CN106686086A - Message pushing method and apparatus - Google Patents
Message pushing method and apparatus Download PDFInfo
- Publication number
- CN106686086A CN106686086A CN201611255475.0A CN201611255475A CN106686086A CN 106686086 A CN106686086 A CN 106686086A CN 201611255475 A CN201611255475 A CN 201611255475A CN 106686086 A CN106686086 A CN 106686086A
- Authority
- CN
- China
- Prior art keywords
- mobile terminal
- user
- push message
- message
- article
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/55—Push-based network services
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/535—Tracking the activity of the user
Abstract
The invention discloses a message pushing method. The message pushing method includes the steps: after receiving the data reported by a mobile terminal, acquiring the user information and article information contained in the data, setting a user label for the user information, and setting an article label for the article information; inputting the user information, the user label, the article information and/or the article label into a preset mathematical model to obtain the user interest data; and according to a custom rule, screening the user interest data, determining a target push message and a corresponding target mobile terminal, and pushing the target push message to the target mobile terminal. The invention also discloses a message pushing apparatus. The message pushing method and apparatus can obtain the user interest data through the mathematical model so as to determine the target push message and push the target push message to the target mobile terminal after classifying the data reported by the mobile terminal and setting labels for the data, thus improving the reliability and efficiency of pushing messages for a server.
Description
Technical field
The present invention relates to server data processing technology field, more particularly to a kind of message pusher and method.
Background technology
At present, PUSH message is usually pushed to all and server and is preserved by server during message push
The mobile terminal of long connection, and because some users are to some message and lose interest in, even if mobile terminal receives this analogizing
Message is sent, user will not also put to drive into go and check, therefore such push mode does not reach expected effect.Or, service
Device targetedly carries out message push according to collecting user behavior data, but the user behavior number that server is collected
According to obtained from only simply being added up to the data of mobile terminal reporting, the user behavior data for obtaining is not
User can be represented interested in some message, reliability is relatively low, cause the less efficient of server push message.
The content of the invention
Present invention is primarily targeted at providing a kind of message pusher and method, it is intended to which improving server carries out message
The reliability and efficiency of push.
For achieving the above object, the invention provides a kind of message pusher, is applied to server end, the message is pushed away
Device is sent to include:
Setting module, for after the data for receiving mobile terminal reporting, obtaining the user's letter included in the data
Breath and Item Information, user tag is set to the user profile, and sets article tag to the Item Information;
Acquisition module, for by the user profile, the user tag, the Item Information and/or the article mark
Sign, be input into preset mathematical model and obtain user interest data;
Pushing module, for being screened to the user interest data according to custom rule, determines that target is pushed and disappears
Breath and corresponding destination mobile terminal, by the target PUSH message destination mobile terminal is pushed to.
Alternatively, the acquisition module includes:
First acquisition unit, for obtaining time index according to the user profile and the Item Information;
Computing unit, for passing through the mathematics according to the time index, the user tag and the article tag
Model calculates user interest data.
Alternatively, when information of the Item Information for reading articles, the first acquisition unit is additionally operable to, according to the use
The total time that all articles of reading that family packet contains are spent reads all article total amounts with corresponding, it is determined that average read speed
Degree;
According to the average reading speed, and reading time, the finger to specifying article that the user profile is included
Determine the content number of words of article, calculate the time index for reading the specified article.
Alternatively, the pushing module includes:
First determining unit, for screening to the user interest data according to custom rule, determines that target is pushed away
Send message and corresponding first object mobile terminal;
Second acquisition unit, presets for obtaining to be more than with the similarity of the user interest data of first object mobile terminal
Second mobile terminal of threshold value;
First push unit, for the target PUSH message to be pushed to into first mobile terminal, when detecting
When stating the first mobile terminal the target PUSH message being browsed, the target PUSH message is pushed to into described second and is moved
Dynamic terminal.
Alternatively, the pushing module also includes:
Second determining unit, for screening to the user interest data according to custom rule, determines the first mesh
Mark PUSH message and corresponding destination mobile terminal;
3rd acquiring unit, for obtain and the first object PUSH message similarity more than the of preset similarity value
Two target PUSH messages;
Second push unit, for first object PUSH message to be pushed to into the destination mobile terminal, when detecting
When stating destination mobile terminal the first object PUSH message being browsed, the second target PUSH message is pushed to into institute
State destination mobile terminal.
Additionally, for achieving the above object, present invention also offers a kind of information push method, is applied to server end, institute
Stating information push method includes:
After the data for receiving mobile terminal reporting, the user profile and Item Information included in the data is obtained,
User tag is set to the user profile, and article tag is set to the Item Information;
By the user profile, the user tag, the Item Information and/or the article tag, it is input into preset
Mathematical model obtains user interest data;
The user interest data are screened according to custom rule, determines target PUSH message and corresponding target
Mobile terminal, by the target PUSH message destination mobile terminal is pushed to.
It is alternatively, described by the user profile, the user tag, the Item Information and/or the article tag,
The preset mathematical model of input obtains user interest data to be included:
Time index is obtained according to the user profile and the Item Information;
User is calculated by the mathematical model according to the time index, the user tag and the article tag emerging
Interesting data.
Alternatively, it is described to be believed according to the user profile and the article when information of the Item Information for reading articles
Breath obtains time index to be included:
It is total with all articles of corresponding reading according to the total time that all articles of reading that the user profile is included are spent
Amount, it is determined that average reading speed;
According to the average reading speed, and reading time, the finger to specifying article that the user profile is included
Determine the content number of words of article, calculate the time index for reading the specified article.
Alternatively, it is described the user interest data are screened according to custom rule, determine target PUSH message
And corresponding destination mobile terminal, the target PUSH message is pushed to into the destination mobile terminal includes:
The user interest data are screened according to custom rule, determines target PUSH message and corresponding first
Destination mobile terminal;
Obtain whole more than the second movement of predetermined threshold value with the similarity of the user interest data of first object mobile terminal
End;
The target PUSH message is pushed to into first mobile terminal, when detecting first mobile terminal to institute
When stating target PUSH message and being browsed, the target PUSH message is pushed to into second mobile terminal.
Alternatively, it is described the user interest data are screened according to custom rule, determine target PUSH message
And corresponding destination mobile terminal, the target PUSH message is pushed to into the destination mobile terminal also includes:
The user interest data are screened according to custom rule, determines first object PUSH message and corresponding
Destination mobile terminal;
Obtain the second target PUSH message for being more than preset similarity value with the similarity of the first object PUSH message;
First object PUSH message is pushed to into the destination mobile terminal, when detecting the destination mobile terminal to institute
When stating first object PUSH message and being browsed, the second target PUSH message is pushed to into the destination mobile terminal.
Embodiment of the present invention server obtains the user's letter included in data after the data for receiving mobile terminal reporting
Breath and Item Information, user tag is set to user profile, and sets article tag to Item Information;By user profile, use
Family label, Item Information and/or article tag, are input into preset mathematical model and obtain user interest data.Then basis is made by oneself
Adopted rule is screened to user interest data, determines target PUSH message and corresponding destination mobile terminal, and target is pushed
Message pushes to destination mobile terminal.So that being classified and being set after label by the data to mobile terminal reporting, pass through
Mathematical model obtains user interest data and pushes to destination mobile terminal determining target PUSH message, and improve server is carried out
Reliability and efficiency that message is pushed.
Description of the drawings
Fig. 1 is the module diagram of message pusher first embodiment of the present invention;
Fig. 2 is the module diagram of message pusher fourth embodiment of the present invention;
Fig. 3 is the module diagram of the embodiment of message pusher the 5th of the present invention;
Fig. 4 is the schematic flow sheet of information push method second embodiment of the present invention;
Fig. 5 is the schematic flow sheet of information push method fourth embodiment of the present invention;
Fig. 6 is the schematic flow sheet of the embodiment of information push method of the present invention 5th.
The realization of the object of the invention, functional characteristics and advantage will be described further referring to the drawings in conjunction with the embodiments.
Specific embodiment
It should be appreciated that specific embodiment described herein is not intended to limit the present invention only to explain the present invention.
As shown in figure 1, showing a kind of message pusher first embodiment of the invention.The message of the embodiment pushes dress
Put including:
Setting module 100, for after the data for receiving mobile terminal reporting, obtaining the user included in the data
Information and Item Information, user tag is set to the user profile, and sets article tag to the Item Information;
In the present embodiment, message pusher is applied to server, and server is receiving the data of mobile terminal reporting
Afterwards, need to carry out pretreatment to data.The type of the mobile terminal can be configured according to actual needs, for example, the movement end
End includes mobile phone, computer etc..
Specifically, during setting module 100 carries out pretreatment to data, data can be classified, is obtained data
In the user profile that includes and Item Information.The user profile include Customs Assigned Number (i.e. ID), work, the age, sex,
Location etc..The user profile also include user behavior information, for example, position, the network for being used, the duration for browsing news,
What news section browsing pages number, each page stay time, total page number, correlation were read browses duration, clicks on number, comment
Duration, comment thumb up number, thumb up are browsed, shared, commented on (comment includes evaluation content, do not like), user profile includes
Also include Negative Feedback information etc..
Then setting module 100 is that user profile sets user tag, and the user tag can pass through what is reported at the beginning
The aspect feedback such as information news that data, the data that third party's social networkies are crawled and follow-up user read is generated.Example
Such as, extract the specific object of user, such as to age segmentations, be divided into 0~20,20~50,>50 grade age brackets, these are belonged to
Property digitized.I.e. user tag includes Customs Assigned Number, the characteristic information, the weight that pass through etc..User tag also includes browsing to the
Several pages, the time started, the end time, the persistent period, whether thumb up, the reasons why do not like, whether share.
The Item Information includes news numbering, headline, the news time, hot spot data, number of clicks, comment are individual
Time, number of visits, download time that number, article length, article title, focus article, the source of article, article are issued etc..
And setting module 100 is that Item Information sets article tag, the article tag can pass through the label for reporting of Domestic News
Information, and server crawls the label for acquiring to Domestic News.For example, article tag includes news numbering, news
Feature etc., news features refer to the classification belonging to news;Or article tag is including application numbers, using feature etc., using spy
Levy is using affiliated classification.
The label of user tag that is, user profile labeling, such as certain user be female, 31 years old, it is married, income 10,000 with
Upper, food of liking to be beautiful, purchase by group intelligent etc..Article tag that is, Item Information labeling.Label for user profile and Item Information
One of free-revving engine is that for convenience server is processed data, for example, can do classified statistic:Like the user of red wine
How many, like M-F in the crowd of red wine to be how many;Data mining work can also be done:Calculated using correlation rule and liked
The people of joyous red wine is generally preferred that any motion brand, people's age bracket distribution situation of red wine is liked using focusing solutions analysis, etc.
Deng.User tag and article tag can lift precision, improve the efficiency of acquisition of information.
Acquisition module 200, for by the user profile, the user tag, the Item Information and/or the article
Label, is input into preset mathematical model and obtains user interest data;
In the present embodiment, server is previously provided with mathematical model, and the mathematical model is used to carry out user interest data
Calculate, the user interest data can be interest-degree of the user to article, and following examples will be described in more detail.Used
After family information, user tag, Item Information and article tag, acquisition module 200 can be by user profile, user tag, article letter
Breath and/or article tag are input into preset mathematical model, and by the mathematical model user interest data are exported.
Pushing module 300, for screening to the user interest data according to custom rule, determines that target is pushed
Message and corresponding destination mobile terminal, by the target PUSH message destination mobile terminal is pushed to.
After user interest data are obtained, user interest data can be sieved according to the custom rule of each business
Choosing, can be ranked up screening or be Integrated Selection etc., to be met the user interest data of certain condition.For example,
Filter out in user interest data 3 before interest-degree highest.
Then pushing module 300 determines that target PUSH message and corresponding target are moved according to the user interest data after screening
Dynamic terminal, by target PUSH message destination mobile terminal is pushed to.The push mode can be it is offline recommend, it is online recommend, it is real
When recommend etc., subsequent server can also be evaluated recommendation results.Server can also be by PUSH message with mail, bullet
The forms such as window, note push to destination mobile terminal, destination mobile terminal after the message for receiving server push, Yong Huke
To select to be checked, downloaded or deleted.It is achieved thereby that the data of effectively utilizes mobile terminal reporting, carry out user characteristicses
Excavate, be to provide recommendation service using center, browser, visual field etc..
It should be noted that in order to improve the reliability that server is pushed to message, server can be with reception
To after the data of multiple mobile terminal reportings, the data of multiple mobile terminal reportings are randomly divided into into A groups and B groups, are then distinguished
Pretreatment is carried out to the data in A groups and B groups.User profile, user tag, Item Information and/or the article mark that A groups are obtained
Sign the first mathematical model of input and obtain user interest data, PUSH message was pushed to into for the first movement end according to user interest data
End, and recommendation results are evaluated, the first evaluating data is obtained, the evaluating data is interest-degree of the user to message.With
And user profile, the user tag, Item Information and/or the article tag obtained to B groups is input into the second mathematical model and obtains
User interest data, the second mobile terminal is pushed to according to user interest data by PUSH message, and recommendation results are evaluated,
Obtain the second evaluating data.When the evaluation of estimate of the first evaluating data is more than the evaluation of estimate of the second evaluating data, by the first mathematics
Model is set to target mathematical model, otherwise the second mathematical model is set to into target mathematical model, will pass through number of targets
It is that PUSH message is pushed that model obtains user interest data.
Embodiment of the present invention server obtains the user's letter included in data after the data for receiving mobile terminal reporting
Breath and Item Information, user tag is set to user profile, and sets article tag to Item Information;By user profile, use
Family label, Item Information and/or article tag, are input into preset mathematical model and obtain user interest data.Then basis is made by oneself
Adopted rule is screened to user interest data, determines target PUSH message and corresponding destination mobile terminal, and target is pushed
Message pushes to destination mobile terminal.So that being classified and being set after label by the data to mobile terminal reporting, pass through
Mathematical model obtains user interest data and pushes to destination mobile terminal determining target PUSH message, and improve server is carried out
Reliability and efficiency that message is pushed.
Further, based on above-mentioned message pusher first embodiment, it is proposed that message pusher second of the present invention
Embodiment, the present embodiment is with the difference of message pusher first embodiment of the present invention, above-mentioned acquisition module in the present embodiment
200 include:
First acquisition unit, for obtaining time index according to the user profile and the Item Information;
Computing unit, for passing through the mathematics according to the time index, the user tag and the article tag
Model calculates user interest data.
In the present embodiment, server during being calculated user interest data according to mathematical model, first by
First acquisition unit obtains time index according to user profile and the Item Information, and following examples will be described in more detail.
Then user interest number is calculated by the mathematical model according to time index, user tag and article tag by computing unit
According to user interest data are user interest degree.
To below be illustrated, and when Item Information is news, data can be quantified by weight calculate
Interest-degree of the user to article, the mathematical model of calculating is as follows:
UC represents user interest data, and 1 represents skew, and NCi represents the classification (i.e. article tag) of i-th Domestic News, n
The news quantity that user sees is represented, the weight coefficient of α express time indexs, β represents the weight coefficient of comment index, and γ is thumb up
The weight coefficient shared, α, β, gamma coefficient can be arranged as the case may be and flexibly.Time express time indexs, comment tables
Show comment index, share represents that thumb up is shared, unlike is represented and do not liked index (i.e. user tag).
The present embodiment server obtains time index according to user profile and Item Information, then according to time index, use
Family label and article tag calculate user interest data by mathematical model, improve the reliability to user interest data acquisition
Property.
Further, based on above-mentioned message pusher second embodiment, it is proposed that message pusher the 3rd of the present invention
Embodiment, the present embodiment is with the difference of message pusher second embodiment of the present invention, above-mentioned when article letter in the present embodiment
When ceasing the information for reading articles, first acquisition unit is additionally operable to, according to all article institutes of reading that the user profile is included
The total time of cost reads all article total amounts with corresponding, it is determined that average reading speed;
According to the average reading speed, and reading time, the finger to specifying article that the user profile is included
Determine the content number of words of article, calculate the time index for reading the specified article.
In the present embodiment, when information of the Item Information for reading articles, first acquisition unit according to user profile and
During Item Information obtains time index, in order to respectively follow-up offline layer is specific to the calculating of feature, can carry out
The statistics of data, such as the quantitative statisticses of browsing time.Spent according to all articles of reading that user profile is included first
Total time t (i) and it is corresponding read all articles total amount size (i), it is determined that average reading speed Ri, this article is with news
Example, computing formula is as follows:
Wherein, RiRepresent the average reading speed of user i, t (i) represent in all news for reading of user the used time between
(5%-95%) the time sum that document is spent, size (i) is that user i is all reads the news documents for meeting t (i) conditions
Size sum (i.e. news total amount).
Then according to average reading speed, and user profile include the reading time to specifying article, specify article
Content number of words, calculates and reads the time index for specifying article, and this article is by taking news as an example, and computing formula is as follows:
Wherein, ti,jRepresent reading times of the user i to news j, RiRepresent the average reading speed of user i, size (j)
Represent the size (i.e. the content number of words of news j) of news documents, timei,jRepresent that user browses the time index of news.
, when reading articles, server can be according to reading total time and the corresponding reading that all articles be spent for the present embodiment
All article total amounts according to average reading speed, the reading time to specifying article and refer to determining average reading speed, then
Determine the content number of words of article, calculate and read the time index for specifying article.Improve what server was determined to time index
Motility and accuracy.
Further, as shown in Fig. 2 being based on the above-mentioned embodiment of message pusher first, second or third, it is proposed that this
Invention message pusher fourth embodiment, the present embodiment and the embodiment of message pusher first, second or third of the present invention
Difference be that above-mentioned pushing module 300 includes in the present embodiment:
First determining unit 310, for screening to the user interest data according to custom rule, determines target
PUSH message and corresponding first object mobile terminal;
Second acquisition unit 320, is more than for obtaining with the similarity of the user interest data of first object mobile terminal
Second mobile terminal of predetermined threshold value;
First push unit 330, for the target PUSH message to be pushed to into first mobile terminal, when detecting
When first mobile terminal is browsed to the target PUSH message, the target PUSH message is pushed to into described second
Mobile terminal.
In the present embodiment, server can be according to the similarity of user interest data come PUSH message, specifically, first by
One determining unit 310 is screened according to custom rule to user interest data, determines target PUSH message and corresponding
One destination mobile terminal, and the phase with the user interest data of first object mobile terminal is obtained by second acquisition unit 320
It is more than the second mobile terminal of predetermined threshold value like degree.The similarity of the user interest data is user's similarity, and user is similar
Degree can be obtained by Pearson's correlation coefficient (Pearson correlation coefficient).Pearson's correlation coefficient
For measuring two mutual relation (linear correlation) between variable X and Y, span between [- 1 ,+1], Pearson came product
Square correlation coefficient is used for measuring the power of two linear variable displacement dependencys.The predetermined threshold value can set as the case may be and flexibly
Put.
Then target PUSH message is pushed to the first mobile terminal by the first push unit 330, and detects for the first movement end
Whether end browses to target PUSH message.When detecting the first mobile terminal target PUSH message being browsed, say
Corresponding user is interested in the target PUSH message for bright first mobile terminal, and due to the corresponding user of the first mobile terminal with
The similarity of the user interest data between the corresponding user of the second mobile terminal is more than predetermined threshold value, and illustrating may the second movement
The corresponding user of terminal also can be interested in the target PUSH message, and now the first push unit 330 pushes away target PUSH message
Deliver to the second mobile terminal.
For example, in the corresponding interesting data of user 1 of the first mobile terminal, user 1 likes article A, article B, article C, the
In the corresponding interesting data of user 2 of two mobile terminals, user 2 also likes article A, article B, article C, and now explanation first is moved
Similarity between the corresponding user 1 of terminal user 2 corresponding with the second mobile terminal is more than predetermined threshold value.When server will be pushed away
Message a is sent to push to after the first mobile terminal, user 1 has checked PUSH message a, now server pushes PUSH message a
To the second mobile terminal.
The present embodiment server determines target PUSH message and corresponding first object mobile terminal, and moves with first object
After the similarity of the user interest data of dynamic terminal is more than the second mobile terminal of predetermined threshold value, target PUSH message is pushed to
First mobile terminal, and when the first mobile terminal is browsed to the target PUSH message, target PUSH message is pushed to
Second mobile terminal.So that server can improve server and offset according to the similarity of user interest data come PUSH message
The motility that breath is pushed.
Further, as shown in figure 3, being based on the above-mentioned embodiment of message pusher the first, second or thirdth, it is proposed that
The embodiment of message pusher the 5th of the present invention, the present embodiment is implemented with message pusher of the present invention first, second or third
The difference of example is that above-mentioned pushing module 300 also includes in the present embodiment:
Second determining unit 340, for screening to the user interest data according to custom rule, determines first
Target PUSH message and corresponding destination mobile terminal;
3rd acquiring unit 350, preset similarity value is more than for obtaining with the similarity of the first object PUSH message
The second target PUSH message;
Second push unit 360, for first object PUSH message to be pushed to into the destination mobile terminal, when detecting
When the destination mobile terminal is browsed to the first object PUSH message, the second target PUSH message is pushed to
The destination mobile terminal.
In the present embodiment, server can be according to the similarity of target PUSH message come PUSH message, specifically, first by
Two determining units 340 are screened according to custom rule to the user interest data, determine first object PUSH message and
Corresponding destination mobile terminal, and obtained with the similarity of first object PUSH message more than pre- by the 3rd acquiring unit 350
If the second target PUSH message of similar value.The similarity of the target PUSH message is article similarity, the article similarity
Obtained by Pearson's correlation coefficient (Pearson correlation coefficient).The preset similarity value can basis
Concrete condition and flexibly arrange.
Then first object PUSH message is pushed to destination mobile terminal by the second push unit 360, and detects target shifting
Whether dynamic terminal browses to first object PUSH message.First object PUSH message is entered when destination mobile terminal is detected
When row is browsed, illustrate that the corresponding user of destination mobile terminal is interested in the first object PUSH message, and due to first object
Similarity between PUSH message and the second target PUSH message is more than preset similarity value, illustrates the corresponding use of destination mobile terminal
Family may also can be interested be in the second target PUSH message, now the second push unit 360 pushes the second target PUSH message
To destination mobile terminal.
For example, the similarity between PUSH message a and PUSH message b is more than preset similarity value, when server disappears push
Breath a is pushed to after destination mobile terminal, and the corresponding user of destination mobile terminal has checked PUSH message a, and now server will
PUSH message b also pushes to the second mobile terminal.
The present embodiment server determines first object PUSH message and corresponding destination mobile terminal, and pushes away with first object
The similarity for sending message is more than after the second target PUSH message of preset similarity value, and first object PUSH message is pushed to into target
Mobile terminal, and when destination mobile terminal is browsed to first object PUSH message, the second target PUSH message is pushed
To destination mobile terminal so that server can improve server pair according to the similarity of target PUSH message come PUSH message
Motility and efficiency that message is pushed.
Accordingly, as shown in figure 4, proposing a kind of information push method first embodiment of the invention.The present embodiment message is pushed away
Delivery method is corresponding with above-mentioned message pusher first embodiment, and the information push method of the embodiment includes:
Step S10, after the data for receiving mobile terminal reporting, obtain the user profile and thing included in the data
Product information, user tag is set to the user profile, and sets article tag to the Item Information;
In the present embodiment, information push method is applied to server, and server is receiving the data of mobile terminal reporting
Afterwards, need to carry out pretreatment to data.The type of the mobile terminal can be configured according to actual needs, for example, the movement end
End includes mobile phone, computer etc..
Specifically, during server carries out pretreatment to data, data can be classified, obtains and included in data
User profile and Item Information.The user profile includes Customs Assigned Number (i.e. ID), work, age, sex, address etc..
The user profile also includes user behavior information, for example, position, the network for being used, the duration for browsing news, Department of Public Information
Point browsing pages number, each page stay time, total page number, it is related read when browsing duration, clicking on number, comment and browse
Grow, comment on thumb up number, thumb up, share, comment on (comment includes evaluation content, do not like), user profile includes also including
Negative Feedback information etc..
Then user tag is set for user profile, the user tag can pass through data, the third party for reporting at the beginning
The aspect feedbacks such as the information news that the data and follow-up user that social networkies are crawled are read are generated.For example, extract user's
Specific object, such as to age segmentations, be divided into 0~20,20~50,>50 grade age brackets, by these attribute digitizeds.Use
Family label includes Customs Assigned Number, the characteristic information, the weight that pass through etc..User tag also includes browsing to which page, starts
Time, the end time, the persistent period, whether thumb up, the reasons why do not like, whether share.
The Item Information includes news numbering, headline, the news time, hot spot data, number of clicks, comment are individual
Time, number of visits, download time that number, article length, article title, focus article, the source of article, article are issued etc..
And article tag is set for Item Information, the article tag can pass through the label information for reporting of Domestic News, and clothes
Business device crawls the label for acquiring to Domestic News.For example, article tag includes news numbering, news features etc., news
Feature refers to the classification belonging to news;Or article tag, including application numbers, using feature etc., application is characterized in that using affiliated
Classification.
The label of user tag that is, user profile labeling, such as certain user be female, 31 years old, it is married, income 10,000 with
Upper, food of liking to be beautiful, purchase by group intelligent etc..Article tag that is, Item Information labeling.Label for user profile and Item Information
One of free-revving engine is that for convenience server is processed data, for example, can do classified statistic:Like the user of red wine
How many, like M-F in the crowd of red wine to be how many;Data mining work can also be done:Calculated using correlation rule and liked
The people of joyous red wine is generally preferred that any motion brand, people's age bracket distribution situation of red wine is liked using focusing solutions analysis, etc.
Deng.User tag and article tag can lift precision, improve the efficiency of acquisition of information.
Step S20, by the user profile, the user tag, the Item Information and/or the article tag, it is defeated
Enter preset mathematical model and obtain user interest data;
In the present embodiment, server is previously provided with mathematical model, and the mathematical model is used to carry out user interest data
Calculate, the user interest data can be interest-degree of the user to article, and following examples will be described in more detail.Used
After family information, user tag, Item Information and article tag, can be by user profile, user tag, Item Information and/or article
Label is input into preset mathematical model, and by the mathematical model user interest data are exported.
Step S30, the user interest data are screened according to custom rule, determine target PUSH message and right
The destination mobile terminal answered, by the target PUSH message destination mobile terminal is pushed to.
After user interest data are obtained, user interest data can be sieved according to the custom rule of each business
Choosing, can be ranked up screening or be Integrated Selection etc., to be met the user interest data of certain condition.For example,
Filter out in user interest data 3 before interest-degree highest.
Then target PUSH message and corresponding destination mobile terminal are determined according to the user interest data after screening, by mesh
Mark PUSH message pushes to destination mobile terminal.The push mode can be it is offline recommend, online recommendation, real-time recommendation etc., after
Continuous server can also be evaluated recommendation results.Server can also be by PUSH message in forms such as mail, pop-up, notes
Destination mobile terminal is pushed to, after the message for receiving server push, user can select to be looked into destination mobile terminal
See, download or delete.It is achieved thereby that the data of effectively utilizes mobile terminal reporting, carry out user characteristicses excavation, it is application
Center, browser, visual field etc. provide recommendation service.
It should be noted that in order to improve the reliability that server is pushed to message, server can be with reception
To after the data of multiple mobile terminal reportings, the data of multiple mobile terminal reportings are randomly divided into into A groups and B groups, are then distinguished
Pretreatment is carried out to the data in A groups and B groups.User profile, user tag, Item Information and/or the article mark that A groups are obtained
Sign the first mathematical model of input and obtain user interest data, PUSH message was pushed to into for the first movement end according to user interest data
End, and recommendation results are evaluated, the first evaluating data is obtained, the evaluating data is interest-degree of the user to message.With
And user profile, the user tag, Item Information and/or the article tag obtained to B groups is input into the second mathematical model and obtains
User interest data, the second mobile terminal is pushed to according to user interest data by PUSH message, and recommendation results are evaluated,
Obtain the second evaluating data.When the evaluation of estimate of the first evaluating data is more than the evaluation of estimate of the second evaluating data, by the first mathematics
Model is set to target mathematical model, otherwise the second mathematical model is set to into target mathematical model, will pass through number of targets
It is that PUSH message is pushed that model obtains user interest data.
Embodiment of the present invention server obtains the user's letter included in data after the data for receiving mobile terminal reporting
Breath and Item Information, user tag is set to user profile, and sets article tag to Item Information;By user profile, use
Family label, Item Information and/or article tag, are input into preset mathematical model and obtain user interest data.Then basis is made by oneself
Adopted rule is screened to user interest data, determines target PUSH message and corresponding destination mobile terminal, and target is pushed
Message pushes to destination mobile terminal.So that being classified and being set after label by the data to mobile terminal reporting, pass through
Mathematical model obtains user interest data and pushes to destination mobile terminal determining target PUSH message, and improve server is carried out
Reliability and efficiency that message is pushed.
Further, based on above-mentioned information push method first embodiment, it is proposed that information push method of the present invention second
Embodiment, the present embodiment information push method is corresponding with above-mentioned message pusher second embodiment.The present embodiment and the present invention
The difference of information push method first embodiment is that above-mentioned steps S20 may include in the present embodiment:
Time index is obtained according to the user profile and the Item Information;
User is calculated by the mathematical model according to the time index, the user tag and the article tag emerging
Interesting data.
In the present embodiment, server during being calculated user interest data according to mathematical model, root first
Time index is obtained according to user profile and the Item Information, following examples will be described in more detail.Then referred to according to the time
Mark, user tag and article tag calculate user interest data by the mathematical model, and it is emerging that user interest data are user
Interesting degree.
To below be illustrated, and when Item Information is news, data can be quantified by weight calculate
Interest-degree of the user to article, the mathematical model of calculating is as follows:
UC represents user interest data, and 1 represents skew, and NCi represents the classification (i.e. article tag) of i-th Domestic News, n
The news quantity that user sees is represented, the weight coefficient of α express time indexs, β represents the weight coefficient of comment index, and γ is thumb up
The weight coefficient shared, α, β, gamma coefficient can be arranged as the case may be and flexibly.Time express time indexs, comment tables
Show comment index, share represents that thumb up is shared, unlike is represented and do not liked index (i.e. user tag).
The present embodiment server obtains time index according to user profile and Item Information, then according to time index, use
Family label and article tag calculate user interest data by mathematical model, improve the reliability to user interest data acquisition
Property.
Further, based on above-mentioned information push method second embodiment, it is proposed that information push method of the present invention 3rd
Embodiment, the present embodiment information push method is corresponding with above-mentioned message pusher 3rd embodiment.The present embodiment and the present invention
The difference of information push method second embodiment is, above-mentioned when information of the Item Information for reading articles in the present embodiment, institute
State includes according to the user profile and Item Information acquisition time index:
It is total with all articles of corresponding reading according to the total time that all articles of reading that the user profile is included are spent
Amount, it is determined that average reading speed;
According to the average reading speed, and reading time, the finger to specifying article that the user profile is included
Determine the content number of words of article, calculate the time index for reading the specified article.
In the present embodiment, when information of the Item Information for reading articles, server is being believed according to user profile and article
During breath obtains time index, in order to respectively follow-up offline layer is specific to the calculating of feature, some data can be carried out
Statistics, such as the quantitative statisticses of browsing time.First according to all articles of reading that user profile is included spent it is total when
Between t (i) and it is corresponding read all articles total amount size (i), it is determined that average reading speed Ri, this article is calculated by taking news as an example
Formula is as follows:
Wherein, RiRepresent the average reading speed of user i, t (i) represent in all news for reading of user the used time between
(5%-95%) the time sum that document is spent, size (i) is that user i is all reads the news documents for meeting t (i) conditions
Size sum (i.e. news total amount).
Then according to average reading speed, and user profile include the reading time to specifying article, specify article
Content number of words, calculates and reads the time index for specifying article, and this article is by taking news as an example, and computing formula is as follows:
Wherein, ti,jRepresent reading times of the user i to news j, RiRepresent the average reading speed of user i, size (j)
Represent the size (i.e. the content number of words of news j) of news documents, timei,jRepresent that user browses the time index of news.
, when reading articles, server can be according to reading total time and the corresponding reading that all articles be spent for the present embodiment
All article total amounts according to average reading speed, the reading time to specifying article and refer to determining average reading speed, then
Determine the content number of words of article, calculate and read the time index for specifying article.Improve what server was determined to time index
Motility and accuracy.
Further, as shown in figure 5, being based on the embodiment of above-mentioned information push method first, second or third, it is proposed that this
Invention information push method fourth embodiment, the present embodiment information push method and above-mentioned message pusher fourth embodiment pair
Should.The present embodiment is with the difference of the embodiment of information push method of the present invention first, second or third, above-mentioned step in the present embodiment
Rapid S30 may include:
Step S31, the user interest data are screened according to custom rule, determine target PUSH message and right
The first object mobile terminal answered;
Step S32, obtain with the similarity of the user interest data of first object mobile terminal more than the of predetermined threshold value
Two mobile terminals;
Step S33, the target PUSH message is pushed to into first mobile terminal, when detecting first movement
When target PUSH message is browsed described in terminal-pair, the target PUSH message is pushed to into second mobile terminal.
In the present embodiment, server specifically, can first be serviced according to the similarity of user interest data come PUSH message
Device is screened according to custom rule to user interest data, determines target PUSH message and the movement of corresponding first object eventually
End, and obtain the second movement for being more than predetermined threshold value with the similarity of the user interest data of first object mobile terminal eventually
End.The similarity of the user interest data is user's similarity, and user's similarity can pass through Pearson's correlation coefficient
(Pearson correlation coefficient) is obtaining.Pearson's correlation coefficient be used for measuring two variable Xs and Y it
Between mutual relation (linear correlation), between [- 1 ,+1], Pearson product-moment correlation coefficient is used for measuring two span
The power of linear variable displacement dependency.The predetermined threshold value can be arranged as the case may be and flexibly.
Then target PUSH message is pushed to the first mobile terminal by server, and whether detects the first mobile terminal to mesh
Mark PUSH message is browsed.When detecting the first mobile terminal target PUSH message being browsed, the first movement is illustrated
The corresponding user of terminal is interested in the target PUSH message, and due to the corresponding user of the first mobile terminal with the second movement eventually
The similarity for holding the user interest data between corresponding user is more than predetermined threshold value, illustrates that possible second mobile terminal is corresponding
User also can be interested in the target PUSH message, now target PUSH message is pushed to into the second mobile terminal.
For example, in the corresponding interesting data of user 1 of the first mobile terminal, user 1 likes article A, article B, article C, the
In the corresponding interesting data of user 2 of two mobile terminals, user 2 also likes article A, article B, article C, and now explanation first is moved
Similarity between the corresponding user 1 of terminal user 2 corresponding with the second mobile terminal is more than predetermined threshold value.When server will be pushed away
Message a is sent to push to after the first mobile terminal, user 1 has checked PUSH message a, now server pushes PUSH message a
To the second mobile terminal.
The present embodiment server determines target PUSH message and corresponding first object mobile terminal, and moves with first object
After the similarity of the user interest data of dynamic terminal is more than the second mobile terminal of predetermined threshold value, target PUSH message is pushed to
First mobile terminal, and when the first mobile terminal is browsed to the target PUSH message, target PUSH message is pushed to
Second mobile terminal.So that server can improve server and offset according to the similarity of user interest data come PUSH message
The motility that breath is pushed.
Further, as shown in fig. 6, being based on the embodiment of above-mentioned information push method first, second or third, it is proposed that this
The embodiment of invention information push method the 5th, the present embodiment information push method and the above-mentioned embodiment pair of message pusher the 5th
Should.The present embodiment is with the difference of the embodiment of information push method of the present invention first, second or third, above-mentioned step in the present embodiment
Rapid S30 may include:
Step S34, the user interest data are screened according to custom rule, determine first object PUSH message
And corresponding destination mobile terminal;
Step S35, acquisition and the similarity of the first object PUSH message are pushed away more than the second target of preset similarity value
Send message;
Step S36, first object PUSH message is pushed to into the destination mobile terminal, when detecting the target movement
When first object PUSH message is browsed described in terminal-pair, the second target PUSH message is pushed to into the target movement
Terminal.
In the present embodiment, server specifically, can first be serviced according to the similarity of target PUSH message come PUSH message
Device is screened according to custom rule to the user interest data, determines that first object PUSH message and corresponding target are moved
Dynamic terminal, and obtain the second target PUSH message for being more than preset similarity value with the similarity of first object PUSH message.Should
The similarity of target PUSH message is article similarity, and the article similarity passes through Pearson's correlation coefficient (Pearson
Correlation coefficient) obtaining.The preset similarity value can be arranged as the case may be and flexibly.
Then first object PUSH message is pushed to destination mobile terminal by server, and whether detects destination mobile terminal
First object PUSH message is browsed.When detecting destination mobile terminal first object PUSH message being browsed,
Illustrate that the corresponding user of destination mobile terminal is interested in the first object PUSH message, and due to first object PUSH message with
Similarity between second target PUSH message is more than preset similarity value, illustrates that the corresponding user of destination mobile terminal may be to the
Two target PUSH messages also can be interested, now the second target PUSH message is pushed to into destination mobile terminal.
For example, the similarity between PUSH message a and PUSH message b is more than preset similarity value, when server disappears push
Breath a is pushed to after destination mobile terminal, and the corresponding user of destination mobile terminal has checked PUSH message a, and now server will
PUSH message b also pushes to the second mobile terminal.
The present embodiment server determines first object PUSH message and corresponding destination mobile terminal, and pushes away with first object
The similarity for sending message is more than after the second target PUSH message of preset similarity value, and first object PUSH message is pushed to into target
Mobile terminal, and when destination mobile terminal is browsed to first object PUSH message, the second target PUSH message is pushed
To destination mobile terminal so that server can improve server pair according to the similarity of target PUSH message come PUSH message
Motility and efficiency that message is pushed.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment side
Method can add the mode of required general hardware platform to realize by software, naturally it is also possible to by hardware, but in many cases
The former is more preferably embodiment.Based on such understanding, technical scheme is substantially done to prior art in other words
Going out the part of contribution can be embodied in the form of software product, and the computer software product is stored in a storage medium
In (such as ROM/RAM, magnetic disc, CD), including some instructions are used so that a station terminal equipment (can be mobile phone, computer takes
Business device, air-conditioner, or network equipment etc.) perform method described in each embodiment of the invention.
The alternative embodiment of the present invention is these are only, the scope of the claims of the present invention is not thereby limited, it is every using this
Equivalent structure or equivalent flow conversion that bright description and accompanying drawing content are made, or directly or indirectly it is used in other related skills
Art field, is included within the scope of the present invention.
Claims (10)
1. a kind of message pusher, is applied to server end, it is characterised in that the message pusher includes:
Setting module, for after the data for receiving mobile terminal reporting, obtain the user profile that includes in the data and
Item Information, user tag is set to the user profile, and sets article tag to the Item Information;
Acquisition module, it is defeated for by the user profile, the user tag, the Item Information and/or the article tag
Enter preset mathematical model and obtain user interest data;
Pushing module, for being screened to the user interest data according to custom rule, determine target PUSH message and
Corresponding destination mobile terminal, by the target PUSH message destination mobile terminal is pushed to.
2. message pusher as claimed in claim 1, it is characterised in that the acquisition module includes:
First acquisition unit, for obtaining time index according to the user profile and the Item Information;
Computing unit, for passing through the mathematical model according to the time index, the user tag and the article tag
Calculate user interest data.
3. message pusher as claimed in claim 2, it is characterised in that when information of the Item Information for reading articles,
The first acquisition unit is additionally operable to, the total time spent according to all articles of reading that the user profile is included with it is corresponding
All article total amounts are read, it is determined that average reading speed;
According to the average reading speed, and reading time, the specified text to specifying article that the user profile is included
The content number of words of chapter, calculates the time index for reading the specified article.
4. the message pusher as described in any one of claim 1-3, it is characterised in that the pushing module includes:
First determining unit, for being screened to the user interest data according to custom rule, determines that target is pushed and disappears
Breath and corresponding first object mobile terminal;
Second acquisition unit, predetermined threshold value is more than for obtaining with the similarity of the user interest data of first object mobile terminal
The second mobile terminal;
First push unit, for the target PUSH message to be pushed to into first mobile terminal, when detecting described
When one mobile terminal is browsed to the target PUSH message, the target PUSH message is pushed to into second movement eventually
End.
5. the message pusher as described in any one of claim 1-3, it is characterised in that the pushing module also includes:
Second determining unit, for screening to the user interest data according to custom rule, determines that first object is pushed away
Send message and corresponding destination mobile terminal;
3rd acquiring unit, for obtaining the second mesh for being more than preset similarity value with the similarity of the first object PUSH message
Mark PUSH message;
Second push unit, for first object PUSH message to be pushed to into the destination mobile terminal, when detecting the mesh
When mark mobile terminal is browsed to the first object PUSH message, the second target PUSH message is pushed to into the mesh
Mark mobile terminal.
6. a kind of information push method, is applied to server end, it is characterised in that the information push method includes following step
Suddenly:
After the data for receiving mobile terminal reporting, the user profile and Item Information included in the data is obtained, to institute
User profile setting user tag is stated, and article tag is set to the Item Information;
By the user profile, the user tag, the Item Information and/or the article tag, preset mathematics is input into
Model obtains user interest data;
The user interest data are screened according to custom rule, determines target PUSH message and the movement of corresponding target
Terminal, by the target PUSH message destination mobile terminal is pushed to.
7. information push method as claimed in claim 6, it is characterised in that described to mark the user profile, the user
Label, the Item Information and/or the article tag, being input into preset mathematical model acquisition user interest data includes:
Time index is obtained according to the user profile and the Item Information;
User interest number is calculated by the mathematical model according to the time index, the user tag and the article tag
According to.
8. information push method as claimed in claim 7, it is characterised in that when information of the Item Information for reading articles,
It is described to be included according to the user profile and Item Information acquisition time index:
All article total amounts are read with corresponding according to the total time that all articles of reading that the user profile is included are spent, really
Allocate equal reading speed;
According to the average reading speed, and reading time, the specified text to specifying article that the user profile is included
The content number of words of chapter, calculates the time index for reading the specified article.
9. the information push method as described in any one of claim 6-8, it is characterised in that it is described according to custom rule to institute
State user interest data to be screened, determine target PUSH message and corresponding destination mobile terminal, the target is pushed and is disappeared
Breath pushes to the destination mobile terminal to be included:
The user interest data are screened according to custom rule, determines target PUSH message and corresponding first object
Mobile terminal;
Obtain the second mobile terminal for being more than predetermined threshold value with the similarity of the user interest data of first object mobile terminal;
The target PUSH message is pushed to into first mobile terminal, when detecting first mobile terminal to the mesh
When mark PUSH message is browsed, the target PUSH message is pushed to into second mobile terminal.
10. the information push method as described in any one of claim 6-8, it is characterised in that described according to custom rule pair
The user interest data are screened, and determine target PUSH message and corresponding destination mobile terminal, and the target is pushed
Message pushes to the destination mobile terminal also to be included:
The user interest data are screened according to custom rule, determines first object PUSH message and corresponding target
Mobile terminal;
Obtain the second target PUSH message for being more than preset similarity value with the similarity of the first object PUSH message;
First object PUSH message is pushed to into the destination mobile terminal, when detecting the destination mobile terminal to described
When one target PUSH message is browsed, the second target PUSH message is pushed to into the destination mobile terminal.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611255475.0A CN106686086A (en) | 2016-12-29 | 2016-12-29 | Message pushing method and apparatus |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611255475.0A CN106686086A (en) | 2016-12-29 | 2016-12-29 | Message pushing method and apparatus |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106686086A true CN106686086A (en) | 2017-05-17 |
Family
ID=58872319
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611255475.0A Pending CN106686086A (en) | 2016-12-29 | 2016-12-29 | Message pushing method and apparatus |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106686086A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107807958A (en) * | 2017-09-30 | 2018-03-16 | 广东南都全媒体网络科技有限公司 | A kind of article list personalized recommendation method, electronic equipment and storage medium |
CN110060425A (en) * | 2019-04-26 | 2019-07-26 | 佛山市南鼎文化传播有限公司 | Resource promotion method and item circulation based on item circulation manage platform |
CN110704750A (en) * | 2019-10-21 | 2020-01-17 | 秒针信息技术有限公司 | Article pushing method, article pushing device, article pushing equipment and storage medium based on user tags |
CN110874639A (en) * | 2018-08-31 | 2020-03-10 | 珠海格力电器股份有限公司 | Method and device for acquiring operation information |
CN113806662A (en) * | 2021-09-23 | 2021-12-17 | 深圳市易平方网络科技有限公司 | Operation data push control method and device, terminal equipment and storage medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102611785A (en) * | 2011-01-20 | 2012-07-25 | 北京邮电大学 | Personalized active news recommending service system and method for mobile phone user |
US9420413B2 (en) * | 2013-04-02 | 2016-08-16 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Push method, system and server based on location information |
-
2016
- 2016-12-29 CN CN201611255475.0A patent/CN106686086A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102611785A (en) * | 2011-01-20 | 2012-07-25 | 北京邮电大学 | Personalized active news recommending service system and method for mobile phone user |
US9420413B2 (en) * | 2013-04-02 | 2016-08-16 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Push method, system and server based on location information |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107807958A (en) * | 2017-09-30 | 2018-03-16 | 广东南都全媒体网络科技有限公司 | A kind of article list personalized recommendation method, electronic equipment and storage medium |
CN107807958B (en) * | 2017-09-30 | 2021-11-19 | 广东南都全媒体网络科技有限公司 | Personalized article list recommendation method, electronic equipment and storage medium |
CN110874639A (en) * | 2018-08-31 | 2020-03-10 | 珠海格力电器股份有限公司 | Method and device for acquiring operation information |
CN110060425A (en) * | 2019-04-26 | 2019-07-26 | 佛山市南鼎文化传播有限公司 | Resource promotion method and item circulation based on item circulation manage platform |
CN110704750A (en) * | 2019-10-21 | 2020-01-17 | 秒针信息技术有限公司 | Article pushing method, article pushing device, article pushing equipment and storage medium based on user tags |
CN113806662A (en) * | 2021-09-23 | 2021-12-17 | 深圳市易平方网络科技有限公司 | Operation data push control method and device, terminal equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106686086A (en) | Message pushing method and apparatus | |
CN106600343B (en) | Video content associated online video advertisement management method and system | |
CN102929959B (en) | A kind of book recommendation method based on user behavior | |
CN102722709B (en) | Method and device for identifying garbage pictures | |
US20100293170A1 (en) | Social network message categorization systems and methods | |
US9600561B2 (en) | Computer-implemented system and method for generating an interest profile for a user from existing online profiles | |
CN108537578A (en) | Advertisement sending method based on big data and device | |
US20140004489A1 (en) | Method and apparatus for providing emotion expression service using emotion expression identifier | |
CN110362740B (en) | Water conservancy portal information hybrid recommendation method | |
CN102521767A (en) | Method and system for publishing network advertising information | |
CN103177384A (en) | Network advertisement putting method based on user interest spectrum | |
CN107292463A (en) | A kind of method and system that the project evaluation is carried out to application program | |
CN103019550A (en) | Real-time display method and system for associated content | |
US20140324571A1 (en) | System and method for selecting and rendering content | |
CN108073667B (en) | Method for generating user browsing attributes, and non-transitory computer readable medium | |
CN104933191A (en) | Spam comment recognition method and system based on Bayesian algorithm and terminal | |
CN106095884A (en) | A kind of relative article information processing method based on picture and device | |
CN104866557A (en) | Customized just-in-time learning support system and method based on constructivist learning theory | |
CN106339891A (en) | Intelligent analysis method and system based on large data acquisition | |
CN103268330A (en) | User interest extraction method based on image content | |
CN103838754A (en) | Information searching device and method | |
CN107277115A (en) | A kind of content delivery method and device | |
CN107203558A (en) | Object recommendation method and apparatus, recommendation information treating method and apparatus | |
CN103399855A (en) | Behavior intention determining method and device based on multiple data sources | |
US9508087B1 (en) | Identifying similar display items for potential placement of content items therein |
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: 20170517 |