
Breath or data etc.) feed back to terminal device.
It should be noted that the method for pushing that the embodiment of the present disclosure is provided can typically be performed by server 105.Accordingly
Ground, the supplying system that the embodiment of the present disclosure is provided can be typically arranged in server 105.What the embodiment of the present disclosure was provided
Method for pushing can also be by can communicate different from server 105 and with terminal device 101,102,103 and/or server 105
Terminal device perform.Correspondingly, the supplying system that the embodiment of the present disclosure is provided can also be arranged at different from server 105
And in the terminal device that can be communicated with terminal device 101,102,103 and/or server 105.
It should be understood that the number of the terminal device, network and server in Fig. 1 is only schematical.According to realizing need
Will, can have any number of terminal device, network and server.
Fig. 2 diagrammatically illustrates the flow chart of the method for pushing according to the embodiment of the present disclosure.
As shown in Fig. 2 this method includes operation S210~S230.
S210 is being operated, is obtaining the image information of the multiple object.
In operation S220, by neutral net, the image of each object is extracted from the image information of the multiple object
Feature.
In operation S230, based on described image feature, the partial objects in the multiple object are pushed.
This method can utilize what is enriched in internet well by using the image information of Processing with Neural Network object
Image information is recommended so that recommendation results are more personalized, and do not need the process of complicated artificial extraction feature, together
When, preferably new object can be recommended.
According to the embodiment of the present disclosure, before this method is implemented, the behavioral data of user can also be obtained, and user and
Each attribute of object, including the source of object, the classification of object, the preference etc. of user.In addition, it can also enter as needed
Row data scrubbing and standardization, for example, being user's unique mark by various user's identity maps, and remove some abnormal datas
Deng.
According to the embodiment of the present disclosure, ob-ject provider would generally provide the image of object, therefore, can be with operation S210
Get the image information of object.
According to the embodiment of the present disclosure, for extracting the characteristics of image of each object from the image information of the multiple object
Neutral net can be convolutional neural networks (Convolutional Neural Networks- abbreviation CNN), the network is kept away
The pretreatment complicated early stage to image is exempted from, original image can be directly inputted.The neutral net for example can include input layer,
The parts such as convolutional layer, pond layer, full articulamentum and output layer.Come original image as input, convolutional layer and pond layer first
Learn the local space structure and texture information in input picture, local message is then convergeed into full articulamentum, full articulamentum
Then learn the global information contained in whole image being more abstracted, the output of full articulamentum is spliced into one by last output layer
Individual multi-C vector, exported as characteristics of image.CNN networks are used to go to attempt feature without artificial with automatic mining feature.
After characteristics of image is obtained, operation S230 is performed, based on described image feature, is pushed in the multiple object
Partial objects.
The embodiment illustrated with reference to Fig. 3 A and Fig. 3 B, operation S230 is illustrated.
Fig. 3 A are diagrammatically illustrated according to the embodiment of the present disclosure based on described image feature, are pushed in the multiple object
Partial objects flow chart.
As shown in Figure 3A, this method includes operation S231, S232 and S233.
S231 is being operated, according to user's history behavior, it is determined that the image information of at least one object related to user.
In operation S232, by neutral net, each object in the multiple object and the described and user are determined
First similarity of related at least one object.
In operation S233, based on first similarity, the partial objects in the multiple object are pushed.
For example, the historical behavior data of user are shown, the user once actively browsed object X1,X2,…,Xn, operating
S231, you can determine the object X related to user1,X2,…,XnImage information A1,A2,…,An.To be more in operation S232
The image information B of the individual object being recommended1,B2,…,BmWith A1,A2,…,AnCommon input neutral net, the figure based on acquisition
As feature, B is determined1,B2,…,BmIn each image information and A1,A2,…,AnSimilarity, be every in multiple objects
Individual object and X1,X2,…,XnThe first similarity.Finally, S233 is being operated, based on first similarity, such as can root
According to sequencing of similarity, the higher partial objects of similarity in the multiple object are pushed.
Historical behavior of this method based on user, it is determined that the object related to user, it is right to obtain other by neutral net
As the similarity with above-mentioned object, and recommended according to this, make recommendation more personalized, meet user's request, improve recommendation
The degree of accuracy.
Fig. 3 B diagrammatically illustrate the multiple right based on described image feature, push according to another embodiment of the disclosure
The flow chart of partial objects as in.
As shown in Figure 3 B, this method includes operation S234, S235 and S236.
In operation S234, according to user's current behavior, the image information of existing object is determined.
In operation S235, by neutral net, each object in the multiple object and the existing object are determined
Second similarity.
In operation S236, based on second similarity, the partial objects in the multiple object are pushed.
For example, user is currently browsing some object X, in operation S234, you can determine the object X related to user
Image information A.S235 is being operated, by the image information B of multiple objects being recommended1,B2,…,BmGod is inputted jointly with A
Through network, the characteristics of image based on acquisition, B is determined1,B2,…,BmIn each image information and A similarity, be multiple
Second similarity of each object and existing object X in object.Finally, S236 is being operated, based on second similarity, example
Such as the partial objects higher with existing object X similarities in the multiple object can be pushed according to sequencing of similarity.
Current behavior of this method based on user, determines existing object, by neutral net obtain other objects with it is current
The similarity of object, and recommended according to this so that recommend the specific aim for user's current demand stronger.
Above-mentioned two embodiment is only schematical, and those skilled in the art can also otherwise, based on the figure
As feature, from the humping section object in the multiple object.
According to the embodiment of the present disclosure, based on described image feature, from the humping section object in the multiple object not
It can be used in combination with mode.Below with reference to embodiment illustrated in Figure 4, the situation of combined use is illustrated.
Fig. 4 diagrammatically illustrates the flow of the partial objects in the multiple object of push according to the embodiment of the present disclosure
Figure.
As shown in figure 4, this method includes operation S410, S420 and S430.
In operation S410, determine partial objects as first set from the multiple object.Pushed away using various ways
In the case of sending, each way can be pushed out partial objects, according to the embodiment of the present disclosure, can push various push modes
Object collect, formed first set.
Operation S420, the attribute information of each object in the first set, the behavioural information of user, with
And/or person, the attribute information of user, by neutral net, obtain the recommended priority of each object.It is real according to the disclosure
Apply example, by the attribute information of each object in the first set, user behavioural information and/or, the attribute of user
Information, which is used as, to be inputted, input neutral net, after Processing with Neural Network, exports the recommended priority of each object.
To achieve these goals, it is necessary to train neutral net.According to the embodiment of the present disclosure, can be trained using training set
The neutral net, the training set include user behavior data, and the recommended priority is the probability that user behavior occurs.It is right
In user and each object, the relevant information for collecting user and object is directed to the operation conduct of each object as input, user
As a result, it is right using the above as training set for example, being clicked on using user and browsing the object related content as basis for estimation
Neutral net is trained., can be with when inputting the relevant information of user and object using the neutral net Jing Guo above-mentioned training
User's operation is predicted, for example, whether user can click on and browse the related content of the object.Prediction result is presented as event
Probability, can be using the probability as recommended priority.
The training method that this method provide can automatically obtain training data, and hold according to the operation occurred in real time
It is trained continuously, saves cost of labor, and being capable of effectively training pattern.
In operation S430, according to the recommended priority, part or all of object conduct is determined from the first set
Second set.For example, can be sorted according to recommended priority, determine that the part that recommended priority is higher in the first set is right
As second set.It is determined that during the higher partial objects of priority, a threshold value can be preset, selects recommended priority high
In the object of predetermined threshold value, the of a relatively high a number of object of recommended priority in first set can also be selected, as
Two set, certainly, both the above mode can also be used in combination.
In operation S440, the object in the second set is pushed.
This method obtains the recommended priority of each object by neutral net, and is pushed according to priority so that
Limited most suitable object can be pushed out.
Fig. 5 diagrammatically illustrates to be believed according to the attribute of each object in the first set of the embodiment of the present disclosure
Breath, user behavioural information and/or, the attribute information of user, by neutral net, obtain pushing away for each object
Recommend the flow chart of priority.
As shown in figure 5, this method includes operation S510, S520 and S530.
In operation S510, according to customer attribute information, user behavior information and object properties information, pass through the depth
Part of neural network is spent, obtains the first result.Deep neural network part can be between learning characteristic intersection, it is extensive to obtain
The first stronger result of property.
In operation S520, according to user behavior information, and object properties information, by the logistic regression part, obtain
Obtain the second result.Logistic regression part has preferable Memorability
In operation S530, handle first result and second result, the recommendation for obtaining each object are preferential
Level.According to the embodiment of the present disclosure, the result of summary first and the second result, return, such as returned using softmax again,
Obtain the recommended priority of each object.
This method is by introducing customer attribute information, and using deep neural network part, while for user behavior and
Commodity use logistic regression so that result can have preferable generalization and Memorability simultaneously.
Fig. 6 diagrammatically illustrates the block diagram of the supplying system 600 according to the embodiment of the present disclosure.
As shown in fig. 6, the supplying system 600 includes image collection module 610, characteristic extracting module 620 and pushing module
630。
Image collection module 610, such as the operation S210 above with reference to Fig. 2 descriptions is performed, it is the multiple right for obtaining
The image information of elephant.
Characteristic extracting module 620, such as perform the operation S220 above with reference to Fig. 2 descriptions, for by neutral net, from
The characteristics of image of each object is extracted in the image information of the multiple object.
Pushing module 630, such as the operation S230 above with reference to Fig. 2 descriptions is performed, for based on described image feature, pushing away
Send the partial objects in the multiple object.
Fig. 7 A diagrammatically illustrate the block diagram of the pushing module 630 according to the embodiment of the present disclosure.
As shown in fig. 7, the pushing module 630 pushes away including the first determination sub-module 710, the first comparison sub-module 720, first
Send submodule 730.
First determination sub-module 710, such as the operation S231 above with reference to Fig. 3 A descriptions is performed, for according to user's history
Behavior, it is determined that the image information of at least one object related to the user.
First comparison sub-module 720, such as the operation S232 above with reference to Fig. 3 A descriptions is performed, for passing through nerve net
Network, determine that each object in the multiple object is similar to the first of at least one object related to the user
Degree.
First push submodule 730, such as the operation S233 above with reference to Fig. 3 A descriptions is performed, for based on described first
Similarity, push the partial objects in the multiple object.
Fig. 7 B diagrammatically illustrate the block diagram of the pushing module 630 according to another embodiment of the disclosure.
As shown in Figure 7 B, the pushing module 630 includes the second determination sub-module 740, the second comparison sub-module 750, second
Push submodule 760.
Second determination sub-module 740, such as the operation S234 above with reference to Fig. 3 B descriptions is performed, for current according to user
Behavior, determine the image information of existing object.
Second comparison sub-module 750, such as the operation S235 above with reference to Fig. 3 B descriptions is performed, for passing through nerve net
Network, determine the second similarity of each object and the existing object in the multiple object.
Second push submodule 760, such as the operation S236 above with reference to Fig. 3 B descriptions is performed, for based on described second
Similarity, push the partial objects in the multiple object.
Fig. 8 diagrammatically illustrates the block diagram of the pushing module 630 according to another embodiment of the disclosure.
As shown in figure 8, the pushing module 630 includes the 3rd determination sub-module 810, priority determination sub-module the 820, the 4th
Determination sub-module the 830, the 3rd pushes submodule 840.
3rd determination sub-module 810, such as the operation S410 above with reference to Fig. 4 descriptions is performed, for from the multiple right
As middle determination partial objects are as first set.
Priority determination sub-module 820, such as the operation S420 above with reference to Fig. 4 descriptions is performed, for according to described the
One set in the attribute information of each object, the behavioural information of user and/or, the attribute information of user, pass through nerve
Network, obtain the recommended priority of each object.
4th determination sub-module 830, such as the operation S430 above with reference to Fig. 4 descriptions is performed, for according to the recommendation
Priority, determine part or all of object as second set from the first set.
3rd push submodule 840, such as the operation S440 above with reference to Fig. 4 descriptions is performed, for pushing described second
Object in set.
Fig. 9 diagrammatically illustrates the block diagram of priority determination sub-module 920 according to the embodiment of the present disclosure.
As shown in figure 9, the priority determination sub-module 820 includes training unit 910.
Training unit 910, for training the neutral net using training set, the training set includes user behavior number
According to the recommended priority is the probability that user behavior occurs.
Figure 10 diagrammatically illustrates the block diagram of priority determination sub-module 820 according to another embodiment of the disclosure.
As shown in Figure 10, the priority determination sub-module 820 includes first processing units 1010, second processing unit 1020
With the 3rd processing unit 1030.
First processing units 1010, such as the operation S510 above with reference to Fig. 5 descriptions is performed, for being believed according to user property
Breath, user behavior information and object properties information, by the deep neural network part, obtain the first result.
Second processing unit 1020, such as the operation S520 above with reference to Fig. 5 descriptions is performed, for being believed according to user behavior
Breath, and object properties information, by the logistic regression part, obtain the second result.
3rd processing unit 1030, such as the operation S530 above with reference to Fig. 5 descriptions is performed, for handling first knot
Fruit and second result, obtain the recommended priority of each object.
Figure 11 diagrammatically illustrates the calculating for being adapted for carrying out information processing method and/or system according to the embodiment of the present disclosure
The block diagram of machine system 1100.
Computer system shown in Figure 11 is only an example, should not be to the function and use range of the embodiment of the present disclosure
Bring any restrictions.
As shown in figure 11, processor 1101 is included according to the computer system 1100 of the embodiment of the present disclosure, it can basis
The program that is stored in read-only storage (ROM) 1102 is loaded into random access storage device (RAM) from storage part 1108
Program in 1103 and perform various appropriate actions and processing.Processor 1101 can for example include general purpose microprocessor (example
Such as CPU), instruction set processor and/or related chip group and/or special microprocessor (for example, application specific integrated circuit (ASIC)),
Etc..Processor 1101 can also include being used for the onboard storage device for caching purposes.Processor 1101 can include being used to perform
Single treatment unit with reference to the different actions of 2~Fig. 5 of figure method flows according to the embodiment of the present disclosure described is either more
Individual processing unit.
In RAM 1103, it is stored with system 1100 and operates required various programs and data.Processor 1101, ROM
1102 and RAM 1103 is connected with each other by bus 1104.Processor 1101 is by performing ROM 1102 and/or RAM 1103
In program come perform above with reference to Fig. 2~Fig. 5 describe information processing method various operations.It is noted that described program
It can also be stored in one or more memories in addition to ROM 1102 and RAM 1103.Processor 1101 can also pass through
Perform and be stored in the program in one or more of memories to perform the information processing side described above with reference to Fig. 2~Fig. 5
The various operations of method.
In accordance with an embodiment of the present disclosure, system 1100 can also include input/output (I/O) interface 1105, input/output
(I/O) interface 1105 is also connected to bus 1104.System 1100 can also include be connected to I/O interfaces 1105 with lower component
It is one or more：Importation 1106 including keyboard, mouse etc.；Including such as cathode-ray tube (CRT), liquid crystal display
Etc. (LCD) and loudspeaker etc. output par, c 1107；Storage part 1108 including hard disk etc.；And including such as LAN card,
The communications portion 1109 of the NIC of modem etc..Communications portion 1109 performs logical via the network of such as internet
Letter processing.Driver 1110 is also according to needing to be connected to I/O interfaces 1105.Detachable media 1111, such as disk, CD, magnetic
CD, semiconductor memory etc., it is arranged on as needed on driver 1110, in order to the computer program read from it
Storage part 1108 is mounted into as needed.
In accordance with an embodiment of the present disclosure, it may be implemented as computer software journey above with reference to the method for flow chart description
Sequence.For example, embodiment of the disclosure includes a kind of computer program product, it includes carrying meter on a computer-readable medium
Calculation machine program, the computer program include the program code for being used for the method shown in execution flow chart.In such embodiments,
The computer program can be downloaded and installed by communications portion 1109 from network, and/or from the quilt of detachable media 1111
Installation.When the computer program is performed by processor 1101, the above-mentioned function of being limited in the system of the embodiment of the present disclosure is performed.
In accordance with an embodiment of the present disclosure, system as described above, unit, module, unit etc. can pass through computer program module
To realize.
It should be noted that the computer-readable medium shown in the disclosure can be computer-readable signal media or meter
Calculation machine readable storage medium storing program for executing either the two any combination.Computer-readable recording medium for example can be --- but not
Be limited to --- electricity, magnetic, optical, electromagnetic, system, device or the device of infrared ray or semiconductor, or it is any more than combination.Meter
The more specifically example of calculation machine readable storage medium storing program for executing can include but is not limited to：Electrical connection with one or more wires, just
Take formula computer disk, hard disk, random access storage device (RAM), read-only storage (ROM), erasable type and may be programmed read-only storage
Device (EPROM or flash memory), optical fiber, portable compact disc read-only storage (CD-ROM), light storage device, magnetic memory device,
Or above-mentioned any appropriate combination.In the disclosure, computer-readable recording medium can any include or store journey
The tangible medium of sequence, the program can be commanded the either device use or in connection of execution system, device.And at this
In open, computer-readable signal media can be included in a base band or the data-signal as carrier wave part propagation,
Wherein carry computer-readable program code.The data-signal of this propagation can take various forms, including but unlimited
In electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be that computer can
Any computer-readable medium beyond storage medium is read, the computer-readable medium, which can send, propagates or transmit, to be used for
By instruction execution system, device either device use or program in connection.Included on computer-readable medium
Program code can be transmitted with any appropriate medium, be included but is not limited to：Wirelessly, electric wire, optical cable, RF etc., or it is above-mentioned
Any appropriate combination.In accordance with an embodiment of the present disclosure, computer-readable medium can include above-described ROM 1102
And/or one or more memories beyond RAM 1103 and/or ROM 1102 and RAM 1103.
Flow chart and block diagram in accompanying drawing, it is illustrated that according to the system of the various embodiments of the disclosure, method and computer journey
Architectural framework in the cards, function and the operation of sequence product.At this point, each square frame in flow chart or block diagram can generation
The part of one module of table, program segment or code, a part for above-mentioned module, program segment or code include one or more
For realizing the executable instruction of defined logic function.It should also be noted that some as replace realization in, institute in square frame
The function of mark can also be with different from the order marked in accompanying drawing generation.For example, two square frames succeedingly represented are actual
On can perform substantially in parallel, they can also be performed in the opposite order sometimes, and this is depending on involved function.Also
It is noted that the combination of each square frame and block diagram in block diagram or flow chart or the square frame in flow chart, can use and perform rule
Fixed function or the special hardware based system of operation are realized, or can use the group of specialized hardware and computer instruction
Close to realize.
As on the other hand, the disclosure additionally provides a kind of computer-readable medium, and the computer-readable medium can be
Included in equipment described in above-described embodiment；Can also be individualism, and without be incorporated the equipment in.Above-mentioned calculating
Machine computer-readable recording medium carries one or more program, when said one or multiple programs are performed by the equipment, makes
Obtain the equipment and perform the method according to the embodiment of the present disclosure described with reference to 2~Fig. 5 of figure.
Embodiment of the disclosure is described above.But the purpose that these embodiments are merely to illustrate that, and
It is not intended to limit the scope of the present disclosure.Although respectively describing each embodiment more than, but it is not intended that each reality
Use can not be advantageously combined by applying the measure in example.The scope of the present disclosure is defined by the appended claims and the equivalents thereof.Do not take off
From the scope of the present disclosure, those skilled in the art can make a variety of alternatives and modifications, and these alternatives and modifications should all fall at this
Within scope of disclosure.
Claims (14)
1. a kind of method for pushing, for from multiple objects humping section object, methods described include：
Obtain the image information of the multiple object；
By neutral net, the characteristics of image of each object is extracted from the image information of the multiple object；And
Based on described image feature, the partial objects in the multiple object are pushed.
2. the method according to claim 11, wherein, it is described to be based on described image feature, push in the multiple object
Partial objects include：
According to user's history behavior, it is determined that the image information of at least one object related to the user；
By neutral net, determine each object in the multiple object to it is described related with the user at least one right
The first similarity of elephant；And
Based on first similarity, the partial objects in the multiple object are pushed.
3. the method according to claim 11, wherein, it is described to be based on described image feature, push in the multiple object
Partial objects include：
According to user's current behavior, the image information of existing object is determined；
By neutral net, the second similarity of each object and the existing object in the multiple object is determined；And
Based on second similarity, the partial objects in the multiple object are pushed.
4. according to the method for claim 1, wherein, the partial objects in the multiple object of push include：
Determine partial objects as first set from the multiple object；
The attribute information of each object in the first set, the behavioural information of user and/or, the category of user
Property information, by neutral net, obtain the recommended priority of each object；
According to the recommended priority, determine part or all of object as second set from the first set；And
Push the object in the second set.
5. according to the method for claim 4, wherein, the attribute of each object in the first set is believed
Breath, user behavioural information and/or, the attribute information of user, by neutral net, obtain pushing away for each object
Recommending priority includes：
The neutral net is trained using training set, the training set includes user behavior data, and the recommended priority is use
The probability that family behavior occurs.
6. according to the method for claim 4, wherein, the neutral net includes deep neural network part and logistic regression
Part, the attribute information of each object in the first set, the behavioural information of user and/or, user
Attribute information, by neutral net, obtaining the recommended priority of each object includes：
According to customer attribute information, user behavior information and object properties information, by the deep neural network part,
Obtain the first result；
According to user behavior information, and object properties information, by the logistic regression part, the second result is obtained；And
First result and second result are handled, obtains the recommended priority of each object.
7. a kind of supplying system, for the humping section object from multiple objects, the system includes：
Image collection module, for obtaining the image information of the multiple object；
Characteristic extracting module, for by neutral net, extracting the figure of each object from the image information of the multiple object
As feature；And
Pushing module, for based on described image feature, pushing the partial objects in the multiple object.
8. system according to claim 7, wherein, the pushing module includes：
First determination sub-module, for according to user's history behavior, it is determined that the figure of at least one object related to the user
As information；
First comparison sub-module, for by neutral net, determine each object in the multiple object with it is described with it is described
First similarity of the related at least one object of user；And
First push submodule, for based on first similarity, pushing the partial objects in the multiple object.
9. system according to claim 7, wherein, the pushing module includes：
Second determination sub-module, for according to user's current behavior, determining the image information of existing object；
Second comparison sub-module, for by neutral net, determine each object in the multiple object with it is described current right
The second similarity of elephant；And
Second push submodule, for based on second similarity, pushing the partial objects in the multiple object.
10. system according to claim 7, wherein, the pushing module includes：
3rd determination sub-module, for determining partial objects as first set from the multiple object；
Priority determination sub-module, the behavior letter of attribute information, user for each object in the first set
Breath and/or, the attribute information of user, by neutral net, obtain the recommended priority of each object；
4th determination sub-module, for according to the recommended priority, part or all of object to be determined from the first set
As second set；And
3rd push submodule, for pushing the object in the second set.
11. system according to claim 10, wherein, the priority determination sub-module includes：
Training unit, for training the neutral net using training set, the training set includes user behavior data, described to push away
It is the probability that user behavior occurs to recommend priority.
12. system according to claim 10, wherein, the neutral net includes deep neural network part and logic is returned
Return part, the priority determination sub-module includes：
First processing units, for according to customer attribute information, user behavior information and object properties information, by described
Deep neural network part, obtain the first result；
Second processing unit, for according to user behavior information, and object properties information, by the logistic regression part,
Obtain the second result；And
3rd processing unit, for handling first result and second result, the recommendation for obtaining each object is excellent
First level.
13. a kind of electronic equipment, including：
One or more processors；
Storage device, for storing one or more programs,
Wherein, when one or more of programs are by one or more of computing devices so that one or more of
Method in computing device such as claim 1~7 as described in any one.
14. a kind of computer-readable medium, is stored thereon with executable instruction, the instruction holds processor when being executed by processor
Method of the row as described in any one in claim 1~7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710747528.9A CN107562845A (en) | 2017-08-25 | 2017-08-25 | Method for pushing, system and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710747528.9A CN107562845A (en) | 2017-08-25 | 2017-08-25 | Method for pushing, system and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107562845A true CN107562845A (en) | 2018-01-09 |
Family
ID=60977141
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710747528.9A Pending CN107562845A (en) | 2017-08-25 | 2017-08-25 | Method for pushing, system and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107562845A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111667340A (en) * | 2020-05-29 | 2020-09-15 | 中国平安财产保险股份有限公司 | Target object recommendation method and device based on big data and computer-readable storage medium |
| CN113343006A (en) * | 2021-06-28 | 2021-09-03 | 北京三快在线科技有限公司 | Picture recommendation method, device, equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1993063A2 (en) * | 2007-05-16 | 2008-11-19 | Canon Kabushiki Kaisha | Image processing apparatus and image retrieval method |
| CN106202362A (en) * | 2016-07-07 | 2016-12-07 | Tcl集团股份有限公司 | Image recommendation method and image recommendation device |
| CN106485567A (en) * | 2016-09-14 | 2017-03-08 | 北京小米移动软件有限公司 | Item recommendation method and device |
-
2017
- 2017-08-25 CN CN201710747528.9A patent/CN107562845A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1993063A2 (en) * | 2007-05-16 | 2008-11-19 | Canon Kabushiki Kaisha | Image processing apparatus and image retrieval method |
| CN106202362A (en) * | 2016-07-07 | 2016-12-07 | Tcl集团股份有限公司 | Image recommendation method and image recommendation device |
| CN106485567A (en) * | 2016-09-14 | 2017-03-08 | 北京小米移动软件有限公司 | Item recommendation method and device |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111667340A (en) * | 2020-05-29 | 2020-09-15 | 中国平安财产保险股份有限公司 | Target object recommendation method and device based on big data and computer-readable storage medium |
| CN113343006A (en) * | 2021-06-28 | 2021-09-03 | 北京三快在线科技有限公司 | Picture recommendation method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105320766B (en) | Information-pushing method and device | |
| US10497013B2 (en) | Purchasing behavior analysis apparatus and non-transitory computer readable medium | |
| CN108805594B (en) | Information pushing method and device | |
| CN109190049B (en) | Keyword recommendation method, system, electronic device and computer readable medium | |
| US11172040B2 (en) | Method and apparatus for pushing information | |
| CN107679217B (en) | Associated content extraction method and device based on data mining | |
| CN108345702A (en) | Entity recommends method and apparatus | |
| CN107172151A (en) | Method and apparatus for pushed information | |
| CN108090162A (en) | Information-pushing method and device based on artificial intelligence | |
| CN105677931B (en) | Information search method and device | |
| CN109145280A (en) | The method and apparatus of information push | |
| CN108804450A (en) | The method and apparatus of information push | |
| CN107992554A (en) | The searching method and device of the polymerization result of question and answer information are provided | |
| CN110827112B (en) | Deep learning commodity recommendation method and device, computer equipment and storage medium | |
| US20200226168A1 (en) | Methods and systems for optimizing display of user content | |
| US20180012284A1 (en) | Information analysis apparatus, information analysis method, and non-transitory computer readable storage medium | |
| JP6976207B2 (en) | Information processing equipment, information processing methods, and programs | |
| CN108256537A (en) | A kind of user gender prediction method and system | |
| US20150379610A1 (en) | Recommendation information presentation device, recommendation information presentation method, and recommendation information presentation program | |
| CN108121814B (en) | Search result ranking model generation method and device | |
| CN111967924A (en) | Commodity recommendation method, commodity recommendation device, computer device, and medium | |
| CN112330382B (en) | Item recommendation method, device, computing equipment and medium | |
| CN115033801B (en) | Article recommendation method, model training method and electronic equipment | |
| CN114154013A (en) | Video recommendation method, device, equipment and storage medium | |
| CN108804448A (en) | The method and apparatus for generating information to be pushed |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |