CN110335100A - Information-pushing method, device, computer and storage medium based on artificial intelligence - Google Patents
Information-pushing method, device, computer and storage medium based on artificial intelligence Download PDFInfo
- Publication number
- CN110335100A CN110335100A CN201910384165.6A CN201910384165A CN110335100A CN 110335100 A CN110335100 A CN 110335100A CN 201910384165 A CN201910384165 A CN 201910384165A CN 110335100 A CN110335100 A CN 110335100A
- Authority
- CN
- China
- Prior art keywords
- user
- information
- pushed
- user behavior
- behavior
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
Abstract
This application involves a kind of information-pushing method based on artificial intelligence, device, computer and storage mediums.This method comprises: obtaining user behavior;User behavior importing user model is learnt, the multidimensional characteristic of user is obtained;Obtain information to be pushed;Information to be pushed is parsed, the feature of information to be pushed is obtained;It by the multidimensional characteristic of the feature of information to be pushed and user, is input in the deep neural network model constructed and is predicted, obtain the prediction probability of each information to be pushed;Based on user behavior, information to be pushed is pushed to user according to obtained prediction probability.The feature of the multidimensional characteristic of user and pushed information is input in neural network model and is predicted, obtain prediction probability, then it is based on user behavior, according to prediction probability to user's pushed information, pushed information is also related to user behavior, pushed information when without operation is avoided to user, effectively avoids that the use of user is caused to perplex.
Description
Technical field
This application involves information advancing technique fields, more particularly to a kind of information-pushing method, device, computer and deposit
Storage media.
Background technique
With the high speed development of the development of the high speed of internet, especially mobile Internet, the various of mobile Internet are answered
With the free life of extreme enrichment people.People can do shopping, entertain and consume with staying indoors.
For the usual consumption habit of user, the main of the push of various applications is had become for user's push consumption information
Method, however at present for user when liking carrying out PUSH message, however it remains the more inaccurate situation of push gives user
Bring biggish puzzlement.
Summary of the invention
Based on this, it is necessary in view of the above technical problems, provide a kind of information-pushing method based on artificial intelligence, dress
It sets, computer and storage medium.
A kind of information-pushing method based on artificial intelligence, which comprises
Obtain user behavior;
User behavior importing user model is learnt, the multidimensional characteristic of user is obtained;
Obtain information to be pushed;
The information to be pushed is parsed, the feature of the information to be pushed is obtained;
By the multidimensional characteristic of the feature of the information to be pushed and user, it is input to the deep neural network constructed
It is predicted in model, obtains the prediction probability of each information to be pushed;
Based on the user behavior, the information to be pushed is pushed to user according to the obtained prediction probability.
It is described in one of the embodiments, to learn user behavior importing user model, obtain user's
The step of multidimensional characteristic includes:
The user model is directed into using the user behavior as sample behavior;
It is trained to described for model, the user model after being trained;
The user model that the user behavior imports after training is learnt, the multidimensional characteristic of user is obtained.
It is described in one of the embodiments, to learn user behavior importing user model, obtain user's
Multidimensional characteristic includes:
User behavior importing user model is learnt, at least one multidimensional characteristic of user is obtained;
Obtain the incidence relation between multiple multidimensional characteristics;
According to the incidence relation and at least one multidimensional characteristic between multiple multidimensional characteristics, the multiple of user are obtained
The multidimensional characteristic.
It is described in one of the embodiments, to be based on the user behavior, according to the obtained prediction probability to user
The step of pushing the information to be pushed include:
Obtain push strategy;
Based on push strategy and the user behavior, according to the obtained prediction probability to user's push it is described to
Pushed information.
It is described in one of the embodiments, to be based on the user behavior, according to the obtained prediction probability to user
The step of pushing the information to be pushed include:
The frequency that type and the user behavior based on the user behavior occur, according to the obtained prediction probability
The information to be pushed is pushed to user.
A kind of information push-delivery apparatus based on artificial intelligence, described device include:
User behavior obtains module, for obtaining user behavior;
Multidimensional characteristic obtains module, for learning user behavior importing user model, obtains the more of user
Dimensional feature;
Pushed information obtains module, for obtaining information to be pushed;
Information characteristics obtain module and obtain the feature of the information to be pushed for parsing the information to be pushed;
Prediction module, for being input to the multidimensional characteristic of the feature of the information to be pushed and user and having constructed
Deep neural network model in predicted, obtain the prediction probability of each information to be pushed;
Info push module pushes institute to user according to the obtained prediction probability for being based on the user behavior
State information to be pushed.
The multidimensional characteristic acquisition module includes: in one of the embodiments,
Import unit, for being directed into the user model for the user behavior as sample behavior;
Training unit, for being trained to described for model, the user model after being trained;
Multidimensional characteristic acquiring unit, the user model for importing the user behavior after training learn,
Obtain the multidimensional characteristic of user.
The multidimensional characteristic acquisition module includes: in one of the embodiments,
Fisrt feature acquiring unit obtains user extremely for learning user behavior importing user model
A few multidimensional characteristic;
Incidence relation acquiring unit, for obtaining the incidence relation between multiple multidimensional characteristics;
Second feature acquiring unit, for according between multiple multidimensional characteristics incidence relation and at least one is more
Dimensional feature obtains multiple multidimensional characteristics of user.
A kind of computer including memory, processor and stores the calculating that can be run on a memory and on a processor
Machine program, the processor perform the steps of when executing the computer program
Obtain user behavior;
User behavior importing user model is learnt, the multidimensional characteristic of user is obtained;
Obtain information to be pushed;
The information to be pushed is parsed, the feature of the information to be pushed is obtained;
By the multidimensional characteristic of the feature of the information to be pushed and user, it is input to the deep neural network constructed
It is predicted in model, obtains the prediction probability of each information to be pushed;
Based on the user behavior, the information to be pushed is pushed to user according to the obtained prediction probability.
A kind of computer readable storage medium, is stored thereon with computer program, and the computer program is held by processor
It is performed the steps of when row
Obtain user behavior;
User behavior importing user model is learnt, the multidimensional characteristic of user is obtained;
Obtain information to be pushed;
The information to be pushed is parsed, the feature of the information to be pushed is obtained;
By the multidimensional characteristic of the feature of the information to be pushed and user, it is input to the deep neural network constructed
It is predicted in model, obtains the prediction probability of each information to be pushed;
Based on the user behavior, the information to be pushed is pushed to user according to the obtained prediction probability.
The above-mentioned information-pushing method based on artificial intelligence, device, computer and storage medium, by the multidimensional characteristic of user
It is input in neural network model and predicts with the feature of pushed information, obtain prediction probability, user behavior is then based on, according to pre-
Probability is surveyed to user's pushed information, since pushed information is pushed according to prediction probability, enables the information of push effective
Meet the behavioural habits of user, in addition, pushed information is also related to user behavior, avoids to user push letter when without operation
Breath, effectively avoids causing to perplex to the use of user.
Detailed description of the invention
Fig. 1 is the applied environment figure of the information-pushing method based on artificial intelligence in one embodiment;
Fig. 2 is the flow diagram of the information-pushing method based on artificial intelligence in one embodiment;
Fig. 3 is the structural block diagram of the information push-delivery apparatus based on artificial intelligence in one embodiment;
Fig. 4 is the internal structure chart of computer equipment in one embodiment.
Specific embodiment
It is with reference to the accompanying drawings and embodiments, right in order to which the objects, technical solutions and advantages of the application are more clearly understood
The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, not
For limiting the application.
Information-pushing method provided by the present application based on artificial intelligence, can be applied to application environment as shown in Figure 1
In.Wherein, terminal 102 with server 104 by network by being communicated.Wherein, terminal 102 can be, but not limited to be various
Personal computer, laptop, smart phone, tablet computer and portable wearable device, server 104 can be with independently
The server cluster of server either multiple servers composition realize.Assemble and interlinking interconnection net application program in terminal 102 is used
Family is operated by the internet application in terminal 102, which is user behavior, and user passes through terminal
Internet application on 102 operates and accesses server 104, and server 104 is obtained and used by communicating with terminal 102
User behavior importing user model is learnt, obtains the multidimensional characteristic of user by family behavior;Then, it obtains wait push
Information;The information to be pushed is parsed, the feature of the information to be pushed is obtained;By the feature of the information to be pushed and user
The multidimensional characteristic, be input in the deep neural network model constructed and predicted, obtain each information to be pushed
Prediction probability;Based on the user behavior, the information to be pushed is pushed to terminal 102 according to the obtained prediction probability.
In one embodiment, as shown in Fig. 2, a kind of information-pushing method based on artificial intelligence is provided, with the party
Method is applied to be illustrated for the terminal in Fig. 1, comprising the following steps:
Step 201, user behavior is obtained.
Specifically, the operation that user behavior, that is, user is carried out by application program, the operation include clicking, sliding, fingerprint
Verifying, face verification, camera shooting etc..For example, clicking the icon in application program, button, the sliding button in application program is slided.
The user behavior is used to match the preset rules in application program, and then obtains the default feedback of application program.For example, the application
Program is to consume the application in store, the user behavior can be in consumption store the checking of commodity, collect, forward, pay the bill,
Purchase etc..
Step 203, user behavior importing user model is learnt, obtains the multidimensional characteristic of user.
Specifically, which is the model that acquisition is trained based on user behavior, which is to instruct in advance
It gets.Study by the user model to user behavior can obtain the multidimensional characteristic of user.The multidimensional characteristic includes
The feature of user behavior and the personal characteristics of user, for example, the multidimensional characteristic includes the commodity of user's payment amount, user's browsing
Type, user collection type of merchandize, collection Brand.The multidimensional characteristic of the user be constitute user behavior it is main because
Element can obtain the usual consumption habit or consumption propensity of user according to the multidimensional characteristic.
The multidimensional characteristic can also can dispense range, validity information, packaging including user group's feature, information (commodity)
Feature, unique identification, other information Relating Characteristic etc..These multidimensional characteristics can learn from the user behavior of user
It obtains.
Step 205, information to be pushed is obtained.
Specifically, it should can be considered the page to be pushed to PUSH message, should can be considered advertisement information to PUSH message.It is worth
One is mentioned that, to PUSH message can be one either it is multiple.In the present embodiment, to PUSH message be it is multiple, it is multiple wait push away
It send message that can be pushed respectively for different users, different user behaviors, improves push precision.
Step 207, the information to be pushed is parsed, the feature of the information to be pushed is obtained.
Specifically, the feature of information to be pushed is matched with the multidimensional characteristic of user, i.e. the feature of information to be pushed and user
Multidimensional characteristic have same alike result or type it is identical.The feature of information to be pushed includes commodity price, type of merchandize, quotient
The brand etc. of product, for example, the commodity price in the feature of information to be pushed can be with the payment amount in the multidimensional characteristic of user
Corresponding, the type of merchandize in the feature of information to be pushed can be with the commodity kind of collection or browsing in the multidimensional characteristic of user
Class is corresponding, and the brand of the commodity in the feature of information to be pushed can be with the Brand pair of the collection in the multidimensional characteristic of user
It answers.In this way, the feature of information to be pushed is corresponding with the multidimensional characteristic of user, convenient for the matching of information to be pushed and user behavior,
So that information to be pushed can be pushed accurately.
Step 209, by the multidimensional characteristic of the feature of the information to be pushed and user, it is input to the depth constructed
It is predicted in neural network model, obtains the prediction probability of each information to be pushed.
Specifically, the deep neural network model of the building is capable of the multidimensional characteristic of feature and user to information to be pushed
Matching degree predicted, so that the prediction probability of each information to be pushed be calculated.In the present embodiment, building one in advance
Deep neural network (DNN) model, is trained using deep neural network model of a large amount of sample data to building, training
An available trained deep neural network model after the completion.The sample data is user behavior.
It is noted that building deep neural network model can be implemented by using the prior art, and the depth nerve net
The framework of network model can also use the prior art.For example, the deep neural network model can be used such as Chinese patent
Mode disclosed in 201611228253.X is constructed, and the structure of the deep neural network model can also be used as Chinese special
Framework disclosed in sharp 201611228253.X.Not burdensome description to this in the present embodiment.
Step 211, it is based on the user behavior, the letter to be pushed is pushed to user according to the obtained prediction probability
Breath.
Specifically, the sequence according to obtained prediction probability from large to small pushes the information to be pushed to user, also
It is to say, preferentially by the larger corresponding information to be pushed of prediction probability to user.In the present embodiment, pushed away based on user behavior to user
It send in the larger corresponding information to be pushed of prediction probability, in such manner, it is possible to make the behavior of information to be pushed matching user, for example,
User is in browsing, and to user's PUSH message, for another example, user, to user's PUSH message, allows users in collection
When browsing or collection, PUSH message is received, without receiving PUSH message when payment, forwarding, so that PUSH message
More meet the current behavior of user, avoid impacting the operation of user, so that the use perception of user is more preferably.
In above-described embodiment, the feature of the multidimensional characteristic of user and pushed information is input in neural network model in advance
It surveys, obtains prediction probability, user behavior is then based on, according to prediction probability to user's pushed information, since pushed information is root
It is predicted that probability push, enables the information of push effectively to meet the behavioural habits of user, in addition, pushed information also with
Family behavior is related, avoids to user pushed information when without operation, effectively avoids causing to perplex to the use of user.
In one embodiment, described to learn user behavior importing user model, obtain the multidimensional of user
The step of feature includes: to be directed into the user model for the user behavior as sample behavior;To it is described for model into
Row training, the user model after being trained;The user model that the user behavior imports after training is learnt, is obtained
To the multidimensional characteristic of user.
In the present embodiment, before being learnt to obtain multidimensional characteristic to user behavior, first using user behavior as sample
Notebook data is directed into user model, and user model is trained.It is noted that the user model is by big
The sample data training of amount obtains, which is user behavior, that is to say, that in the present embodiment, is obtained first by newest
The user behavior obtained is input to user model, is trained to user model, after the user model after being trained, to user
Behavior is learnt, and the multidimensional characteristic of the user behavior is obtained, in such manner, it is possible to train user model again, is made
It is more accurate to obtain user model.When obtaining user behavior each time, user model is once trained first, is constantly repaired
Positive user model, and the user behavior of the secondary acquisition is learnt after training each time, corresponding multidimensional characteristic is obtained,
So that the multidimensional characteristic is coming for user model study after training, so that the multidimensional characteristic obtained each time can
It is more accurate.
In one embodiment, described to learn user behavior importing user model, obtain the multidimensional of user
Feature includes: to learn user behavior importing user model, obtains at least one multidimensional characteristic of user;It obtains
Incidence relation between multiple multidimensional characteristics;According between multiple multidimensional characteristics incidence relation and at least one
Multidimensional characteristic obtains multiple multidimensional characteristics of user.
Specifically, multidimensional characteristic refers to the feature of multiple dimensions, in the present embodiment, there is association between multidimensional characteristic,
Namely incidence relation.Incidence relation between multidimensional characteristic can be user model and learn to obtain, and be also possible to user's input
's.In the present embodiment, user behavior learns in user model, it will a multidimensional characteristic is obtained, and according to this multidimensional spy
The incidence relation of sign and other multidimensional characteristics, can obtain the multidimensional characteristic being more associated, in such manner, it is possible to effectively improve
The acquisition efficiency and accuracy of multiple multidimensional characteristics.
According to other multidimensional characteristics that incidence relation obtains, user interest or consumption habit can be widely adapted to.
It should be understood that the acquisition about incidence relation, can be user's manual setting, for example, user setting waits pushing away
Send the relevance of commodity or information, for example articles for babies and baby milk are associated with, and if user buys articles for babies, then buy baby
A possibility that youngster's milk powder, is bigger, is also possible to what user model learnt, for example after push, user has purchase, then judges
User has the demand of this respect;The preferential Securities of commodity for another example, for example may have demand to certain commodity by other features of user, but
User's point has been opened certain commodity and has but never been bought, or never searches for certain commodity, it may be possible to because price is pre- higher than user
Time value, can push commodity coupon, and in the present embodiment, discount coupon is the multidimensional for having incidence relation with user's search commercial articles
Special medical treatment should;It then may be because of the commodity valence that commodity price is bought relative to other approach if user has used discount coupon purchase
Lattice are slightly higher, can continue to push the preferential Securities of commodity, or suitably do commercial activities and price adjustment;Additionally it is possible to by some marketing strategies
It is implanted into information characteristics model, then the marketing strategy is the multidimensional characteristic for having incidence relation with the multidimensional characteristic of user behavior.
In the present embodiment, it is based on artificial intelligence technology, the information of user preference is pushed to user, or user is lost interest in
Information by the implantation marketing methods of user demand test model, pushed information associated with user behavior is sent out
It send to user, participation or the consumption for transferring user are warm and sincere, more fully serve target user, more accurately push to user
Information can be used for electric business platform, information platform, or large-scale forum platform etc..
In one embodiment, described to be based on the user behavior, it is pushed according to the obtained prediction probability to user
The step of information to be pushed includes: to obtain push strategy;Based on push strategy and the user behavior, according to obtaining
The prediction probability push the information to be pushed to user.
In the present embodiment, push strategy for control PUSH message strategy, for example, push time, push duration, push away
The frequency etc. of the mode, push sent.According to push strategy to user's PUSH message, the operation of user can be more effectively adapted to
Habit and consumption habit, avoid causing to perplex to user.
In order to obtain push strategy, one embodiment is, obtains push strategy according to user behavior, and one embodiment is,
The multidimensional characteristic learnt according to user model, acquisition push corresponding with multidimensional characteristic is tactful, in the present embodiment, according to not
Same user behavior, or different multiple push strategies are preset according to different user's multidimensional characteristics, in this way, push plan
Summary can more meet the behavior of user, meet the operating habit and consumption habit of user, in such manner, it is possible to effectively avoid to
It causes to perplex in family.In the present embodiment, multiple push strategies are preset, then according to the mostly feature acquisition pair of user behavior
The push strategy answered, so that push strategy more meets the habit of user.
For example, user behavior occurs after the 23:00 at night, and it is desk lamp that user, which buys commodity, then passes through user
Model learning, the multidimensional characteristic that user can be obtained are night amusement and recreation, need headlamp, then pushing the time then selects at night
22:00 to 24:00 push.It avoids and is pushed in user job on daytime to user, effectively avoid that user is caused to perplex.
In addition, going back multidimensional characteristic further include: user group's feature, information (commodity) can dispense range, validity information, packet
Fill feature, unique identification, other information Relating Characteristic etc..User group's feature may include age of user, gender, workability
Matter, active period etc., in this way, corresponding push strategy can be obtained, and then be more bonded user's according to user group's feature
Browsing habit, consumption habit are to user's PUSH message.
It is noted that multiple multidimensional characteristics respectively correspond different push strategies, and the multidimensional of different users
The corresponding push strategy of feature combination, that is to say, that different multidimensional characteristics will obtain different push strategies, and different
Multidimensional characteristic combination will also obtain different push strategies, in this way, multidimensional characteristic combination is corresponded to push strategy, can make
Strategy, which must be pushed, can more meet the habit of user, avoid causing to perplex to user.
In one embodiment, described to be based on the user behavior, it is pushed according to the obtained prediction probability to user
The step of information to be pushed includes: the frequency of type and user behavior generation based on the user behavior, according to
The obtained prediction probability pushes the information to be pushed to user.
In the present embodiment, the occurrence frequency of type and user behavior based on user behavior can to user's PUSH message
It is more bonded the behavior of user, can effectively avoid that user is caused to perplex.Specifically, the type of user behavior includes sliding screen
Curtain, click screen, click button etc., in this way, the mode of operation of PUSH message is then pushed according to user behavior type, so that
The opening or browsing that obtain the PUSH message meet the mode of operation of user.One embodiment is the type based on user behavior,
It should be to PUSH message to user's push according to obtained prediction probability, wherein should be to the mode of operation and user's row of PUSH message
For type matching.In such manner, it is possible to make the operating habit for more meeting user to the operation of PUSH message, user is effectively improved
Perception.
Specifically, user behavior occurrence frequency is higher, then corresponds to the PUSH message in such a way that the user behavior is corresponding, than
Such as, user behavior is sliding screen, and the frequency for sliding screen is higher, then makes the mode of operation to PUSH message for sliding.
One embodiment is the higher user behavior of the frequency based on generation, pushes institute to user according to the obtained prediction probability
State information to be pushed, wherein corresponding to the mode of operation of PUSH message and the higher user behavior of the frequency of generation.In this way, energy
Enough behavioural habits for more agreeing with user, so that the operation of PUSH message more meets the behavioural habits of user.
It should be understood that although each step in the flow chart of Fig. 2 is successively shown according to the instruction of arrow, this
A little steps are not that the inevitable sequence according to arrow instruction successively executes.Unless expressly state otherwise herein, these steps
It executes there is no the limitation of stringent sequence, these steps can execute in other order.Moreover, at least part in Fig. 2
Step may include that perhaps these sub-steps of multiple stages or stage are executed in synchronization to multiple sub-steps
It completes, but can execute at different times, the execution sequence in these sub-steps or stage, which is also not necessarily, successively to be carried out,
But it can be executed in turn or alternately at least part of the sub-step or stage of other steps or other steps.
It is noted that the information-pushing method based on artificial intelligence is referred to as information push in each embodiment
Method, the information push-delivery apparatus based on artificial intelligence are referred to as information push-delivery apparatus.
In one embodiment, as shown in figure 3, providing a kind of information push-delivery apparatus based on artificial intelligence, comprising: use
Family behavior obtains module 301, multidimensional characteristic obtains module 303, pushed information obtains module 305, information characteristics obtain module
307, prediction module 309 and info push module 311, in which:
User behavior obtains module, for obtaining user behavior;
Multidimensional characteristic obtains module, for learning user behavior importing user model, obtains the more of user
Dimensional feature;
Pushed information obtains module, for obtaining information to be pushed;
Information characteristics obtain module and obtain the feature of the information to be pushed for parsing the information to be pushed;
Prediction module, for being input to the multidimensional characteristic of the feature of the information to be pushed and user and having constructed
Deep neural network model in predicted, obtain the prediction probability of each information to be pushed;
Info push module pushes institute to user according to the obtained prediction probability for being based on the user behavior
State information to be pushed.
In one embodiment, the multidimensional characteristic acquisition module includes:
Import unit, for being directed into the user model for the user behavior as sample behavior;
Training unit, for being trained to described for model, the user model after being trained;
Multidimensional characteristic acquiring unit, the user model for importing the user behavior after training learn,
Obtain the multidimensional characteristic of user.
In one embodiment, the multidimensional characteristic acquisition module includes:
Fisrt feature acquiring unit obtains user extremely for learning user behavior importing user model
A few multidimensional characteristic;
Incidence relation acquiring unit, for obtaining the incidence relation between multiple multidimensional characteristics;
Second feature acquiring unit, for according between multiple multidimensional characteristics incidence relation and at least one is more
Dimensional feature obtains multiple multidimensional characteristics of user.
In one embodiment, info push module includes:
Tactful acquiring unit is pushed, for obtaining push strategy;
Information push unit, it is general according to the obtained prediction for being based on the push strategy and the user behavior
Rate pushes the information to be pushed to user.
In one embodiment, the info push module is also used to type and the user based on the user behavior
The frequency that behavior occurs pushes the information to be pushed to user according to the obtained prediction probability.
Specific about information push-delivery apparatus limits the restriction that may refer to above for information-pushing method, herein not
It repeats again.Modules in above- mentioned information driving means can be realized fully or partially through software, hardware and combinations thereof.On
Stating each module can be embedded in the form of hardware or independently of in the processor in computer equipment, can also store in a software form
In memory in computer equipment, the corresponding operation of the above modules is executed in order to which processor calls.
In one embodiment, a kind of computer equipment is provided, the computer equipment, that is, computer, the computer equipment
It can be server, internal structure chart can be as shown in Figure 4.The computer equipment includes the processing connected by system bus
Device, memory, network interface and database.Wherein, the processor of the computer equipment is for providing calculating and control ability.It should
The memory of computer equipment includes non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operation
System, computer program and database.The built-in storage is operating system and computer program in non-volatile memory medium
Operation provide environment.The database of the computer equipment disappears for storing user model, deep neural network model and push
Breath.The network interface of the computer equipment is used to communicate with external terminal by network connection.The computer program is processed
To realize a kind of information-pushing method when device executes.
It will be understood by those skilled in the art that structure shown in Fig. 4, only part relevant to application scheme is tied
The block diagram of structure does not constitute the restriction for the computer equipment being applied thereon to application scheme, specific computer equipment
It may include perhaps combining certain components or with different component layouts than more or fewer components as shown in the figure.
In one embodiment, a kind of computer equipment is provided, including memory, processor and storage are on a memory
And the computer program that can be run on a processor, processor perform the steps of when executing computer program
Obtain user behavior;
User behavior importing user model is learnt, the multidimensional characteristic of user is obtained;
Obtain information to be pushed;
The information to be pushed is parsed, the feature of the information to be pushed is obtained;
By the multidimensional characteristic of the feature of the information to be pushed and user, it is input to the deep neural network constructed
It is predicted in model, obtains the prediction probability of each information to be pushed;
Based on the user behavior, the information to be pushed is pushed to user according to the obtained prediction probability.
In one embodiment, it is also performed the steps of when processor executes computer program
The user model is directed into using the user behavior as sample behavior;
It is trained to described for model, the user model after being trained;
The user model that the user behavior imports after training is learnt, the multidimensional characteristic of user is obtained.
In one embodiment, it is also performed the steps of when processor executes computer program
User behavior importing user model is learnt, at least one multidimensional characteristic of user is obtained;
Obtain the incidence relation between multiple multidimensional characteristics;
According to the incidence relation and at least one multidimensional characteristic between multiple multidimensional characteristics, the multiple of user are obtained
The multidimensional characteristic.
In one embodiment, it is also performed the steps of when processor executes computer program
Obtain push strategy;
Based on push strategy and the user behavior, according to the obtained prediction probability to user's push it is described to
Pushed information.
In one embodiment, it is also performed the steps of when processor executes computer program
The frequency that type and the user behavior based on the user behavior occur, according to the obtained prediction probability
The information to be pushed is pushed to user.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
Machine program performs the steps of when being executed by processor
Obtain user behavior;
User behavior importing user model is learnt, the multidimensional characteristic of user is obtained;
Obtain information to be pushed;
The information to be pushed is parsed, the feature of the information to be pushed is obtained;
By the multidimensional characteristic of the feature of the information to be pushed and user, it is input to the deep neural network constructed
It is predicted in model, obtains the prediction probability of each information to be pushed;
Based on the user behavior, the information to be pushed is pushed to user according to the obtained prediction probability.
In one embodiment, it is also performed the steps of when computer program is executed by processor
The user model is directed into using the user behavior as sample behavior;
It is trained to described for model, the user model after being trained;
The user model that the user behavior imports after training is learnt, the multidimensional characteristic of user is obtained.
In one embodiment, it is also performed the steps of when computer program is executed by processor
User behavior importing user model is learnt, at least one multidimensional characteristic of user is obtained;
Obtain the incidence relation between multiple multidimensional characteristics;
According to the incidence relation and at least one multidimensional characteristic between multiple multidimensional characteristics, the multiple of user are obtained
The multidimensional characteristic.
In one embodiment, it is also performed the steps of when computer program is executed by processor
Obtain push strategy;
Based on push strategy and the user behavior, according to the obtained prediction probability to user's push it is described to
Pushed information.
In one embodiment, it is also performed the steps of when computer program is executed by processor
The frequency that type and the user behavior based on the user behavior occur, according to the obtained prediction probability
The information to be pushed is pushed to user.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
To any reference of memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
Each technical characteristic of above embodiments can be combined arbitrarily, for simplicity of description, not to above-described embodiment
In each technical characteristic it is all possible combination be all described, as long as however, the combination of these technical characteristics be not present lance
Shield all should be considered as described in this specification.
The several embodiments of the application above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the concept of this application, various modifications and improvements can be made, these belong to the protection of the application
Range.Therefore, the scope of protection shall be subject to the appended claims for the application patent.
Claims (10)
1. a kind of information-pushing method based on artificial intelligence, which comprises
Obtain user behavior;
User behavior importing user model is learnt, the multidimensional characteristic of user is obtained;
Obtain information to be pushed;
The information to be pushed is parsed, the feature of the information to be pushed is obtained;
By the multidimensional characteristic of the feature of the information to be pushed and user, it is input to the deep neural network model constructed
In predicted, obtain the prediction probability of each information to be pushed;
Based on the user behavior, the information to be pushed is pushed to user according to the obtained prediction probability.
2. the method according to claim 1, wherein described import user model for the user behavior
It practises, the step of obtaining the multidimensional characteristic of user includes:
The user model is directed into using the user behavior as sample behavior;
The user model is trained, the user model after being trained;
The user model that the user behavior imports after training is learnt, the multidimensional characteristic of user is obtained.
3. the method according to claim 1, wherein described import user model for the user behavior
It practises, obtain the multidimensional characteristic of user includes:
User behavior importing user model is learnt, at least one multidimensional characteristic of user is obtained;
Obtain the incidence relation between multiple multidimensional characteristics;
According to the incidence relation and at least one multidimensional characteristic between multiple multidimensional characteristics, the multiple described of user is obtained
Multidimensional characteristic.
4. the method according to claim 1, wherein described be based on the user behavior, according to obtaining
Prediction probability to user push the information to be pushed the step of include:
Obtain push strategy;
It is described wait push to user's push according to the obtained prediction probability based on the push strategy and the user behavior
Information.
5. the method according to claim 1, which is characterized in that it is described to be based on the user behavior, according to
The obtained prediction probability to user push the information to be pushed the step of include:
The frequency that type based on the user behavior and the user behavior occur, according to the obtained prediction probability to
Family pushes the information to be pushed.
6. a kind of information push-delivery apparatus based on artificial intelligence, which is characterized in that described device includes:
User behavior obtains module, for obtaining user behavior;
Multidimensional characteristic obtains module, and for learning user behavior importing user model, the multidimensional for obtaining user is special
Sign;
Pushed information obtains module, for obtaining information to be pushed;
Information characteristics obtain module and obtain the feature of the information to be pushed for parsing the information to be pushed;
Prediction module, for being input to the depth constructed for the multidimensional characteristic of the feature of the information to be pushed and user
It is predicted in degree neural network model, obtains the prediction probability of each information to be pushed;
Info push module, for being based on the user behavior, according to the obtained prediction probability to user's push it is described to
Pushed information.
7. device according to claim 6, which is characterized in that the multidimensional characteristic obtains module and includes:
Import unit, for being directed into the user model for the user behavior as sample behavior;
Training unit, for being trained to described for model, the user model after being trained;
Multidimensional characteristic acquiring unit, the user model for importing the user behavior after training learn, obtain
The multidimensional characteristic of user.
8. device according to claim 6, which is characterized in that the multidimensional characteristic obtains module and includes:
Fisrt feature acquiring unit obtains at least the one of user for learning user behavior importing user model
A multidimensional characteristic;
Incidence relation acquiring unit, for obtaining the incidence relation between multiple multidimensional characteristics;
Second feature acquiring unit, for according between multiple multidimensional characteristics incidence relation and at least one multidimensional it is special
Sign, obtains multiple multidimensional characteristics of user.
9. a kind of computer including memory, processor and stores the computer that can be run on a memory and on a processor
Program, which is characterized in that the processor realizes side described in any one of claims 1 to 5 when executing the computer program
The step of method.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any one of claims 1 to 5 is realized when being executed by processor.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910384165.6A CN110335100A (en) | 2019-05-09 | 2019-05-09 | Information-pushing method, device, computer and storage medium based on artificial intelligence |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910384165.6A CN110335100A (en) | 2019-05-09 | 2019-05-09 | Information-pushing method, device, computer and storage medium based on artificial intelligence |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110335100A true CN110335100A (en) | 2019-10-15 |
Family
ID=68139822
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910384165.6A Pending CN110335100A (en) | 2019-05-09 | 2019-05-09 | Information-pushing method, device, computer and storage medium based on artificial intelligence |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110335100A (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110751524A (en) * | 2019-10-22 | 2020-02-04 | 电子科技大学 | Intelligent coupon dispatching method based on reinforcement learning |
CN110867030A (en) * | 2019-11-28 | 2020-03-06 | 中国银行股份有限公司 | Data processing method and device, electronic equipment and computer storage medium |
CN111061956A (en) * | 2019-12-24 | 2020-04-24 | 北京百度网讯科技有限公司 | Method and apparatus for generating information |
CN111464641A (en) * | 2020-03-31 | 2020-07-28 | 深圳前海微众银行股份有限公司 | Message pushing optimization method and device and readable storage medium |
CN111556155A (en) * | 2020-04-29 | 2020-08-18 | 中国银行股份有限公司 | Information pushing method and device |
CN111553736A (en) * | 2020-04-30 | 2020-08-18 | 中国银行股份有限公司 | Method and device for processing preference information, readable medium and equipment |
CN112288447A (en) * | 2020-10-30 | 2021-01-29 | 北京每日优鲜电子商务有限公司 | Article information display method and device, electronic equipment and computer readable medium |
CN112418924A (en) * | 2020-11-17 | 2021-02-26 | 单高峰 | Advertisement pushing method based on big data and cloud computing and artificial intelligence platform |
CN112818169A (en) * | 2021-02-04 | 2021-05-18 | 北京易车互联信息技术有限公司 | Personalized PUSH platform system |
WO2021138823A1 (en) * | 2020-01-08 | 2021-07-15 | 深圳市欢太科技有限公司 | Information pushing method and apparatus, server, and storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102360386A (en) * | 2011-10-12 | 2012-02-22 | 朱一超 | Intelligent shopping guide system and method of electronic commerce website |
CN106649774A (en) * | 2016-12-27 | 2017-05-10 | 北京百度网讯科技有限公司 | Artificial intelligence-based object pushing method and apparatus |
US9864951B1 (en) * | 2015-03-30 | 2018-01-09 | Amazon Technologies, Inc. | Randomized latent feature learning |
-
2019
- 2019-05-09 CN CN201910384165.6A patent/CN110335100A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102360386A (en) * | 2011-10-12 | 2012-02-22 | 朱一超 | Intelligent shopping guide system and method of electronic commerce website |
US9864951B1 (en) * | 2015-03-30 | 2018-01-09 | Amazon Technologies, Inc. | Randomized latent feature learning |
CN106649774A (en) * | 2016-12-27 | 2017-05-10 | 北京百度网讯科技有限公司 | Artificial intelligence-based object pushing method and apparatus |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110751524A (en) * | 2019-10-22 | 2020-02-04 | 电子科技大学 | Intelligent coupon dispatching method based on reinforcement learning |
CN110867030B (en) * | 2019-11-28 | 2021-06-25 | 中国银行股份有限公司 | Data processing method and device, electronic equipment and computer storage medium |
CN110867030A (en) * | 2019-11-28 | 2020-03-06 | 中国银行股份有限公司 | Data processing method and device, electronic equipment and computer storage medium |
CN111061956A (en) * | 2019-12-24 | 2020-04-24 | 北京百度网讯科技有限公司 | Method and apparatus for generating information |
CN111061956B (en) * | 2019-12-24 | 2022-08-16 | 北京百度网讯科技有限公司 | Method and apparatus for generating information |
US11412070B2 (en) | 2019-12-24 | 2022-08-09 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for generating information |
WO2021138823A1 (en) * | 2020-01-08 | 2021-07-15 | 深圳市欢太科技有限公司 | Information pushing method and apparatus, server, and storage medium |
CN111464641A (en) * | 2020-03-31 | 2020-07-28 | 深圳前海微众银行股份有限公司 | Message pushing optimization method and device and readable storage medium |
CN111556155A (en) * | 2020-04-29 | 2020-08-18 | 中国银行股份有限公司 | Information pushing method and device |
CN111553736A (en) * | 2020-04-30 | 2020-08-18 | 中国银行股份有限公司 | Method and device for processing preference information, readable medium and equipment |
CN111553736B (en) * | 2020-04-30 | 2023-06-20 | 中国银行股份有限公司 | Method and device for processing preferential information, readable medium and equipment |
CN112288447A (en) * | 2020-10-30 | 2021-01-29 | 北京每日优鲜电子商务有限公司 | Article information display method and device, electronic equipment and computer readable medium |
CN112418924A (en) * | 2020-11-17 | 2021-02-26 | 单高峰 | Advertisement pushing method based on big data and cloud computing and artificial intelligence platform |
CN112418924B (en) * | 2020-11-17 | 2021-08-20 | 上海东方财富金融数据服务有限公司 | Advertisement pushing method based on big data and cloud computing and artificial intelligence platform |
CN112818169A (en) * | 2021-02-04 | 2021-05-18 | 北京易车互联信息技术有限公司 | Personalized PUSH platform system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110335100A (en) | Information-pushing method, device, computer and storage medium based on artificial intelligence | |
Brynjolfsson et al. | Crowd-Squared | |
CN109783730A (en) | Products Show method, apparatus, computer equipment and storage medium | |
CN109712001A (en) | Information recommendation method, device, computer equipment and storage medium | |
CN109934619A (en) | User's portrait tag modeling method, apparatus, electronic equipment and readable storage medium storing program for executing | |
EP4181026A1 (en) | Recommendation model training method and apparatus, recommendation method and apparatus, and computer-readable medium | |
CN110008397B (en) | Recommendation model training method and device | |
US20160125028A1 (en) | Systems and methods for query rewriting | |
CN105074687A (en) | Re-ranking story content | |
CN110247974A (en) | Information-pushing method, device, computer and storage medium based on block chain | |
US9268859B2 (en) | Method and system for customizing a web site | |
CN111339404A (en) | Content popularity prediction method and device based on artificial intelligence and computer equipment | |
WO2022228303A1 (en) | Video processing method, and storage medium and processor | |
Zhao | Frontiers of big data business analytics: Patterns and cases in online marketing | |
CN109245996A (en) | Mail push method, device, computer equipment and storage medium | |
CN110033331A (en) | A kind of distribution method of discount coupon, system and terminal device | |
US20160275806A1 (en) | Learning apparatus, learning method, and non-transitory computer readable storage medium | |
Wang et al. | Personalized recommendation for new questions in community question answering | |
CN111506821B (en) | Recommendation model, recommendation method, recommendation device, recommendation equipment and recommendation storage medium | |
CN109801101A (en) | Label determines method, apparatus, computer equipment and storage medium | |
CN109417644A (en) | The income optimization launched across screen advertisement | |
Hong et al. | Selective residual learning for visual question answering | |
CN110175892B (en) | Bidirectional thermal ball recommendation method and system based on consumer behaviors | |
CN111522886A (en) | Information recommendation method, terminal and storage medium | |
CN102739698A (en) | Method for using community sharing to assist mobile terminal advertisement publishing and system thereof |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191015 |