CN109840793A - Method for pushing and device, storage medium, the electronic device of promotion message - Google Patents
Method for pushing and device, storage medium, the electronic device of promotion message Download PDFInfo
- Publication number
- CN109840793A CN109840793A CN201711217712.9A CN201711217712A CN109840793A CN 109840793 A CN109840793 A CN 109840793A CN 201711217712 A CN201711217712 A CN 201711217712A CN 109840793 A CN109840793 A CN 109840793A
- Authority
- CN
- China
- Prior art keywords
- account number
- media information
- promotion message
- user
- media
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 65
- 238000003860 storage Methods 0.000 title claims abstract description 22
- 238000012549 training Methods 0.000 claims description 31
- 230000015654 memory Effects 0.000 claims description 22
- 238000012545 processing Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 4
- 230000032258 transport Effects 0.000 claims 1
- 238000005516 engineering process Methods 0.000 abstract description 11
- 239000000047 product Substances 0.000 description 17
- 238000010801 machine learning Methods 0.000 description 15
- 238000010586 diagram Methods 0.000 description 11
- 230000006870 function Effects 0.000 description 10
- 230000008569 process Effects 0.000 description 9
- 230000005540 biological transmission Effects 0.000 description 5
- 238000004891 communication Methods 0.000 description 5
- 230000002123 temporal effect Effects 0.000 description 5
- 238000010276 construction Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 235000013305 food Nutrition 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 230000006872 improvement Effects 0.000 description 3
- 244000291564 Allium cepa Species 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 238000013145 classification model Methods 0.000 description 2
- 238000012790 confirmation Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 101001084254 Homo sapiens Peptidyl-tRNA hydrolase 2, mitochondrial Proteins 0.000 description 1
- 102100030867 Peptidyl-tRNA hydrolase 2, mitochondrial Human genes 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000005304 joining Methods 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000005389 magnetism Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000005498 polishing Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of method for pushing of promotion message and device, storage medium, electronic device.Wherein, this method comprises: obtaining the first media information that the first account number was shared, wherein for the first media information for being shared between the account number of target application, the account number of target application includes the first account number;The first account number set belonging to the first account number is determined based on the first media information;Using with the first promotion message of the first account number sets match as the promotion message wait be pushed to the first account number, wherein the first promotion message is used to indicate the first object to be promoted;The first promotion message is pushed to the first account number.The present invention solves and searches the technical problem lower with the efficiency of the matched promotion message of the user when pushing promotion message to user in the related technology.
Description
Technical field
The present invention relates to internet areas, and the method for pushing and device, storage in particular to a kind of promotion message are situated between
Matter, electronic device.
Background technique
With the development of internet, user is increasing to the quantity demand of browsing media information, and businessman, video production are public
Department etc. generally requires the media promotion message in the media upslide amplification quantity such as internet platform, TV, these media promotion messages
Including but not limited to: film, TV play, short-movie, advertisement video, advertising pictures, propaganda film, animation video etc..
Due to user preferences difference, the media information of release can not necessarily obtain the hobby of user.In the related art,
Qualitative mode was often used under line to be analyzed, to be determined for compliance with each user demand according to the information investigated out
Media information.
Although can accurately be contacted using qualitative fashion under line and distinguish different user (such as different types of viewing person,
Advertising user, TV play spectators etc.) demand, but it is long the time required to the demand that mode determines different user under line, such as manually look into
The web page browsing record of certain user is seen, so that browsing record according to it determines its favorite media promotion message type of institute, in turn
It is pushed according to the type, whole process takes time and effort, low efficiency, at high cost.
For searched when pushing promotion message to user in the related technology with the efficiency of the matched promotion message of the user compared with
Low technical problem, currently no effective solution has been proposed.
Summary of the invention
The embodiment of the invention provides a kind of method for pushing of promotion message and device, storage medium, electronic device, so that
It is few to solve to search the efficiency lower skill with the matched promotion message of the user when pushing promotion message to user in the related technology
Art problem.
According to an aspect of an embodiment of the present invention, a kind of method for pushing of promotion message is provided, this method comprises: obtaining
Take the first media information that the first account number was shared, wherein the first media information between the account number of target application for carrying out
Share, the account number of target application includes the first account number;The first account number collection belonging to the first account number is determined based on the first media information
It closes;Using with the first promotion message of the first account number sets match as the promotion message wait be pushed to the first account number, wherein first
Promotion message is used to indicate the first object to be promoted;The first promotion message is pushed to the first account number.
According to another aspect of an embodiment of the present invention, a kind of driving means of promotion message is additionally provided, which includes:
Acquiring unit, the first media information shared for obtaining the first account number, wherein the first media information is used in target application
Account number between shared, the account number of target application includes the first account number;Determination unit, for true based on the first media information
First account number set belonging to fixed first account number;Processing unit, for by the first promotion message with the first account number sets match
As the promotion message wait be pushed to the first account number, wherein the first promotion message is used to indicate the first object to be promoted;Push
Unit, for pushing the first promotion message to the first account number.
According to another aspect of an embodiment of the present invention, a kind of storage medium is additionally provided, which includes storage
Program, program execute above-mentioned method when running.
According to another aspect of an embodiment of the present invention, it additionally provides a kind of electronic device, including memory, processor and deposits
The computer program that can be run on a memory and on a processor is stored up, processor executes above-mentioned side by computer program
Method.
In embodiments of the present invention, in determining promotion message matched with the first account number, the first account number point can first be obtained
The first media information enjoyed, the first media information are based on the first media for being shared between the account number of target application
Information determines the first account number set belonging to the first account number;And using with the first promotion message of the first account number sets match as to
It is pushed to the promotion message of the first account number, the first promotion message is used to indicate the first object to be promoted;It is pushed to the first account number
First promotion message can solve to search when pushing promotion message to user in the related technology and believe with the matched popularization of the user
The lower technical problem of the efficiency of breath, and then reach the technical effect for improving and searching with the matched promotion message of user.
Detailed description of the invention
The drawings described herein are used to provide a further understanding of the present invention, constitutes part of this application, this hair
Bright illustrative embodiments and their description are used to explain the present invention, and are not constituted improper limitations of the present invention.In the accompanying drawings:
Fig. 1 is the schematic diagram of the hardware environment of the method for pushing of promotion message according to an embodiment of the present invention;
Fig. 2 is a kind of flow chart of the method for pushing of optional promotion message according to an embodiment of the present invention;
Fig. 3 is a kind of schematic diagram at optional circle of friends interface according to an embodiment of the present invention;
Fig. 4 is a kind of schematic diagram at optional circle of friends interface according to an embodiment of the present invention;
Fig. 5 is a kind of flow chart of optional construction training data according to an embodiment of the present invention;
Fig. 6 is a kind of schematic diagram of optional image tag sequence according to an embodiment of the present invention;
Fig. 7 is a kind of schematic diagram of the network structure of optional model according to an embodiment of the present invention;
Fig. 8 is a kind of schematic diagram of the network structure of optional model according to an embodiment of the present invention;
Fig. 9 is a kind of schematic diagram of the driving means device of optional promotion message according to an embodiment of the present invention;And
Figure 10 is a kind of structural block diagram of terminal according to an embodiment of the present invention.
Specific embodiment
In order to enable those skilled in the art to better understand the solution of the present invention, below in conjunction in the embodiment of the present invention
Attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only
The embodiment of a part of the invention, instead of all the embodiments.Based on the embodiments of the present invention, ordinary skill people
The model that the present invention protects all should belong in member's every other embodiment obtained without making creative work
It encloses.
It should be noted that description and claims of this specification and term " first " in above-mentioned attached drawing, "
Two " etc. be to be used to distinguish similar objects, without being used to describe a particular order or precedence order.It should be understood that using in this way
Data be interchangeable under appropriate circumstances, so as to the embodiment of the present invention described herein can in addition to illustrating herein or
Sequence other than those of description is implemented.In addition, term " includes " and " having " and their any deformation, it is intended that cover
Cover it is non-exclusive include, for example, the process, method, system, product or equipment for containing a series of steps or units are not necessarily limited to
Step or unit those of is clearly listed, but may include be not clearly listed or for these process, methods, product
Or other step or units that equipment is intrinsic.
Firstly, the part noun or term that occur during the embodiment of the present invention is described are suitable for as follows
It explains:
Embedding: by data information from a space reflection to another space, it is however generally that the dimension in latter space
Degree is more previous, and space is smaller, but retains the essential attribute of data information, for example, vocabulary word Embedding is exactly from a height
Dimension space is mapped to low dimensional space.
LSTM (Long Short-Term Memory), Chinese is shot and long term memory network, is a kind of specific cycle
The type of neural network RNN (Recurrent Neural Networks), can learn long-term Dependency Specification.
One side according to an embodiment of the present invention provides a kind of embodiment of the method for the method for pushing of promotion message.
Optionally, in the present embodiment, the method for pushing of above-mentioned promotion message can be applied to as shown in Figure 1 by servicing
In the hardware environment that device 102 and terminal 104 are constituted.As shown in Figure 1, server 102 is connected by network with terminal 104
Connect, above-mentioned network includes but is not limited to: wide area network, Metropolitan Area Network (MAN) or local area network, terminal 104 are not limited to PC, mobile phone, plate electricity
Brain etc..The method for pushing of the promotion message of the embodiment of the present invention can be executed by server 102, can also by terminal 104
It executes, can also be and executed jointly by server 102 and terminal 104.Wherein, terminal 104 executes the popularization of the embodiment of the present invention
The method for pushing of information is also possible to be executed by client mounted thereto.
Entire flow including the present processes is described in detail for executing the present processes on the server below:
Step S11, the account numbers such as the first account number, the second account number issued state in the client of target application.The state can
To be video, picture, audio, one or more the combination in text.
Step S12, the state of account number publication is by the server of client synchronization to target application.
Database is arrived in step S13, the state that server issues each account number (or keyword) storage as unit of account number.
Step S14, after triggering the present processes, server obtains all shapes of each account number in historical time section
State.
Step S15, server is according to stateful the first promotion message of determination of institute in historical time section.Specific method is such as
Under:
Step S151 obtains the first media information that the first account number was shared, wherein the first media information is used in target
Shared between the account number of application, the account number of target application includes the first account number;
Step S152 determines the first account number set belonging to the first account number based on the first media information;
Step S153, using with the first promotion message of the first account number sets match as the popularization wait be pushed to the first account number
Information, wherein the first promotion message is used to indicate the first object to be promoted.
Step S16, the client where server to the first account number send the first promotion message, open visitor in the first account number
First promotion message can be shown behind the end of family in the client.
The application is concentrated mainly on step S15 and the step related to step S15 to the improvement of above-mentioned entire flow, below
Improved place is described in detail in conjunction with Fig. 2:
Fig. 2 is a kind of flow chart of the method for pushing of optional promotion message according to an embodiment of the present invention, such as Fig. 2 institute
Show, this method may comprise steps of:
Step S202, obtains the first media information that the first account number was shared, and the first media information is used in target application
Account number between shared, the account number of target application includes the first account number.
Above-mentioned target application includes but is not limited to: instant messaging application, social application, microblogging application, video class are answered
With, web application, picture share application, music share application etc..The account number of above-mentioned target application allows in target application
Platform on the account number that uses in (such as applications client, webpage client), two types can be divided into for source,
First is that registration target application using account number, the second is other than target application application in register and by target application
Allow the account number used in target application;Two kinds can also be divided into for account number type, one is allowing to be made for a long time by user
Register account number, the second is when user is using target application as the interim account number of user's distribution.
The first above-mentioned media information is all media informations that the user of the first account number shared in historical time section
(namely the quantity of the first media information can be multiple), which can be the pervious time at current time,
But using current time as the period of the fixation duration of finish time or the time of current time pervious fixed duration
Section.
Step S204 determines the first account number set belonging to the first account number based on the first media information.
All account numbers or a part of account number can be divided into according to characteristic value of the account number in multiple characteristic dimensions more
A account number set.
Step S206, using with the first promotion message of the first account number sets match as the popularization wait be pushed to the first account number
Information, wherein the first promotion message is used to indicate the first object to be promoted.
Above-mentioned promotion message is information (the first popularization letter as being used to promote the first object for being used to promote some object
Breath), the object can for virtual or entity object, the commodity such as promoted, promote certain thought or moral idea, film,
Content of courses etc..
Step S208 pushes the first promotion message to the first account number, namely to the first client for being logged in the first account number
The first promotion message is pushed, the first client is the client of target application.
During above-mentioned the first promotion message of confirmation, all steps can be run in the form of program code, and nothing
It need to manually participate in, therefore the determining efficiency with the matched promotion message type of the user can be greatly improved, meanwhile, with account number collection
Unit is combined into uniformly to determine that the promotion message of the set can be further improved matching efficiency;In addition, by thus according to first
The historical information (the first media information shared) of account number to select suitable first promotion message therefore can be most for it
Limits match with the hobby of user, such as meet film, the animation of its hobby, some aspects learnt
Content etc. is practised, the experience of user can be improved in which.
S202 to step S208 through the above steps can be obtained first in determining promotion message matched with the first account number
The first media information that first account number was shared, the first media information between the account number of target application for being shared, base
The first account number set belonging to the first account number is determined in the first media information;And it will be promoted with the first of the first account number sets match
For information as the promotion message wait be pushed to the first account number, the first promotion message is used to indicate the first object to be promoted;To
One account number pushes the first promotion message, can solve and has searched and the user when pushing promotion message to user in the related technology
The lower technical problem of the efficiency for the promotion message matched, and then reach to improve to search and be imitated with the technology of the matched promotion message of user
Fruit.
For convenience, in the application by taking target application is instant messaging application as an example, for promotion message is advertisement
It is illustrated, embodiments herein, which is equivalent to, at this time provides a kind of advertisement dispensing based on media informations such as circle of friends images
Method provides a kind of new mode to launch for circle of friends advertisement.In the application that target application is remaining type and/or promote
When information is the information of remaining type, implementation is same or similar with mode described in the application subsequent embodiment, this
Application repeats no more.
Using the technical solution of the application, some user behaviors logs of the user on network can be utilized when targeted ads are launched
And the basis portrait of user to carry out analysis mining to user, so that user's portrait is constructed, according to user's portrait come by user
With ad content carry out similarity mode, and for user launch similarity it is higher, user may interested advertisement, certainly also
Can content-based recommendation, use the recommended method based on collaborative filtering.It, can benefit for the circle of friends advertisement of instant messaging
With the operation log data of user in instant communication applications etc. portrait building carried out to user, and circle of friends delivers picture number
According to the interest and hobby for also having fed back user to a certain extent, very important effect is built with to user's portrait, below
It is described in detail in conjunction with the application technical solution shown in Fig. 2.
In the technical solution that step S202 is provided, the first media information that the first account number was shared is obtained, wherein first
For media information for being shared between the account number of target application, the account number of target application includes the first account number.
The condition that triggering above-mentioned steps S202 is executed includes but is not limited to:
1) before issuing advertising media's promotion message, server shifts to an earlier date (such as executing according to the timing of set time) basis
The media information shared executes the present processes;
2) when user enables the client of target application, server executes the application according to the media information shared
Method;
3) for user when sharing the first media information in the client of target application, server is according to the media shared
Information executes the present processes.
An object of the application application can be instant messaging application, in this way, obtaining the first media letter that the first account number was shared
Breath be obtain the first account number publication, forwarding, sharing various types of status informations, specifically may include it is following at least it
One:
1) obtain the first account number shared be picture type the first media information;
2) obtain the first account number shared be video type the first media information;
3) obtain the first account number shared be audio types the first media information;
4) obtain the first account number shared be text type the first media information.
It should be noted that specifically, the first media information can be in picture, video, audio and text information
One kind, be also possible to two or more combination therein, such as audio and video, video and text.
For example, if the first media information is picture, then can with labeled picture come training machine learning model, into
And it can use the identification that training machine learning model completes the first media information of picture type;If the first media information is view
Frequently, the first media information can be handled for plurality of pictures (as interception video in picture frame), identification method just and picture category
First media information of type is identical;If the first media information is audio, then (can have audio types with labeled text
The first media information convert) carry out training machine learning model, using speech recognition technology by audio types when identification
The first media information be converted to text input machine learning model and identified, it is subsequent to be described in detail by taking picture as an example.
In the technical solution that step S204 is provided, the first account number belonging to the first account number is determined based on the first media information
Set;
In the embodiment of step S204, media information classification engine and machine learning model (i.e. the application can be passed through
First model) it realizes, machine learning model may include Embedding layers.Optionally, in step S204 or step S204
Before, can train in advance needs media information classification engine to be used and machine learning model.
By taking media information is image as an example, this method obtains the circle of friends image sent in user's a period of time recently first
Sequence (such as step S202), and user is obtained according to existing image classification model (media information classification engine) and tag along sort
Circle of friends image tag sequence in nearly a period of time, and send sequence label according to the circle of friends image of user and pick out and meet
User's (or image) of circle of friends advertisement tag (label of object) constructs training data, is trained.Specific step is as follows:
About the training of media information classification engine, the media letter of each media information in identifying multiple media informations
Before breath mark, the media information classification engine of media information mark for identification can be trained as follows:
Step S21, the media information for obtaining the second media information and the second media information identify (i.e. training data), herein
The second media information and the first media information belong to same major class (be such as picture, be video), specifically, second
Media information can be an image, and media information is identified as image tag, if the quantity of the second media information be it is multiple, it is right
The training data answered is the tag along sort sequence of image sequence and image.
Above-mentioned media information mark is that the content (or target object) for carrying to the second media information is identified
Information, such as the second media information is picture, has the contents such as " fried mutton slices with green scallion ", " dumbbell ", " mountain-climbing " in picture, then identifying and being
Information for being identified to " fried mutton slices with green scallion ", " dumbbell ", " mountain-climbing ".Target object herein can classify for media information
What engine succeeded in school in advance.
Step S22 draws the media information of the second media information and the second media information mark as media information classification
The input held up, so that media information classification engine learns to corresponding between characteristics of image and the media information mark of media information
Relationship.
It should be noted that the feature that above-mentioned media information classification engine learns can be diversified, such as figure
The label of picture is cuisines, then it is cuisines which characteristics of image media information classification engine, which may be learned, so that class occurs in next time
The label of " cuisines " can be stamped when like image for the image;Similarly, for the image of the labels such as electronic product, movement, media
Corresponding characteristics of image also may be learned in information classification engine.
Optionally, training media information classification engine and machine learning model, (training of machine learning model is rear
It is described in detail in text) after, step S204's determines that the first account number set can belonging to the first account number based on the first media information
It is achieved by the steps of:
Step S2041, being classified according to multiple account numbers of multiple media informations to target application, (multiple account numbers include the
One account number), obtain multiple account number set, multiple media informations include in the first media information and multiple account numbers except the first account number with
Outer account number carries out the media information shared between the account number of target application, namely including multiple account numbers in target application
The media information shared is carried out between account number.
Above-mentioned multiple account numbers are filtered out from all account numbers of target application, including but not limited to according to such as
One of under type a variety of is filtered: being filtered according to area belonging to account number, according to the circle of friends liveness of account number
(being determined by the duration logged in nearest a period of time, the quantity interacted in circle of friends etc.) is filtered, according to the chat of account number
Liveness is filtered.
Step S2042 searches the first account number set in multiple account number set where the first account number, such as according to the first account number
Account number mark (such as account number itself, account number pet name information) searched in each account number set, find including the account
Number mark set be the first account number set.
Optionally, classify according to multiple account numbers of multiple media informations to target application, obtain multiple account number collection
When conjunction, it can realize in accordance with the following steps:
Step S31 identifies each media information in multiple media informations by above-mentioned media information classification engine
Media information mark, specifically, can be using media information as the input of media information classification engine, exporting result is matchmaker
Body message identification.
Step S32 determines that each account number exists according to the media information of the media information of account number each in multiple account numbers mark
Characteristic value in multiple characteristic dimensions.It should be noted that feature of each account number obtained in multiple characteristic dimensions herein
Value can be determined by above-mentioned machine learning model.
It should be noted that schematically being carried out below to the scheme in step S32 for the ease of the understanding to the application
Description (feature of actually machine learning model study is not limited thereto, and can be the term vector feature of text): above-mentioned use
Family vector dimension specifically can according to the number or type of promotion message determine, as advertisement include mother and baby's product, electronic product,
Kitchen products, food, sports apparatus etc., then multiple characteristic dimensions can for mother and baby's product, electronic product, kitchen products,
The corresponding dimension such as food, sports apparatus, dimension values are used to characterize the user to the favorable rating of corresponding product, such as certain user
Repeatedly publication cuisines picture, then its characteristic value in this characteristic dimension of food is just higher, it is then lower on the contrary.
This means, the number of types or number determination of multiple characteristic dimensions and popularization object (have several classes or several popularization objects
Just have several characteristic dimensions), the feature of each characteristic dimension is used to describe the popularization object of one or one type, such as popularization pair
As for electronic product, then the feature of corresponding dimension is to be used to indicate the characteristics of image (brand of such as electronic product of electronic product
Mark, shape etc.), popularization object is sports apparatus, then the feature of corresponding dimension is to be used to indicate that the image of sports apparatus to be special
Sign.
Step S33 clusters multiple account numbers according to characteristic value of each account number in multiple characteristic dimensions, obtains more
A account number set.
The circle of friends of user can be sent to the sequence label input model of image, and extract Embedding layers of output
(characteristic value i.e. in multiple characteristic dimensions) is used as user vector, is clustered using Kmeans to user vector, thus right
User carries out a point group, obtains liking same or similar user group.
Step S206 provide technical solution in, using with the first promotion message of the first account number sets match as wait push away
The promotion message of the first account number is given, the first promotion message is used to indicate the first object to be promoted.
Optionally, using with the first promotion message of the first account number sets match as the popularization wait be pushed to the first account number
Before information, machine learning model (the second model) can be trained as follows:
Step S41, obtain third media information, third media information media information mark and third media information with
The matching degree of each promotion message in multiple promotion messages, wherein the account number of target application includes the 4th account number;Third herein
Media information is similar with the second media information, can also be exactly the second media information itself, can also be believed with identical media
For breath to train media information classification engine and machine learning model, third media information can be an image, media information
Be identified as image tag, if the quantity of third media information be it is multiple, corresponding training data be image sequence, image point
Class sequence label, advertisement tag sequence.
It should be noted that the first media information refers to the media information that the first account number was shared;Second media information is
Refer to the media information for training media information classification engine;Third media information refers to the matchmaker for training machine learning model
Body information.
Optionally, above-mentioned matching degree can be " 0 " and " 1 ", and " 0 " indicates product that the image and the advertisement are promoted not
Matching, " 1 " indicates that the image is matched with the product that the advertisement is promoted.
Step S42, by the media information of third media information, third media information mark and third media information and more
Input of the matching degree of each promotion message as the second model in a promotion message, so as to first network layer in the second model
It is initialized with the parameter in the second network layer, the first model after obtaining parameter initialization.
To machine learning model training complete after, using with the first promotion message of the first account number sets match as
When being pushed to the promotion message of the first account number, it can be accomplished in that
Step S2062, it is every from multiple account numbers using the first network layer (can be Embedding layers) of the first model
Characteristic value of the study to each account number in multiple characteristic dimensions in the media information mark of the media information of a account number.
Step S2064, using the second network layer (can be LSTM layers) of the first model, according to each account number in multiple spies
Characteristic value in sign dimension determines the matching degree of each promotion message in each account number and multiple promotion messages.
Step S2066 searches the first promotion message in multiple promotion messages, wherein in multiple account numbers each account number with
The sum of matching degree of first promotion message promotes letter not less than removing first in each account number in multiple account numbers and multiple promotion messages
The sum of matching degree of promotion message other than breath.
In the above-described embodiments, it when the circle of friends advertisement of instant communication applications is launched, can first the circle of friends to user send out
Image tag sequence inputting model is sent, and the output for extracting Embedding layers is clustered using Kmeans and calculated as user vector
Method clusters user vector, to carry out a point group to user, to belonging to the user in group's (namely account number set),
The score (namely matching degree) that each user corresponds to each advertisement tag is calculated using trained LSTM model, and institute is useful
Family sums (namely the sum of matching degree) to the score of each advertisement tag respectively, obtains the group and divides each advertisement tag
Number, the highest advertisement tag of score are then used as the dispensing advertisement of the group.
In the technical solution that step S208 is provided, to first the first promotion message of client push of the first account number, the
One client is the client of target application.
The first above-mentioned promotion message can be one of picture, text, video information, be also possible to it is therein at least
Two kinds of combination.
As shown in figure 3, can show the first above-mentioned promotion message in " advertising area " A at circle of friends interface, need
It is bright, the partial information (such as information of cover, summary, body matter) of " the first promotion message " can be only shown herein, when
It when user clicks the region, then completely shows the first promotion message, such as completely plays entire advertisement video, advertising pictures.
It can be realized in the manner illustrated in fig. 4 when completely playing the first promotion message, when the user clicks " the advertisement in Fig. 3
When the A of region ", an advertising area B is rebuild at the center of entire circle of friends, in the first promotion message of the complete broadcasting in the region.
Optionally, when executing step S208, server can push promotion message as unit of account number set, such as by first
Promotion message pushes to the client of all account numbers in the first account number set.
In embodiments herein, it is equivalent to and provides a kind of advertisement placement method for delivering image based on circle of friends,
This method can be used as a kind of new mode of circle of friends advertisement dispensing.By excavating the hobby of user, to construct user
Portrait, the similarity height drawn a portrait and launched between advertisement further according to user to carry out advertisement recommendation to user, here mainly
The feature used mainly includes the media information that user delivers in circle of friends, may also include the primary attribute feature of user, such as
Age, gender etc. may also include some users behavioural characteristic on network, such as web page browsing behavior etc..
The circle of friends of user delivers image, it will be appreciated that for a kind of embodiment of user interest hobby, passes through the friend to user
Loop graph picture is excavated, to construct user's portrait, is launched for circle of friends advertisement and is provided a kind of new mode.In addition, according to
The user of circle of friends picture construction draws a portrait, and can be used as foundation characteristic and is applied in existing advertisement dispensing, supplements as feature.
As a kind of optional embodiment, below to the specific reality that the application is described in detail so that media information is image information as an example
Apply mode.
This application provides a kind of advertisement placement methods based on circle of friends image, mainly launch for the advertisement of circle of friends
A kind of new putting mode is provided, this method obtains the circle of friends image sequence sent in user's a period of time recently first, and
The circle of friends image tag sequence in user's nearly a period of time is obtained according to trained image classification model and tag along sort, and
The user for meeting circle of friends advertisement tag is picked out according to the circle of friends image of user transmission sequence label, constructs training data.
Then Embedding dimensionality reduction is first carried out to the sequence of user again, using the LSTM construction classification temporal model of multi-tag two, passes through instruction
Get the multi-tag disaggregated model of user.
When circle of friends carries out advertisement dispensing, first by the transmission image tag sequence inputting model of the circle of friends of user, and
The output for extracting Embedding layers clusters user vector using Kmeans as user vector, thus to user into
Row divides group.To belonging to the user in group, each user is calculated using trained LSTM model and corresponds to each advertisements
The score of (i.e. the label of promotion message) is signed, and score of all users to each advertisement tag is summed respectively, obtains the group
Body is then used as the dispensing advertisement of the group to the score of each advertisement tag, the highest advertisement tag of score.It will retouch in detail below
State the implementation process of the technical solution.
Advertisement placement method based on circle of friends image can be divided into three phases, carry out specifically to these three stages below
It is bright.
(1) first stage, namely mark training data construct the stage
It is modeled in order to which the circle of friends to user sends image sequence, it can be in nearest a period of time of first structuring user's
Temporal aspect.The application utilizes circle of friends image classification engine (media information classification engine), carrys out friend's loop graph of structuring user's
As sequence label.Under specific implementation such as Fig. 5:
Step S501, extract user for a period of time in deliver circle of friends image configuration sequence S1 (also referred to as image sequence,
Multiple media informations including the user).
The circle of friends obtained in the nearest a period of time t of user delivers image, obtains image sequence S1, the image root of sequence S1
According to time order and function ordered arrangement, multiple images are once delivered, image sequence sequence when circle of friends, then benefit are delivered according to user
Classified with image classification engine to every image in S1, tag along sort (media information mark) is obtained, by tag along sort
Construction circle of friends delivers image tag sequence S.
Step S502, using image classification engine, according to image sequence S1 structuring user's image tag sequence S, image mark
Each label (namely media information mark) of label sequence is used to identify the content in a corresponding image.
Enabling image category sum in image classification engine is that N gives respectively to each of image sequence S1 image is delivered
A fixed tag along sort, such as baby's label, cuisines label etc., if i-th of image is classified engine in image sequence S1
It is divided into jth (1≤j≤N) class, then the value of S [i] is j in image tag sequence, to obtain image tag sequence S.
Step S503 selects the user data for meeting circle of friends advertisement tag according to user tag sequence as training number
According to.
Obtain user deliver image tag sequence S after, the use for meeting circle of friends advertisement tag is picked out according to the sequence
Family, such as mother and baby user, sports fan, food enthusiasts etc. stamp mother and baby's label, movement hobby mark to the user respectively
Label, cuisines like label etc., to construct training data.
Step S504, by the shorter sequence of all length in training data to the longest sequence polishing of length.
Since the image number sum that different user is delivered in nearest a period of time t may be inconsistent, each user's
Image tag sequence S length is inconsistent.Here enabling M, (such as 8, BIT1 to BIT8) is all user S sequence length maximum values, right
In the user tag sequence S for being shorter in length than M, then face mends 0 behind, so that the image tag sequence length of each user is consistent,
Schematic diagram can refer to Fig. 6, such as sequence label for account number 2, and latter three are mended 0.In addition, for delivering image in a period of time t
The excessive user of data, it is believed that the probability for belonging to difference quotient user is larger, this portion of user data can cause to do to model training
It disturbs, therefore filters out this portion of user data.
(2) second stage namely user model construct the stage
After obtaining training data, temporal model can be constructed and modeled to send circle of friends image sequence to user, from
And advertisement tag classification is carried out to user.The network structure design of model is as shown in Figure 7 and Figure 8: first network layer is
Embedding layers, mainly to transmission circle of friends image tag one Feature Mapping of Sequence Learning of user, and dimensionality reduction is carried out,
128 dimensions may be selected in Embedding layers of output size, and the second network layer is LSTM layers, and network output is 256 dimensions, third layer
For full articulamentum, the output size of this layer is advertisement tag number m, and activation primitive is that (Sigmoid is believing sigmoid function
In breath science, since it has the properties such as single increasing and the increasing of inverse function list, therefore Sigmoid function is used as the threshold of neural network
Value function, by variable mappings between 0 and 1) so that the output size of network is limited in section [0,1].In view of each use
Family is likely to be assigned in any one label, i.e., be between each advertisement tag it is independent, therefore loss function using pair
Number loss function binary_crossentropy (also referred to as cross entropy cost function), i.e., do two classification to each advertisement tag,
The output probability of network can regard that user matches the probability of the series advertisements as.
After obtaining user model, model can be trained using the training data of first stage mark, thus
To trained user's temporal model.
(3) stage is launched in phase III namely circle of friends advertisement
After obtaining trained temporal model, when circle of friends advertisement is launched, the circle of friends of user is first sent into image mark
Sequence inputting model is signed, and extracts Embedding layer of output as user vector, it is that user is special that this vector, which can be regarded,
A kind of expression of sign, clusters user vector using Kmeans, so that user is divided into k group, k is obtained by real data
It arrives.To belonging to the user in group, each user is calculated using trained LSTM model and corresponds to each advertisement tag
Score of the users all in group to the advertisement tag is summed for each group and each advertisement, obtains the group by score
Body is then used as the dispensing advertisement of the group, friend to the score of the advertisement tag, multiple highest advertisement tags of advertisement mid-score
Enclosing advertisement dispensing, specific step is as follows:
Step S51 obtains the user friend to be launched and encloses image sequence;
Image sequence is input in the trained user model of second stage by step S52, obtains Embedding layers
The m of 128 dimension outputs and last full articulamentum ties up output;
Embedding layers of 128 dimension outputs are clustered, obtain k user group by step S53;
Step S54 calculates the score summation under its each advertisement tag to belonging to the user in user group, point
The highest advertisement tag of number is then the advertisement that the user group should launch.
In the technical solution of the application, 1) a kind of advertisement placement method based on circle of friends image is provided, building is used
Family circle of friends image tag sequence picks out the user for meeting advertisement tag and tagged, to construct training data, then structure
It builds user's timing disaggregated model and sequence progress model training is delivered to the circle of friends image of user, and by Embedding layers of network
Output as user draw a portrait, the output of the full articulamentum of network the last layer is corresponded to the score of each advertisement, base as user
It draws a portrait in the user that Embedding layers export and carries out cluster grouping, calculate the ad score of each user group, to realize wide
It accuses and launches;2) feature by Embedding layers of user's timing disaggregated model of output as building user's portrait is supplemented, and utilizes use
A kind of feature representation of Embedding layers of the family timing disaggregated model output as user, can construct user's portrait, can also make
It is supplemented for the feature of other advertisement placement methods.
It should be noted that for the various method embodiments described above, for simple description, therefore, it is stated as a series of
Combination of actions, but those skilled in the art should understand that, the present invention is not limited by the sequence of acts described because
According to the present invention, some steps may be performed in other sequences or simultaneously.Secondly, those skilled in the art should also know
It knows, the embodiments described in the specification are all preferred embodiments, and related actions and modules is not necessarily of the invention
It is necessary.
Through the above description of the embodiments, those skilled in the art can be understood that according to above-mentioned implementation
The method of example can be realized by means of software and necessary general hardware platform, naturally it is also possible to by hardware, but it is very much
In the case of the former be more preferably embodiment.Based on this understanding, technical solution of the present invention is substantially in other words to existing
The part that technology contributes can be embodied in the form of software products, which is stored in a storage
In medium (such as ROM/RAM, magnetic disk, CD), including some instructions are used so that a terminal device (can be mobile phone, calculate
Machine, server or network equipment etc.) execute method described in each embodiment of the present invention.
Other side according to an embodiment of the present invention additionally provides a kind of for implementing the push side of above-mentioned promotion message
The driving means device of the promotion message of method.Fig. 9 is a kind of push dress of optional promotion message according to an embodiment of the present invention
Set the schematic diagram of device, as shown in figure 9, the apparatus may include: acquiring unit 901, determination unit 903, processing unit 905 with
And push unit 907.
Acquiring unit 901, the first media information shared for obtaining the first account number, wherein the first media information is used
Shared between the account number in target application, the account number of target application includes the first account number.
Determination unit 903, for determining the first account number set belonging to the first account number based on the first media information.
Processing unit 905, for using with the first promotion message of the first account number sets match as wait be pushed to the first account
Number promotion message, wherein the first promotion message is used to indicate the first object to be promoted.
Push unit 907, for pushing the first promotion message to the first account number.
During above-mentioned the first promotion message of confirmation, it can be run in the form of program code, without manually joining
With, therefore the efficiency searched with the matched promotion message type of the user can be greatly improved, meanwhile, as unit of account number set
Uniformly to determine that the promotion message of the set can be further improved matching efficiency;In addition, by going through thus according to the first account number
History information (the first media information shared) to select suitable first promotion message therefore can be maximum for it
Match with the hobby of user, such as meets film, the animation of its hobby, the learning Content of some aspects learnt
Deng the experience of user can be improved in which.
It should be noted that the acquiring unit 901 in the embodiment can be used for executing the step in the embodiment of the present application
S202, the determination unit 903 in the embodiment can be used for executing the step S204 in the embodiment of the present application, in the embodiment
Processing unit 905 can be used for executing the step S206 in the embodiment of the present application, and the push unit 907 in the embodiment can be used
Step S208 in execution the embodiment of the present application.
Herein it should be noted that above-mentioned module is identical as example and application scenarios that corresponding step is realized, but not
It is limited to above-described embodiment disclosure of that.It should be noted that above-mentioned module as a part of device may operate in as
In hardware environment shown in FIG. 1, hardware realization can also be passed through by software realization.
It can first obtain the first account number at determining promotion message matched with the first account number by above-mentioned module and share
The first media information, the first media information is based on the first media information for being shared between the account number of target application
Determine the first account number set belonging to the first account number;And using with the first promotion message of the first account number sets match as wait push
To the promotion message of the first account number, the first promotion message is used to indicate the first object to be promoted;To the first account number push first
Promotion message can solve and search when pushing promotion message to user in the related technology and the matched promotion message of the user
The lower technical problem of efficiency, and then reach the technical effect for improving and searching with the matched promotion message of user.
Above-mentioned determination unit can include: categorization module, for multiple accounts according to multiple media informations to target application
Number classify, obtain multiple account number set, wherein multiple media informations include in the first media information and multiple account numbers except the
Account number other than one account number carries out the media information shared between the account number of target application;First searching module, for looking into
Look for the first account number set in multiple account number set where the first account number.
Optionally, above-mentioned categorization module can include: identification submodule, for identification each matchmaker in multiple media informations out
The media information of body information identifies;Determine submodule, the media for the media information according to account number each in multiple account numbers are believed
Breath mark determines characteristic value of each account number in multiple characteristic dimensions;Classification submodule, is used for according to each account number multiple
Characteristic value in characteristic dimension clusters multiple account numbers, obtains multiple account number set.
Optionally, the device of the application may also include the first training unit, for every in identifying multiple media informations
Before the media information mark of a media information, the training media information that media information identifies for identification divides as follows
Class engine: the media information mark of the second media information and the second media information is obtained;By the second media information and the second media
The media information of information identifies the input as media information classification engine, so that the study of media information classification engine is believed to media
Corresponding relationship between characteristics of image and the media information mark of breath.
The processing unit of the application can include: study module, for the first network layer using the first model, from multiple accounts
Characteristic value of the study to each account number in multiple characteristic dimensions in the media information mark of the media information of each account number in number;
Determining module, it is true according to characteristic value of each account number in multiple characteristic dimensions for the second network layer using the first model
The matching degree of each promotion message in fixed each account number and multiple promotion messages;Second searching module, for searching multiple popularizations
The first promotion message in information, wherein the sum of each account number and the matching degree of the first promotion message are not less than in multiple account numbers
Each account number and the sum of the matching degree of promotion message in multiple promotion messages in addition to the first promotion message in multiple account numbers.
Optionally, the device of the application may also include the second training unit, for by with the first account number sets match
Before first promotion message is as the promotion message wait be pushed to the first account number, training obtains the first model as follows:
It is every in acquisition third media information, the media information mark of third media information and third media information and multiple promotion messages
The matching degree of a promotion message;By the media information mark and third media information of third media information, third media information
Input with the matching degree of promotion message each in multiple promotion messages as the second model, so as to the first net in the second model
Parameter in network layers and the second network layer is initialized, the first model after obtaining parameter initialization.
Optionally, above-mentioned target application includes instant messaging application, wherein acquiring unit can be used for executing it is following at least
One of: obtain the first account number shared be picture type the first media information;Obtaining the first account number to share is video
First media information of type;Obtain the first account number and shared be audio types the first media information.
In the technical solution of the application, 1) a kind of advertisement placement method based on circle of friends image is provided, building is used
Family circle of friends image tag sequence picks out the user for meeting advertisement tag and tagged, to construct training data, then structure
It builds user's timing disaggregated model and sequence progress model training is delivered to the circle of friends image of user, and by Embedding layers of network
Output as user draw a portrait, the output of the full articulamentum of network the last layer is corresponded to the score of each advertisement, base as user
It draws a portrait in the user that Embedding layers export and carries out cluster grouping, calculate the ad score of each user group, to realize wide
It accuses and launches;2) feature by Embedding layers of user's timing disaggregated model of output as building user's portrait is supplemented, and utilizes use
A kind of feature representation of Embedding layers of the family timing disaggregated model output as user, can construct user's portrait, can also make
It is supplemented for the feature of other advertisement placement methods.
Herein it should be noted that above-mentioned module is identical as example and application scenarios that corresponding step is realized, but not
It is limited to above-described embodiment disclosure of that.It should be noted that above-mentioned module as a part of device may operate in as
In hardware environment shown in FIG. 1, hardware realization can also be passed through by software realization, wherein hardware environment includes network
Environment.
Other side according to an embodiment of the present invention additionally provides a kind of for implementing the push side of above-mentioned promotion message
The server or terminal of method.
Figure 10 is a kind of structural block diagram of terminal according to an embodiment of the present invention, and as shown in Figure 10, which may include:
One or more (one is only shown in Figure 10) processors 1001, memory 1003 and (such as above-mentioned implementation of transmitting device 1005
Sending device in example), as shown in Figure 10, which can also include input-output equipment 1007.
Wherein, memory 1003 can be used for storing software program and module, such as the promotion message in the embodiment of the present invention
Method for pushing and the corresponding program instruction/module of device, processor 1001 by operation be stored in it is soft in memory 1003
Part program and module realize the push side of above-mentioned promotion message thereby executing various function application and data processing
Method.Memory 1003 may include high speed random access memory, can also include nonvolatile memory, such as one or more magnetism
Storage device, flash memory or other non-volatile solid state memories.In some instances, memory 1003 can further comprise
The memory remotely located relative to processor 1001, these remote memories can pass through network connection to terminal.Above-mentioned net
The example of network includes but is not limited to internet, intranet, local area network, mobile radio communication and combinations thereof.
Above-mentioned transmitting device 1005 is used to that data to be received or sent via network, can be also used for processor with
Data transmission between memory.Above-mentioned network specific example may include cable network and wireless network.In an example,
Transmitting device 1005 includes a network adapter (Network Interface Controller, NIC), can pass through cable
It is connected with other network equipments with router so as to be communicated with internet or local area network.In an example, transmission dress
1005 are set as radio frequency (Radio Frequency, RF) module, is used to wirelessly be communicated with internet.
Wherein, specifically, memory 1003 is for storing application program.
The application program that processor 1001 can call memory 1003 to store by transmitting device 1005, it is following to execute
Step:
Obtain the first media information that the first account number was shared, wherein the first media information is used for the account in target application
Shared between number, the account number of target application includes the first account number;
The first account number set belonging to the first account number is determined based on the first media information;
Using with the first promotion message of the first account number sets match as the promotion message wait be pushed to the first account number,
In, the first promotion message is used to indicate the first object to be promoted;
The first promotion message is pushed to the first account number.
Processor 1001 is also used to execute following step:
Using the first network layer of the first model, the media information mark of the media information of each account number from multiple account numbers
Characteristic value of the middle study to each account number in multiple characteristic dimensions;
Using the second network layer of the first model, determined according to characteristic value of each account number in multiple characteristic dimensions each
The matching degree of each promotion message in account number and multiple promotion messages;
Search the first promotion message in multiple promotion messages, wherein each account number is promoted with first and believed in multiple account numbers
The sum of matching degree of breath is not less than each account number and pushing away in addition to the first promotion message in multiple promotion messages in multiple account numbers
The sum of the matching degree of Guangxin breath.
First account number point can first be obtained in determining promotion message matched with the first account number using the embodiment of the present invention
The first media information enjoyed, the first media information are based on the first media for being shared between the account number of target application
Information determines the first account number set belonging to the first account number;And using with the first promotion message of the first account number sets match as to
It is pushed to the promotion message of the first account number, the first promotion message is used to indicate the first object to be promoted;It is pushed to the first account number
First promotion message can solve to search when pushing promotion message to user in the related technology and believe with the matched popularization of the user
The lower technical problem of the efficiency of breath, and then reach the technical effect for improving and searching with the matched promotion message of user.
Optionally, the specific example in the present embodiment can be with reference to example described in above-described embodiment, the present embodiment
Details are not described herein.
It will appreciated by the skilled person that structure shown in Fig. 10 is only to illustrate, terminal can be smart phone
(such as Android phone, iOS mobile phone), tablet computer, palm PC and mobile internet device (Mobile Internet
Devices, MID), the terminal devices such as PAD.Figure 10 it does not cause to limit to the structure of above-mentioned electronic device.For example, terminal is also
May include than shown in Figure 10 more perhaps less component (such as network interface, display device) or have and Figure 10 institute
Show different configurations.
Those of ordinary skill in the art will appreciate that all or part of the steps in the various methods of above-described embodiment is can
It is completed with instructing the relevant hardware of terminal device by program, which can store in a computer readable storage medium
In, storage medium may include: flash disk, read-only memory (Read-Only Memory, ROM), random access device (Random
Access Memory, RAM), disk or CD etc..
The embodiments of the present invention also provide a kind of storage mediums.Optionally, in the present embodiment, above-mentioned storage medium can
With the program code of the method for pushing for executing promotion message.
Optionally, in the present embodiment, above-mentioned storage medium can be located at multiple in network shown in above-described embodiment
On at least one network equipment in the network equipment.
Optionally, in the present embodiment, storage medium is arranged to store the program code for executing following steps:
S61 obtains the first media information that the first account number was shared, wherein the first media information is used in target application
Account number between shared, the account number of target application includes the first account number;
S62 determines the first account number set belonging to the first account number based on the first media information;
S63, using with the first promotion message of the first account number sets match as the promotion message wait be pushed to the first account number,
Wherein, the first promotion message is used to indicate the first object to be promoted;
S64 pushes the first promotion message to the first account number.
Optionally, storage medium is also configured to store the program code for executing following steps:
S71, using the first network layer of the first model, the media information of the media information of each account number from multiple account numbers
Characteristic value of the study to each account number in multiple characteristic dimensions in mark;
S72 is determined using the second network layer of the first model according to characteristic value of each account number in multiple characteristic dimensions
The matching degree of each promotion message in each account number and multiple promotion messages;
S73 searches the first promotion message in multiple promotion messages, wherein each account number is pushed away with first in multiple account numbers
The sum of matching degree of Guangxin breath not less than in each account number in multiple account numbers and multiple promotion messages in addition to the first promotion message
The sum of the matching degree of promotion message.
Optionally, the specific example in the present embodiment can be with reference to example described in above-described embodiment, the present embodiment
Details are not described herein.
Optionally, in the present embodiment, above-mentioned storage medium can include but is not limited to: USB flash disk, read-only memory (ROM,
Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or
The various media that can store program code such as CD.
The serial number of the above embodiments of the invention is only for description, does not represent the advantages or disadvantages of the embodiments.
If the integrated unit in above-described embodiment is realized in the form of SFU software functional unit and as independent product
When selling or using, it can store in above-mentioned computer-readable storage medium.Based on this understanding, skill of the invention
Substantially all or part of the part that contributes to existing technology or the technical solution can be with soft in other words for art scheme
The form of part product embodies, which is stored in a storage medium, including some instructions are used so that one
Platform or multiple stage computers equipment (can be personal computer, server or network equipment etc.) execute each embodiment institute of the present invention
State all or part of the steps of method.
In the above embodiment of the invention, it all emphasizes particularly on different fields to the description of each embodiment, does not have in some embodiment
The part of detailed description, reference can be made to the related descriptions of other embodiments.
In several embodiments provided herein, it should be understood that disclosed client, it can be by others side
Formula is realized.Wherein, the apparatus embodiments described above are merely exemplary, such as the division of the unit, and only one
Kind of logical function partition, there may be another division manner in actual implementation, for example, multiple units or components can combine or
It is desirably integrated into another system, or some features can be ignored or not executed.Another point, it is shown or discussed it is mutual it
Between coupling, direct-coupling or communication connection can be through some interfaces, the INDIRECT COUPLING or communication link of unit or module
It connects, can be electrical or other forms.
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 software functional units.
The above is only a preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art
For member, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications are also answered
It is considered as protection scope of the present invention.
Claims (13)
1. a kind of method for pushing of promotion message characterized by comprising
Obtain the first media information that the first account number was shared, wherein first media information is used for the account in target application
Shared between number, the account number of the target application includes first account number;
The first account number set belonging to first account number is determined based on first media information;
Using with the first promotion message of the first account number sets match as the promotion message wait be pushed to first account number,
Wherein, first promotion message is used to indicate the first object to be promoted;
First promotion message is pushed to first account number.
2. the method according to claim 1, wherein determining first account number based on first media information
The first affiliated account number set includes:
Classified according to multiple account numbers of multiple media informations to the target application, obtains multiple account number set, wherein institute
Stating multiple media informations includes that the multiple account number carries out the media information shared, institute between the account number of the target application
Stating multiple account numbers includes first account number;
Search the first account number set where the first account number described in the multiple account number set.
3. according to the method described in claim 2, it is characterized in that, according to multiple media informations to the multiple of the target application
Account number is classified, and is obtained multiple account number set and is included:
Identify the media information mark of each media information in the multiple media information;
Determine each account number multiple according to the media information of the media information of account number each in the multiple account number mark
Characteristic value in characteristic dimension;
The multiple account number is clustered according to characteristic value of each account number in multiple characteristic dimensions, is obtained described more
A account number set.
4. according to the method described in claim 3, it is characterized in that, each media are believed in identifying the multiple media information
Before the media information mark of breath, the method also includes: the training media information point that the media information identifies for identification
Class engine, comprising:
Obtain the media information mark of the second media information and second media information;
The media information of second media information and second media information is identified and is drawn as media information classification
The input held up, so that media information classification engine study is between characteristics of image and the media information mark of media information
Corresponding relationship.
5. method as claimed in any of claims 2 to 4, which is characterized in that will be with the first account number set
The first promotion message matched includes: as the promotion message wait be pushed to first account number
Using the first network layer of the first model, the media information mark of the media information of each account number from the multiple account number
Characteristic value of the middle study to each account number in multiple characteristic dimensions;
Using the second network layer of first model, determined according to characteristic value of each account number in multiple characteristic dimensions
The matching degree of each promotion message in each account number and multiple promotion messages;
Search first promotion message in the multiple promotion message, wherein each account number and institute in the multiple account number
The sum of matching degree of the first promotion message is stated not less than removing in each account number in the multiple account number and the multiple promotion message
The sum of the matching degree of promotion message other than first promotion message.
6. according to the method described in claim 5, it is characterized in that, will be promoted with the first of the first account number sets match
Before information is as the promotion message wait be pushed to first account number, the method also includes:
Obtain third media information, the third media information media information mark and the third media information with it is described
The matching degree of each promotion message in multiple promotion messages;
By the media information of third media information, third media information mark and the third media information and described more
Input of the matching degree of each promotion message as the second model in a promotion message, so as to the first net in second model
Parameter in network layers and the second network layer is initialized, first model after obtaining parameter initialization.
7. the method according to claim 1, wherein the target application includes instant messaging application, wherein obtain
The first media information for taking the first account number to share includes at least one of:
Obtain first account number and shared be picture type first media information;
Obtain first account number and shared be video type first media information;
Obtain first account number and shared be audio types first media information.
8. a kind of driving means of promotion message characterized by comprising
Acquiring unit, the first media information shared for obtaining the first account number, wherein first media information is used for
Shared between the account number of target application, the account number of the target application includes first account number;
Determination unit, for determining the first account number set belonging to first account number based on first media information;
Processing unit, for using with the first promotion message of the first account number sets match as wait be pushed to first account
Number promotion message, wherein first promotion message is used to indicate the first object to be promoted;
Push unit, for pushing first promotion message to first account number.
9. device according to claim 8, which is characterized in that the determination unit includes:
Categorization module obtains multiple accounts for classifying according to multiple account numbers of multiple media informations to the target application
Number set, wherein the multiple media information includes that the multiple account number is shared between the account number of the target application
The media information crossed, the multiple account number include first account number;
First searching module, for searching the first account number collection where the first account number described in the multiple account number set
It closes.
10. device according to claim 9, which is characterized in that the categorization module includes:
Identify submodule, the media information of each media information identifies in the multiple media information out for identification;
Submodule is determined, for according to the mark determination of the media information of the media information of account number each in the multiple account number
Characteristic value of each account number in multiple characteristic dimensions;
Classification submodule, for being carried out according to characteristic value of each account number in multiple characteristic dimensions to the multiple account number
Cluster, obtains the multiple account number set.
11. device according to claim 9 or 10, which is characterized in that the processing unit includes:
Study module, for the first network layer using the first model, the media information of each account number from the multiple account number
Media information mark in characteristic value of the study to each account number in multiple characteristic dimensions;
Determining module, for the second network layer using first model, according to each account number in multiple characteristic dimensions
On characteristic value determine the matching degree of each promotion message in each account number and multiple promotion messages;
Second searching module, for searching first promotion message in the multiple promotion message, wherein the multiple account
In number the sum of each account number and the matching degree of first promotion message not less than each account number in the multiple account number with it is described
The sum of the matching degree of promotion message in multiple promotion messages in addition to first promotion message.
12. a kind of storage medium, which is characterized in that the storage medium includes the program of storage, wherein when described program is run
Execute method described in 1 to 7 any one of the claims.
13. a kind of electronic device, including memory, processor and it is stored on the memory and can transports on the processor
Capable computer program, which is characterized in that the processor executes the claims 1 to 7 by the computer program
Method described in one.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711217712.9A CN109840793B (en) | 2017-11-28 | 2017-11-28 | Promotion information pushing method and device, storage medium and electronic device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711217712.9A CN109840793B (en) | 2017-11-28 | 2017-11-28 | Promotion information pushing method and device, storage medium and electronic device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109840793A true CN109840793A (en) | 2019-06-04 |
CN109840793B CN109840793B (en) | 2023-11-17 |
Family
ID=66881165
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711217712.9A Active CN109840793B (en) | 2017-11-28 | 2017-11-28 | Promotion information pushing method and device, storage medium and electronic device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109840793B (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110516153A (en) * | 2019-08-28 | 2019-11-29 | 腾讯科技(深圳)有限公司 | Intelligently pushing method and apparatus, storage medium and the electronic device of video |
CN110555736A (en) * | 2019-09-05 | 2019-12-10 | 广东聚宝屋传媒有限公司 | control method and device for advertising based on published demand information |
CN110879851A (en) * | 2019-10-15 | 2020-03-13 | 北京三快在线科技有限公司 | Video dynamic cover generation method and device, electronic equipment and readable storage medium |
CN111598159A (en) * | 2020-05-14 | 2020-08-28 | 清华大学 | Training method, device, equipment and storage medium of machine learning model |
CN112035613A (en) * | 2020-08-28 | 2020-12-04 | 光大科技有限公司 | Information pushing method and device based on multiple rounds of conversations, storage medium and electronic device |
CN113010777A (en) * | 2021-03-05 | 2021-06-22 | 腾讯科技(深圳)有限公司 | Data pushing method, device, equipment and storage medium |
CN113240468A (en) * | 2021-05-13 | 2021-08-10 | 北京沃东天骏信息技术有限公司 | Information processing method, device and storage medium |
CN113886713A (en) * | 2021-12-09 | 2022-01-04 | 北京达佳互联信息技术有限公司 | Data processing method, device, server and storage medium |
WO2022206103A1 (en) * | 2021-04-01 | 2022-10-06 | 重庆邮电大学 | User portrait construction method based on microblog heterogeneous information |
CN116485475A (en) * | 2023-05-06 | 2023-07-25 | 湖北巨字传媒有限公司 | Internet of things advertisement system, method and device based on edge calculation |
CN116521908A (en) * | 2023-06-28 | 2023-08-01 | 图林科技(深圳)有限公司 | Multimedia content personalized recommendation method based on artificial intelligence |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100262501A1 (en) * | 2007-12-16 | 2010-10-14 | Softbank Corp. | Advertisement system and advertising method |
CN102739698A (en) * | 2011-04-01 | 2012-10-17 | 上海亿动信息技术有限公司 | Method for using community sharing to assist mobile terminal advertisement publishing and system thereof |
CN104766080A (en) * | 2015-05-06 | 2015-07-08 | 苏州搜客信息技术有限公司 | Image multi-class feature recognizing and pushing method based on electronic commerce |
CN104992342A (en) * | 2015-05-11 | 2015-10-21 | 腾讯科技(北京)有限公司 | Method for determining the effectiveness of popularizing information release, monitoring server, and terminal |
CN105847114A (en) * | 2016-03-21 | 2016-08-10 | 腾讯科技(深圳)有限公司 | Dynamic information display system, method and device |
CN105912650A (en) * | 2016-04-07 | 2016-08-31 | 腾讯科技(深圳)有限公司 | Method and apparatus of recommending songs |
CN106055617A (en) * | 2016-05-26 | 2016-10-26 | 乐视控股(北京)有限公司 | Data pushing method and device |
CN106357517A (en) * | 2016-09-27 | 2017-01-25 | 腾讯科技(北京)有限公司 | Directional label generation method and device |
CN106446078A (en) * | 2016-09-08 | 2017-02-22 | 乐视控股(北京)有限公司 | Information recommendation method and recommendation apparatus |
CN106612230A (en) * | 2015-10-27 | 2017-05-03 | 腾讯科技(北京)有限公司 | Media information promotion method, client and server |
-
2017
- 2017-11-28 CN CN201711217712.9A patent/CN109840793B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100262501A1 (en) * | 2007-12-16 | 2010-10-14 | Softbank Corp. | Advertisement system and advertising method |
CN102739698A (en) * | 2011-04-01 | 2012-10-17 | 上海亿动信息技术有限公司 | Method for using community sharing to assist mobile terminal advertisement publishing and system thereof |
CN104766080A (en) * | 2015-05-06 | 2015-07-08 | 苏州搜客信息技术有限公司 | Image multi-class feature recognizing and pushing method based on electronic commerce |
CN104992342A (en) * | 2015-05-11 | 2015-10-21 | 腾讯科技(北京)有限公司 | Method for determining the effectiveness of popularizing information release, monitoring server, and terminal |
CN106612230A (en) * | 2015-10-27 | 2017-05-03 | 腾讯科技(北京)有限公司 | Media information promotion method, client and server |
CN105847114A (en) * | 2016-03-21 | 2016-08-10 | 腾讯科技(深圳)有限公司 | Dynamic information display system, method and device |
CN105912650A (en) * | 2016-04-07 | 2016-08-31 | 腾讯科技(深圳)有限公司 | Method and apparatus of recommending songs |
CN106055617A (en) * | 2016-05-26 | 2016-10-26 | 乐视控股(北京)有限公司 | Data pushing method and device |
CN106446078A (en) * | 2016-09-08 | 2017-02-22 | 乐视控股(北京)有限公司 | Information recommendation method and recommendation apparatus |
CN106357517A (en) * | 2016-09-27 | 2017-01-25 | 腾讯科技(北京)有限公司 | Directional label generation method and device |
Non-Patent Citations (1)
Title |
---|
王培志;: "场景视频的H5创新--以央广两会H5"王小艺的朋友圈"为例", 中国报业, no. 07 * |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110516153A (en) * | 2019-08-28 | 2019-11-29 | 腾讯科技(深圳)有限公司 | Intelligently pushing method and apparatus, storage medium and the electronic device of video |
CN110516153B (en) * | 2019-08-28 | 2023-05-30 | 腾讯科技(深圳)有限公司 | Intelligent video pushing method and device, storage medium and electronic device |
CN110555736B (en) * | 2019-09-05 | 2022-06-03 | 广东聚宝屋传媒有限公司 | Control method and device for advertising based on published demand information |
CN110555736A (en) * | 2019-09-05 | 2019-12-10 | 广东聚宝屋传媒有限公司 | control method and device for advertising based on published demand information |
CN110879851A (en) * | 2019-10-15 | 2020-03-13 | 北京三快在线科技有限公司 | Video dynamic cover generation method and device, electronic equipment and readable storage medium |
CN111598159A (en) * | 2020-05-14 | 2020-08-28 | 清华大学 | Training method, device, equipment and storage medium of machine learning model |
CN111598159B (en) * | 2020-05-14 | 2024-04-26 | 清华大学 | Training method, device, equipment and storage medium of machine learning model |
CN112035613A (en) * | 2020-08-28 | 2020-12-04 | 光大科技有限公司 | Information pushing method and device based on multiple rounds of conversations, storage medium and electronic device |
CN112035613B (en) * | 2020-08-28 | 2023-09-12 | 光大科技有限公司 | Information pushing method and device based on multi-round dialogue, storage medium and electronic device |
CN113010777A (en) * | 2021-03-05 | 2021-06-22 | 腾讯科技(深圳)有限公司 | Data pushing method, device, equipment and storage medium |
WO2022206103A1 (en) * | 2021-04-01 | 2022-10-06 | 重庆邮电大学 | User portrait construction method based on microblog heterogeneous information |
CN113240468A (en) * | 2021-05-13 | 2021-08-10 | 北京沃东天骏信息技术有限公司 | Information processing method, device and storage medium |
CN113886713A (en) * | 2021-12-09 | 2022-01-04 | 北京达佳互联信息技术有限公司 | Data processing method, device, server and storage medium |
CN116485475A (en) * | 2023-05-06 | 2023-07-25 | 湖北巨字传媒有限公司 | Internet of things advertisement system, method and device based on edge calculation |
CN116521908A (en) * | 2023-06-28 | 2023-08-01 | 图林科技(深圳)有限公司 | Multimedia content personalized recommendation method based on artificial intelligence |
CN116521908B (en) * | 2023-06-28 | 2024-01-09 | 图林科技(深圳)有限公司 | Multimedia content personalized recommendation method based on artificial intelligence |
Also Published As
Publication number | Publication date |
---|---|
CN109840793B (en) | 2023-11-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109840793A (en) | Method for pushing and device, storage medium, the electronic device of promotion message | |
CN110166827B (en) | Video clip determination method and device, storage medium and electronic device | |
CN110737783B (en) | Method and device for recommending multimedia content and computing equipment | |
CN110909176A (en) | Data recommendation method and device, computer equipment and storage medium | |
CN110035314A (en) | Methods of exhibiting and device, storage medium, the electronic device of information | |
CN109104639A (en) | Live broadcast system, the method, apparatus and electronic equipment for determining live video theme | |
CN108694602A (en) | Promotional literature generation method and device | |
CN106204165A (en) | A kind of advertisement placement method and device | |
CN102073956A (en) | Data mining-based directional advertisement release method, system and equipment | |
CN108108996A (en) | Advertisement placement method, device, computer equipment and readable medium in video | |
CN108959323B (en) | Video classification method and device | |
CN113382301A (en) | Video processing method, storage medium and processor | |
CN108876430B (en) | Advertisement pushing method based on crowd characteristics, electronic equipment and storage medium | |
CN109947510A (en) | A kind of interface recommended method and device, computer equipment | |
US20150131967A1 (en) | Computerized systems and methods for generating models for identifying thumbnail images to promote videos | |
CN110008999A (en) | Determination method, apparatus, storage medium and the electronic device of target account number | |
CN111259257A (en) | Information display method, system, device, electronic equipment and storage medium | |
CN111144936B (en) | Similar crowd expansion method and device based on user labels | |
CN116894711A (en) | Commodity recommendation reason generation method and device and electronic equipment | |
CN114501105B (en) | Video content generation method, device, equipment and storage medium | |
CN110197386A (en) | Media resource method for pushing and device, storage medium and electronic device | |
CN110020123B (en) | Popularization information delivery method, device, medium and equipment | |
CN114283349A (en) | Data processing method and device, computer equipment and storage medium | |
WO2021184153A1 (en) | Summary video generation method and device, and server | |
CN111787042B (en) | Method and device for pushing information |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |