CN105045831B - A kind of information push method and device - Google Patents
A kind of information push method and device Download PDFInfo
- Publication number
- CN105045831B CN105045831B CN201510372471.XA CN201510372471A CN105045831B CN 105045831 B CN105045831 B CN 105045831B CN 201510372471 A CN201510372471 A CN 201510372471A CN 105045831 B CN105045831 B CN 105045831B
- Authority
- CN
- China
- Prior art keywords
- user
- message
- push
- time
- user group
- 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.)
- Active
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
Abstract
The embodiment of the invention discloses a kind of information push method and device, and platform is pushed applied to message, the described method includes：PUSH message is treated in acquisition；PUSH message the first user group interested is treated in acquisition；According to time of return disaggregated model trained in advance, the period divided in advance corresponding to the time interval of each user's return message push platform in the first user group is predicted；The period divided in advance corresponding to time interval that message push platform is returned to according to each user in the first user group of prediction, obtain the corresponding time parameter values r of each useri；According to the time parameter riUser in first user group is ranked up, according to ranking results, selection is with the default user for pushing user volume respective numbers as the potential user group for treating PUSH message；Treat that PUSH message is pushed to identified potential user group by described.This just effectively improves the exposure rate of message, reduces the waste of push platform resource.
Description
Technical field
The present invention relates to field of information processing, more particularly to a kind of information push method and device.
Background technology
With the development of information technology, the message push technology based on network and communication instead of traditionally on paper and propagate
The mode of message.Message push technology be by automatically delivering message to user, to reduce the time that user is used for message search,
Therefore, suitable message is pushed to suitable potential user group, it is necessary to which message and potential user group are made sufficiently
Analysis, and both are effectively matched.
Existing information push method, is the targeted customer that message push is chosen according to user's degree interested in message
Group.Its operating process is as follows：The interest level of message is ranked up according to user first, secondly choosing to offset
The most interested top n user of breath is potential user group, and message finally is pushed to potential user group.But since message has
Timeliness, selected N number of user can not necessarily log within the message effective period, cause the wave of push platform resource
Take, on the other hand, the lower slightly user of interest level may then log in message valid time section and but can not receive message.
As it can be seen that there is not necessarily can be when message is effective due to selected user for the method for above-mentioned message push
Between log in section caused by source of exposure rate it is low, and the shortcomings that waste platform resource.
The content of the invention
The embodiment of the invention discloses a kind of information push method and device, the method for solving existing message push
Source of exposure rate is low, and the shortcomings that waste platform resource；Concrete technical scheme is as follows：
A kind of information push method, platform is pushed applied to message, the described method includes：
PUSH message is treated in acquisition；
PUSH message the first user group interested is treated in acquisition；
Obtain the behavioural characteristic of each user in first user group；
According to time of return disaggregated model trained in advance, predict that it is flat to return to message push by each user in the first user group
The period divided in advance corresponding to the time interval of platform；Included in the time of return disaggregated model：The behavior of user is special
Sign returns to the relation of the period divided in advance corresponding to the time interval of message push platform with user；
It is advance corresponding to the time interval for returning to message push platform according to each user in the first user group of prediction
The period of division, obtains the corresponding time parameter values r of each useri；The time parameter values riDisappear for identifying user and returning
The length of the time interval of breath push platform；
According to the time parameter values riUser in first user group is ranked up, according to ranking results, selection with
The user of default push user volume respective numbers is as the potential user group for treating PUSH message；
Treat that PUSH message is pushed to identified potential user group by described.
In a kind of preferred embodiment of this bright embodiment, the training process of the time of return disaggregated model includes：
By user twice in succession login message push platform time interval be divided into T period, T >=2；
M user is obtained in the behavior record sample before given time and returns to after given time message first and pushes
The time interval of platform, M >=2；
Determine the corresponding user vector set of each user in M user, the user is included in the user vector set
Identification information, the behavioural characteristic that is extracted from the user's behavior record sample and the user return and disappear first after given time
The period divided in advance corresponding to the time interval of breath push platform；
Identified N number of user vector set is trained by default grader, obtains time of return classification mould
Type.
Wherein, the grader includes：One kind in random forest, logistic regression and support vector machine classifier.
In a kind of preferred embodiment of this bright embodiment, each user correspondence one is emerging in first user group
Interesting parameter value di；The interest parameter value diFor identified user interest degree；
It is described according to the time parameter values riUser in first user group is ranked up, according to ranking results, choosing
The user with default push user volume respective numbers is selected as the potential user group for treating PUSH message, including：
According to the interest parameter value diWith the time parameter values riUser in first user group is ranked up, root
According to ranking results, selection is with the default user for pushing user volume respective numbers as the potential user group for treating PUSH message.
Wherein, it is described according to the interest parameter value diWith the time parameter values riTo the user in the first user group into
Row sequence, according to ranking results, selection is with presetting the user of push user volume respective numbers as the mesh for treating PUSH message
User group is marked, including：
According to interest parameter value diWith time parameter values riCalculate w1×di+w2×ri, according to result of calculation to the first user
User in group is ranked up, and according to ranking results, selects the user with default push user volume respective numbers to be treated as described
The potential user group of PUSH message；Wherein, w1For interest parameter value diCorresponding weight, w2For time parameter values riCorresponding power
Weight.
Corresponding to above method embodiment, present invention also offers a kind of message pusher, is pushed applied to message
Platform, described device include：
Message obtains module, and PUSH message is treated for obtaining；
First user group obtains module, and PUSH message the first user group interested is treated for obtaining；
Behavioural characteristic obtains module, for obtaining the behavioural characteristic of each user in first user group；
Time interval prediction module, for according to time of return disaggregated model trained in advance, predicting in the first user group
Each user returns to the period divided in advance corresponding to the time interval of message push platform；The time of return classification mould
Included in type：The behavioural characteristic of user returns to the time divided in advance corresponding to the time interval of message push platform with user
The relation of section；
Time parameter values obtain module, and message push platform is returned for each user in the first user group according to prediction
Time interval corresponding to the period divided in advance, obtain the corresponding time parameter values r of each useri；The time ginseng
Numerical value riThe length for returning to the time interval of message push platform for identifying user；
Potential user group determining module, for according to the time parameter values riUser in first user group is arranged
Sequence, according to ranking results, selection and the user of default push user volume respective numbers use as the target for treating PUSH message
Family group；
Message pushing module, for treating that PUSH message is pushed to identified potential user group by described.
In a kind of preferred embodiment of this bright embodiment, the training for training time of return disaggregated model is further included
Module, the training module include：
Period divide submodule, for by user twice in succession login message push platform time interval be divided into T
A period, T >=2；
User behavior acquisition submodule, for obtaining behavior record sample of the M user before given time and referring to
Timing returns to the time interval of message push platform, M >=2 first after carving；
User vector set determination sub-module, for determining the corresponding user vector set of each user, institute in M user
State the identification information comprising the user in user vector set, the behavioural characteristic extracted from the user's behavior record sample and be somebody's turn to do
User returns to the period divided in advance corresponding to the time interval of message push platform first after given time；
Model determination sub-module, for identified N number of user vector set to be trained by default grader,
Obtain time of return disaggregated model.
Wherein, the grader includes：One kind in random forest, logistic regression and support vector machine classifier.
In a kind of preferred embodiment of this bright embodiment, each user correspondence one is emerging in first user group
Interesting parameter value di；The interest parameter value diFor identified user interest degree；
The potential user group determining module, is specifically used for：
According to the interest parameter value diWith the time parameter values riUser in first user group is ranked up, root
According to ranking results, selection is with the default user for pushing user volume respective numbers as the potential user group for treating PUSH message.
Wherein, potential user group determining module, is specifically used for：
According to interest parameter value diWith time parameter values riCalculate w1×di+w2×ri, according to result of calculation to the first user
User in group is ranked up, and according to ranking results, selects the user with default push user volume respective numbers to be treated as described
The potential user group of PUSH message；Wherein, w1For interest parameter value diCorresponding weight, w2For time parameter values riCorresponding power
Weight.
The technical solution of the embodiment of the present invention, treats the interest level of PUSH message by user and the possibility of user is returned
Return the potential user group of the time interval two of message push platform because usually determining to treat PUSH message；
The potential user group determined by this method, not only has higher interest-degree to the message pushed, while again
Have it is larger may be logged in message valid time section, this just effectively improves the exposure rate of message, while reduces and disappear in storage
The situation that user does not log in breath effective period of time, reduces the waste of push platform resource.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
There is attached drawing needed in technology description to be briefly described, it should be apparent that, drawings in the following description are only this
Some embodiments of invention, for those of ordinary skill in the art, without creative efforts, can be with
Other attached drawings are obtained according to these attached drawings.
Fig. 1 is a kind of flow chart of information push method provided in an embodiment of the present invention；
Fig. 2 is a kind of structure diagram of message pusher provided in an embodiment of the present invention.
Embodiment
Below in conjunction with the attached drawing in the embodiment of the present invention, the technical solution in the embodiment of the present invention is carried out clear, complete
Site preparation describes, it is clear that described embodiment is only part of the embodiment of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, those of ordinary skill in the art are obtained every other without creative efforts
Embodiment, belongs to the scope of protection of the invention.
The embodiment of the present invention provide firstly a kind of information push method, and platform is pushed applied to message, as shown in Figure 1,
This method can include：
PUSH message is treated in S101, acquisition；
Described treats that PUSH message can be needed according to the business of message push platform to determine, the specific implementation of this step
Method can realize that the present invention is not especially limited herein using the related art scheme of the prior art.
PUSH message the first user group interested is treated in S102, acquisition；
The specific implementation method of this step can be realized using the related art scheme of the prior art, and the present invention is herein not
Make specific limit.
For example, in actual application, it can determine that all users treat the sense of PUSH message according to default rule
Level of interest；In the prior art, for determining that user has much the method for message interest level, the present invention does not make to have with this
Body limits.
Then it is the first user group that the preceding S user being most interested in message is chosen further according to default threshold value.Described
Default threshold value can be predefined by research staff, and the present invention is not limited thereto.
S103, obtains the behavioural characteristic of each user in first user group；
In actual application, the behavioural characteristic of user can be obtained according to the behavior record of user；Need to illustrate
, the behavior record of described user is to push the relevant behavior record of platform with message.Behavior obtained in this step
It is characterized in the behavioural characteristic corresponding to time of return disaggregated model；Specifically, the behavioural characteristic obtained in this step should
This is corresponding with the behavioural characteristic utilized during training time of return disaggregated model.
For example, the behavioural characteristic utilized during training time of return disaggregated model includes：User is at the appointed time
Interior viewing amount of video, the channel of viewing；So obtaining the behavioural characteristic of each user in first user group should also wrap
Include the viewing amount of video of user within a specified time, the channel of viewing.
S104, according to time of return disaggregated model trained in advance, predicts that each user returns to message in the first user group
Push the period divided in advance corresponding to the time interval of platform；Included in the time of return disaggregated model：User's
Behavioural characteristic returns to the relation of the period divided in advance corresponding to the time interval of message push platform with user；
In practical applications, the training process of described time of return disaggregated model can be as follows：
1st, by user twice in succession login message push platform time interval be divided into T period, T >=2；
The division of period can be determined according to the actual needs for treating PUSH message, if treating that PUSH message compares the time
Sensitivity, can carry out period division in units of hour, day, can also be by more days even with week, the moon if sensitiveness is poor
The period is divided Deng for unit.
For example, when the timeliness for treating PUSH message is shorter i.e. more sensitive to the time, the period can be divided into：" 5 it is small when
Within ", " when 5-12 is small ", when small " 12-1 day ", "-2 days 1 day " and " more than 2 days ".
2nd, M user is obtained in the behavior record sample before given time and return to message first after given time and push away
Send the time interval of platform, M >=2；
The quantity M of described user is can be predefined by research staff.The present invention is not especially limited with this.Can be with
Understand, M is bigger, and obtained model can be more accurate, but at the same time, calculation amount can also increase.
Described given time is also predetermined by research staff., can when obtaining M user behavior record sample
To obtain M user since given time, and the row for message push platform before given time in certain time
To record sample.
3rd, determine the corresponding user vector set of each user in M user, the use is included in the user vector set
The identification information at family, the behavioural characteristic extracted from the user's behavior record sample and the user return first after given time
The period divided in advance corresponding to the time interval of message push platform；
In practical applications, can be that the period divided in advance, class indication was set, for example, when the period is divided into：
When small " 5 within ", " when 5-12 is small ", when small " 12-1 day ", "-2 days 1 day " and when " more than 2 days ", can will when small " 5 within "
Setting is identified as 1, and " when 5-12 is small " sets and be identified as 2, and so on, last setting in " more than 2 days " is identified as 5.
In this case, M user vector set can be expressed as：
{ user id1, feature 1, feature 2, feature 3 ... .1 }；
{ user id2, feature 1, feature 2, feature 3 ... .4 }；
….
{ user idM, feature 1, feature 2, feature 3 ... .5 }；
In above-mentioned user vector set, first is classified as the unique identification information of user, last is classified as the time of user
The class indication of the interval corresponding period divided in advance；
It should be noted that in user vector set the order of each element be can be predetermined by research staff, use
The identification information at family can be in secondary series, or other row.But corresponding element should be in each user vector set
Same position in each user vector set.For example, if the unique identification information of user is located in user vector set
Two, then the unique identification information of M user should all be located at the second in each user vector set.
The behavioural characteristic extracted in sample is recorded from user behavior, can be true previously according to being actually needed by research staff
Fixed, the present invention is not especially limited herein.It is understood that the behavioural characteristic corresponding to different types of message push platform
May be different.
4th, identified N number of user vector set is trained by default grader, obtains time of return classification
Model.
Described grader can use correlation classifier of the prior art, for example, random forest grader, logic are returned
Return grader or support vector machine classifier.The concrete form present invention of grader is not limited thereto, can be by research staff
Made choice according to actual needs.
In practical applications, a number of other user vectors can also be gathered again with the method for above-mentioned definite user vector set
Set, for carrying out the test of generalization ability to obtained time of return disaggregated model.
After time of return disaggregated model is obtained, according to the behavioural characteristic of each user obtained in step S103, so that it may
To predict that each user returns to the period divided in advance corresponding to the time interval of message push platform.
S105, according to corresponding to the time interval of each user's return message push platform in the first user group of prediction
The period divided in advance, obtains the corresponding time parameter values r of each useri；The time parameter values riReturned for identifying user
Return the length of the time interval of message push platform；
In practical applications, the correspondence of divided period and score value can be predefined, and the time divided
The time for the return message push platform that segment table shows is shorter, its score value is higher；For example, when the period is divided into：" 5 it is small when with
It is interior ", " when 5-12 is small ", when small " 12-1 day ", "-2 days 1 day " and when " more than 2 days ", when small " 5 within " represent the user not
Come 5 it is small when it is interior be likely to again return to platform, which will correspond to a higher score value, corresponding point of " when 5-12 is small "
Be worth it is lower slightly, and so on, " more than 2 days " corresponding score value is minimum.
The corresponding time parameter values r of useriThe user exactly predicted by time of return disaggregated model returns to message
Push the score value of the period divided in advance corresponding to the time interval of platform.
S106, according to the time parameter values riUser in first user group is ranked up, according to ranking results, choosing
The user with default push user volume respective numbers is selected as the potential user group for treating PUSH message；
In practical applications, the embodiment of this step can be：By the time of each user in the first user group
Parameter value riArranged from small to large, then according to default push user volume, from time parameter values riMinimum user starts,
Selection is with the default user for pushing user volume respective numbers as the potential user group for treating PUSH message；For example, default push away
It is 10 to send user volume, then selects time parameter values ri10 minimum users are as the potential user group for treating PUSH message.
In addition, in practical applications, each user can correspond to an interest parameter value d in the first user groupi；It is described
Interest parameter value diFor identified user interest degree；
At this time, step S106 can also be：According to the interest parameter value diWith the time parameter values riTo the first user
User in group is ranked up, and according to ranking results, selects the user with default push user volume respective numbers to be treated as described
The potential user group of PUSH message.
Specifically, can be according to interest parameter value diWith time parameter values riCalculate w1×di+w2×ri, according to result of calculation
User in first user group is ranked up, according to ranking results, selection and the user of default push user volume respective numbers
As the potential user group for treating PUSH message；Wherein, w1For interest parameter value diCorresponding weight, w2For time parameter values
riCorresponding weight.
Using w1×di+w2×riResult when being ranked up, result can also be arranged from small to large, then
According to default push user volume, since the user of result minimum, the user with default push user volume respective numbers is selected to make
For the potential user group for treating PUSH message；For example, default push user volume is 10, then 10 users of selection result minimum
As the potential user group for treating PUSH message.
It should be noted that w1And w2It can be determined by research staff according to business needs, the present invention does not make specific herein
Limit.
Explanation is needed further exist for, for interest parameter value diFor, the level of interest that user treats PUSH message is got over
Height, interest parameter value diIt is bigger.Interest parameter value diDefinite method, can be using the related art scheme of the prior art come real
Existing, the present invention is not especially limited herein.PUSH message the first user group interested is treated for example, being obtained in step S102
At the same time, it is possible to determine the interest parameter value d of each user in the first user groupi。
S107, treats that PUSH message is pushed to identified potential user group by described.
This step can realize that the present invention is not especially limited herein using the related art scheme of the prior art.
It can be seen that the technical solution of the embodiment of the present invention by above-mentioned embodiment of the method, push treated by user
The time interval two that the interest level of message and the possibility of user return to message push platform treats that push disappears because usually determining
The potential user group of breath；
The potential user group determined by this method, not only has higher interest-degree to the message pushed, while again
Have it is larger may be logged in message valid time section, this just effectively improves the exposure rate of message, while reduces and disappear in storage
The situation that user does not log in breath effective period of time, reduces the waste of push platform resource.
Corresponding to above method embodiment, present invention also offers a kind of message pusher, as shown in Fig. 2, using
Platform is pushed in message, described device includes：
Message obtains module 101, and PUSH message is treated for obtaining；
First user group obtains module 102, and PUSH message the first user group interested is treated for obtaining；
Behavioural characteristic obtains module 103, for obtaining the behavioural characteristic of each user in first user group；
Time interval prediction module 104, for according to time of return disaggregated model trained in advance, predicting the first user group
In each user return to period divided in advance corresponding to the time interval of message push platform；The time of return classification
Included in model：The behavioural characteristic of user returned with user corresponding to the time interval of message push platform divide in advance when
Between section relation；
Time parameter values obtain module 105, and message push is returned for each user in the first user group according to prediction
The period divided in advance corresponding to the time interval of platform, obtains the corresponding time parameter values ri of each user；When described
Between parameter value ri be used for identify user return message push platform time interval length；
Potential user group determining module 106, for according to the time parameter values ri to the user in the first user group into
Row sequence, according to ranking results, selection is with presetting the user of push user volume respective numbers as the mesh for treating PUSH message
Mark user group；
Message pushing module 107, for treating that PUSH message is pushed to identified potential user group by described.
In practical applications, which can also include being used for the training module for training time of return disaggregated model, described
Training module includes：
Period divide submodule, for by user twice in succession login message push platform time interval be divided into T
A period, T >=2；
User behavior acquisition submodule, for obtaining behavior sample of the M user before given time and when specified
The time interval of message push platform, M >=2 are returned to after quarter first；
User vector set determination sub-module, for determining the corresponding user vector set of each user, institute in M user
State the identification information comprising the user in user vector set, the behavioural characteristic extracted from the user's behavior sample and the user
Return to the period divided in advance corresponding to the time interval of message push platform first after given time；
Model determination sub-module, for identified N number of user vector set to be trained by default grader,
Obtain time of return disaggregated model.
Wherein, the grader can include：One kind in random forest, logistic regression and support vector machine classifier.
In the embodiment of above-mentioned each device, each user corresponds to an interest parameter value in first user group
di；The interest parameter value di is used for identified user interest degree；
The potential user group determining module 106, specifically can be used for：
The user in the first user group is ranked up according to the interest parameter value di and the time parameter values ri, root
According to ranking results, selection is with the default user for pushing user volume respective numbers as the potential user group for treating PUSH message.
In practical applications, w1 × di+w2 × ri can be calculated according to interest parameter value di and time parameter values ri, according to
Result of calculation is ranked up the user in the first user group, according to ranking results, selection and default push user volume respective counts
The user of amount is as the potential user group for treating PUSH message；Wherein, w1 is the corresponding weights of interest parameter value di, when w2 is
Between the corresponding weights of parameter value ri.
It can be seen that the technical solution of the embodiment of the present invention by above-mentioned embodiment of the method, push treated by user
The time interval two that the interest level of message and the possibility of user return to message push platform treats that push disappears because usually determining
The potential user group of breath；
The potential user group determined by this method, not only has higher interest-degree to the message pushed, while again
Have it is larger may be logged in message valid time section, this just effectively improves the exposure rate of message, while reduces and disappear in storage
The situation that user does not log in breath effective period of time, reduces the waste of push platform resource.
It should be noted that for device embodiment, since it is substantially similar to embodiment of the method, so description
Fairly simple, the relevent part can refer to the partial explaination of embodiments of method.
For convenience of description, it is divided into various modules during description apparatus above with function to describe respectively.Certainly, this is being implemented
The function of each module can be realized in same or multiple softwares and/or hardware during invention.
It should be noted that herein, relational terms such as first and second and the like are used merely to a reality
Body or operation are distinguished with another entity or operation, are deposited without necessarily requiring or implying between these entities or operation
In any this actual relation or order.Moreover, term " comprising ", "comprising" or its any other variant are intended to
Non-exclusive inclusion, so that process, method, article or equipment including a series of elements not only will including those
Element, but also including other elements that are not explicitly listed, or further include as this process, method, article or equipment
Intrinsic key element.In the absence of more restrictions, the key element limited by sentence "including a ...", it is not excluded that
Also there are other identical element in process, method, article or equipment including the key element.
Each embodiment in this specification is described using relevant mode, identical similar portion between each embodiment
Divide mutually referring to what each embodiment stressed is the difference with other embodiment.It is real especially for device
For applying example, since it is substantially similar to embodiment of the method, so description is fairly simple, related part is referring to embodiment of the method
Part explanation.
Can one of ordinary skill in the art will appreciate that realizing that all or part of step in above method embodiment is
To instruct relevant hardware to complete by program, the program can be stored in computer read/write memory medium,
The storage medium designated herein obtained, such as：ROM/RAM, magnetic disc, CD etc..
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the scope of the present invention.It is all
Any modification, equivalent replacement, improvement and so within the spirit and principles in the present invention, are all contained in protection scope of the present invention
It is interior.
Claims (10)
1. a kind of information push method, platform is pushed applied to message, it is characterised in that the described method includes：
PUSH message is treated in acquisition；
PUSH message the first user group interested is treated in acquisition；
Obtain the behavioural characteristic of each user in first user group；
According to time of return disaggregated model trained in advance, predict that each user returns to message push platform in the first user group
The period divided in advance corresponding to time interval；Included in the time of return disaggregated model：The behavioural characteristic of user with
User returns to the relation of the period divided in advance corresponding to the time interval of message push platform；
It is advance corresponding to the time interval for returning to message push platform according to each user in first user group of prediction
The period of division, obtains the corresponding time parameter values r of each useri；The time parameter values riDisappear for identifying user and returning
The length of the time interval of breath push platform；
According to the time parameter values riUser in first user group is ranked up, according to ranking results, selection is pushed away with default
The user of user volume respective numbers is sent as the potential user group for treating PUSH message；
Treat that PUSH message is pushed to identified potential user group by described.
2. the method as described in claim 1, it is characterised in that the training process of the time of return disaggregated model includes：
By user twice in succession login message push platform time interval be divided into T period, T >=2；
M user is obtained in the behavior record sample before given time and returns to after given time message first and pushes platform
Time interval, M >=2；
Determine the corresponding user vector set of each user in M user, the mark of the user is included in the user vector set
Know information, the behavioural characteristic extracted from the user's behavior record sample and the user to return to message first after given time and push away
Send the period divided in advance corresponding to the time interval of platform；
Identified N number of user vector set is trained by default grader, obtains time of return disaggregated model.
3. method as claimed in claim 2, it is characterised in that the grader includes：Random forest, logistic regression and support
One kind in vector machine classifier.
4. the method as described in claim 1, it is characterised in that each user corresponds to an interest in first user group
Parameter value di；The interest parameter value diFor identified user interest degree；
It is described according to the time parameter values riUser in first user group is ranked up, according to ranking results, is selected and pre-
If the user of user volume respective numbers is pushed as the potential user group for treating PUSH message, including：
According to the interest parameter value diWith the time parameter values riUser in first user group is ranked up, according to row
Sequence is as a result, selection pushes the user of user volume respective numbers with presetting as the potential user group for treating PUSH message.
5. method as claimed in claim 4, it is characterised in that described according to the interest parameter value diWith the time parameter
Value riUser in first user group is ranked up, according to ranking results, selection and default push user volume respective numbers
User as the potential user group for treating PUSH message, including：
According to interest parameter value diWith time parameter values riCalculate w1×di+w2×ri, according to result of calculation in the first user group
User be ranked up, according to ranking results, selection is waited to push with the default user for pushing user volume respective numbers as described
The potential user group of message；Wherein, w1For interest parameter value diCorresponding weight, w2For time parameter values riCorresponding weight.
6. a kind of message pusher, pushes platform, it is characterised in that described device includes applied to message：
Message obtains module, and PUSH message is treated for obtaining；
First user group obtains module, and PUSH message the first user group interested is treated for obtaining；
Behavioural characteristic obtains module, for obtaining the behavioural characteristic of each user in first user group；
Time interval prediction module, for according to time of return disaggregated model trained in advance, predicting each in the first user group
User returns to the period divided in advance corresponding to the time interval of message push platform；In the time of return disaggregated model
Comprising：The behavioural characteristic of user returns to the period divided in advance corresponding to the time interval of message push platform with user
Relation；
Time parameter values obtain module, and message push platform is returned for each user in first user group according to prediction
Time interval corresponding to the period divided in advance, obtain the corresponding time parameter values r of each useri；The time ginseng
Numerical value riThe length for returning to the time interval of message push platform for identifying user；
Potential user group determining module, for according to the time parameter values riUser in first user group is ranked up, root
According to ranking results, selection is with the default user for pushing user volume respective numbers as the potential user group for treating PUSH message；
Message pushing module, for treating that PUSH message is pushed to identified potential user group by described.
7. device as claimed in claim 6, it is characterised in that further include the training mould for training time of return disaggregated model
Block, the training module include：
Period divide submodule, for by user twice in succession login message push platform time interval be divided into T when
Between section, T >=2；
User behavior acquisition submodule, for obtaining behavior record sample of the M user before given time and when specified
The time interval of message push platform, M >=2 are returned to after quarter first；
User vector set determination sub-module, for determining the corresponding user vector set of each user, the use in M user
Identification information comprising the user, the behavioural characteristic extracted from the user's behavior record sample and the user in the vector set of family
Return to the period divided in advance corresponding to the time interval of message push platform first after given time；
Model determination sub-module, for identified N number of user vector set to be trained by default grader, obtains
Time of return disaggregated model.
8. device as claimed in claim 7, it is characterised in that the grader includes：Random forest, logistic regression and support
One kind in vector machine classifier.
9. device as claimed in claim 6, it is characterised in that each user corresponds to an interest in first user group
Parameter value di；The interest parameter value diFor identified user interest degree；
The potential user group determining module, is specifically used for：
According to the interest parameter value diWith the time parameter values riUser in first user group is ranked up, according to row
Sequence is as a result, selection pushes the user of user volume respective numbers with presetting as the potential user group for treating PUSH message.
10. device as claimed in claim 9, it is characterised in that potential user group determining module, is specifically used for：
According to interest parameter value diWith time parameter values riCalculate w1×di+w2×ri, according to result of calculation in the first user group
User be ranked up, according to ranking results, selection is waited to push with the default user for pushing user volume respective numbers as described
The potential user group of message；Wherein, w1For interest parameter value diCorresponding weight, w2For time parameter values riCorresponding weight.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510372471.XA CN105045831B (en) | 2015-06-30 | 2015-06-30 | A kind of information push method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510372471.XA CN105045831B (en) | 2015-06-30 | 2015-06-30 | A kind of information push method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105045831A CN105045831A (en) | 2015-11-11 |
| CN105045831B true CN105045831B (en) | 2018-04-13 |
Family
ID=54452378
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510372471.XA Active CN105045831B (en) | 2015-06-30 | 2015-06-30 | A kind of information push method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105045831B (en) |
Families Citing this family (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105656762A (en) * | 2016-01-26 | 2016-06-08 | 四川长虹电器股份有限公司 | Message pushing method and message pushing system for mobile office software |
| CN105872050A (en) * | 2016-03-30 | 2016-08-17 | 北京奇艺世纪科技有限公司 | Message pushing method, device and system |
| CN107465741B (en) * | 2017-08-02 | 2020-09-04 | 北京星选科技有限公司 | Information pushing method and device |
| CN107577736B (en) * | 2017-08-25 | 2021-12-17 | 武汉数字智能信息科技有限公司 | File recommendation method and system based on BP neural network |
| CN107818479B (en) * | 2017-10-27 | 2021-03-05 | 携程旅游网络技术（上海）有限公司 | Preferential information pushing method, system, storage medium and electronic equipment |
| CN108052597A (en) * | 2017-12-12 | 2018-05-18 | 百度在线网络技术（北京）有限公司 | Method, apparatus, equipment and the computer-readable medium of news push |
| CN108111399B (en) * | 2017-12-27 | 2021-06-04 | Oppo广东移动通信有限公司 | Message processing method, device, terminal and storage medium |
| CN108171267B (en) * | 2017-12-28 | 2022-03-22 | 阿里巴巴（中国）有限公司 | User group division method and device and message pushing method and device |
| CN109995798B (en) * | 2017-12-29 | 2021-05-04 | Tcl科技集团股份有限公司 | Information pushing method and equipment thereof |
| CN108833521B (en) * | 2018-06-06 | 2021-01-08 | 广州视源电子科技股份有限公司 | Message pushing method, device, system, computer equipment and storage medium |
| CN111062734B (en) * | 2018-10-16 | 2023-08-25 | 北京字节跳动网络技术有限公司 | Method and device for reordering delivery information, electronic equipment and readable medium |
| CN109936498B (en) * | 2019-02-26 | 2021-11-19 | 企知道网络技术有限公司 | Full ordering method for prompt of instant communication message |
| CN110264276A (en) * | 2019-06-25 | 2019-09-20 | 广州视源电子科技股份有限公司 | Determine method, apparatus, equipment and the storage medium of push user group |
| CN110957044A (en) * | 2019-09-20 | 2020-04-03 | 上海派拉软件股份有限公司 | Health management method based on improved logistic regression model |
| CN111369281A (en) * | 2020-02-28 | 2020-07-03 | 深圳前海微众银行股份有限公司 | Online message processing method, device, equipment and readable storage medium |
| WO2021168830A1 (en) * | 2020-02-28 | 2021-09-02 | 深圳市欢太科技有限公司 | Content pushing method and apparatus, server and storage medium |
| CN112202663A (en) * | 2020-09-28 | 2021-01-08 | 厦门美柚股份有限公司 | Message pushing method, device, terminal and medium |
| CN112990324A (en) * | 2021-03-23 | 2021-06-18 | 李光伟 | Resource pushing method based on big data online mode and deep learning service system |
| CN115134765A (en) * | 2021-03-29 | 2022-09-30 | 中兴通讯股份有限公司 | Message processing method, message processing apparatus, distribution control apparatus, and storage medium |
| CN113556572B (en) * | 2021-07-21 | 2023-03-14 | 云南腾云信息产业有限公司 | Live broadcast starting method and device and electronic equipment |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103685502A (en) * | 2013-12-09 | 2014-03-26 | 腾讯科技（深圳）有限公司 | Message pushing method, device and system |
| CN104462270A (en) * | 2014-11-24 | 2015-03-25 | 华为软件技术有限公司 | Information recommendation method and device |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140024335A1 (en) * | 2012-07-17 | 2014-01-23 | Chidopi Co., Ltd. | Methods for collecting usage information of portable devices, and utilizing methods and platforms thereof |
| US20150095456A1 (en) * | 2013-09-30 | 2015-04-02 | Ashok Deepak Shah | Realtime status data display with push advertising |
-
2015
- 2015-06-30 CN CN201510372471.XA patent/CN105045831B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103685502A (en) * | 2013-12-09 | 2014-03-26 | 腾讯科技（深圳）有限公司 | Message pushing method, device and system |
| CN104462270A (en) * | 2014-11-24 | 2015-03-25 | 华为软件技术有限公司 | Information recommendation method and device |
Non-Patent Citations (4)
| Title |
|---|
| A universal push service for mobile devices;Jonas Brustel等;《Complex, Intelligent and Software Intensive Systems (CISIS), 2012 Sixth International Conference on》;20120723;第40-45页 * |
| Research on the advertisement effect of push type mobile advertisement;Jung Woo Lee等;《Cooperation and Promotion of Information Resources in Science and Technology, 2009. COINFO "09. Fourth International Conference on》;20091228;第137-142页 * |
| 基于Epoll的消息推送系统的设计与实现;余光远;《中国优秀硕士学位论文全文数据库 信息科技辑》;20130715(第07期);第I138-248页 * |
| 基于动态权值优先级队列的移动消息推送策略;蒋溢等;《计算机工程与设计》;20131031;第34卷(第10期);第3520-3524页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105045831A (en) | 2015-11-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105045831B (en) | A kind of information push method and device | |
| CN103685502B (en) | A kind of information push method, apparatus and system | |
| CN108108821A (en) | Model training method and device | |
| CN108733706B (en) | Method and device for generating heat information | |
| CN110472879B (en) | Resource effect evaluation method and device, electronic equipment and storage medium | |
| CN107689008A (en) | A kind of user insures the method and device of behavior prediction | |
| CN105023165A (en) | Method, device and system for controlling release tasks in social networking platform | |
| CN108536608A (en) | Page versions test method and device | |
| CN103647800A (en) | Method and system of recommending application resources | |
| CN110222975A (en) | A kind of loss customer analysis method, apparatus, electronic equipment and storage medium | |
| CN106294830A (en) | The recommendation method and device of multimedia resource | |
| CN107885796A (en) | Information recommendation method and device, equipment | |
| CN105335876A (en) | Effect tracking method and apparatus of advertisement put on media | |
| CN110457566B (en) | Information screening method and device, electronic equipment and storage medium | |
| Hatton | Early prioritisation of goals | |
| CN110442788A (en) | A kind of information recommendation method and device | |
| CN104993962A (en) | Method and system for obtaining use state of terminal | |
| CN108765094A (en) | Insurance data processing method, device, computer equipment and storage medium | |
| CN109165691A (en) | Training method, device and the electronic equipment of the model of cheating user for identification | |
| CN108345601A (en) | Search result ordering method and device | |
| Carlson et al. | Cost-effective sampling design applied to large-scale monitoring of boreal birds | |
| CN106130756A (en) | A kind of method and device predicting access content clicking rate | |
| Waage et al. | Nonstationary water planning: an overview of several promising planning methods 1 | |
| CN110264276A (en) | Determine method, apparatus, equipment and the storage medium of push user group | |
| CN110428276A (en) | Choose method, apparatus and storage medium that region is launched in advertisement under line |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |