CN104281594B - The reminding method and device of user's overlay capacity information - Google Patents
The reminding method and device of user's overlay capacity information Download PDFInfo
- Publication number
- CN104281594B CN104281594B CN201310279704.2A CN201310279704A CN104281594B CN 104281594 B CN104281594 B CN 104281594B CN 201310279704 A CN201310279704 A CN 201310279704A CN 104281594 B CN104281594 B CN 104281594B
- Authority
- CN
- China
- Prior art keywords
- classification
- user
- target
- overlay capacity
- accumulation
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
This application discloses the reminding method and device of user's overlay capacity information, methods described includes：Obtain basic data；According to each classification, the difference of Sorting distance is classified to classification in classification sequence；Classification in same subsequence is divided into multiple classification combinations, the combination of each classification is calculated relative to user's overlay capacity growth rate that previous classification is combined；According to each user's overlay capacity growth rate obtained in same subsequence and preset Function Format, the fitting function under correspondence Sorting distance is obtained；After information dispensing side have selected target keyword, multiple target classifications in system are mapped as, the Sorting distance of each objective cross is calculated；Each self-corresponding fitting function is utilized respectively, accumulation user's overlay capacity of target classification combination is estimated out, to be supplied to information dispensing side.By the application, amount of calculation can be reduced, and solve the problem of same user repeats in different classifications.
Description
Technical field
The application is related to the information alert technical field during point of interest orientation impression information, and more particularly to user is covered
The reminding method and device of lid amount information.
Background technology
It is specific that some are delivered in some internet service platforms, in the webpage for often having information dispensing direction business platform
Information, while concern to the webpage by user, these customizing messages can be also concerned about, or even be clicked, and imported into letter
The page of breath dispensing side oneself brings flow for it.Initially, the customizing messages delivered in same webpage is usually fixed, but
For large-scale business platform, its user (viewer) is large number of, for the different access user of same webpage,
The point of user's concern is probably different.For example, for certain E-commerce transaction platform, its business object provided can be from more
Individual dimension is divided into multiple classifications, such as clothing, digital product class, and some users may be interested in clothes category information,
And other users may logarithmic code product it is interested etc..Now, if thrown into same webpage (such as certain website homepage)
Fixed customizing messages is put, then means that only certain customers can be interested in the information, for other users, equivalent to
Waste the space of a whole page where the customizing messages.
Therefore, the information that " point of interest orientation " is proposed in some systems delivers mode, also information dispensing side is selected
The oriented approach that matches of keyword and the possible point of interest of user.Information dispensing side can be by inputting its customizing messages
The modes such as description information obtain the keyword of system recommendation, can also these keywords be screened, according to the selection result,
Keyword can be mapped to the classification in system by system, and the keyword of information dispensing side selection is " cost performance is high ", then system
" cost performance is high " can be mapped as classification " digital product ", and if the keyword of selection is " comfortable feel ", then system can
So that " comfortable feel " is mapped as into classification " household articles ", etc., so can be with the point of interest of awareness information dispensing side, this interest
Point can be represented with each corresponding classification of selected keyword.
Simultaneity factor will analyze the current browse webpage content and history focus of each information browse user, obtain each
The focus of individual user, this focus can equally be represented with the classification in custom system.For example, the focus bag of user's first
Include " clothes ", " digital product " etc..In this manner it is possible to the focus of the point of interest of information dispensing side and user is matched,
The user crowd that the customizing messages of information dispensing side matches is delivered.That is, for same webpage, pin
To different IDs, itself it can be seen that customizing messages be probably different, but the focus all substantially with user matches,
It therefore, it can the space of a whole page for making full use of netpage user to show customizing messages so that same webpage can be different letters simultaneously
Bring customer flow in breath dispensing side.
In the information dispensing mode of above-mentioned this point of interest orientation, when information dispensing side have selected some or some passes
After keyword, system, which can also estimate corresponding classification combination, can great user's overlay capacity, and be shown to information dispensing
Side, and then information dispensing side can decide whether that these keywords of selection are delivered accordingly.However, flat for large-scale business
For platform, wherein the classification quantity of the business object included is usually very many, if will accurate various possible classifications
Corresponding user's overlay capacity is combined, it is necessary to huge amount of calculation.For example assume there is N number of classification, then need to calculate 2^N classification group
User's overlay capacity of conjunction, for the N of magnitudes thousands of or even up to ten thousand, this amount of calculation is even for existing extensive distribution
Formula computing system is all difficult to.
In addition, when calculating user's overlay capacity of classification combination, simply corresponding it can't cover each classification is independent
Lid amount is added, because same user may pay close attention to different classifications, such as user A is electrical type user, is mother and baby colony again
User, is sports fan again, if the interest classification of information dispensing side selection includes these three classifications, that user A can only
A user is calculated, and three times can not be calculated.That is flow is estimated and also to solve same user and repeated out in different classifications
Existing the problem of.
In a word, in the urgent need to the technical problem that those skilled in the art solve is that：How to carry out what flow was estimated
During, amount of calculation is reduced, and solve the problem of same user repeats in different classifications.
The content of the invention
This application provides the reminding method and device of user's overlay capacity information, amount of calculation can be reduced, and solve same
The problem of user repeats in different classifications.
This application provides following scheme：
A kind of reminding method of user's overlay capacity information, including：
The category information that each user according to getting in advance pays close attention to respectively, obtains basic data；The basic data
Include the single user's overlay capacity of each classification, according to the classification sequence obtained after the single descending sequence of user's overlay capacity
Row, and sequence number of each classification in classification sequence；
According to each classification, the difference of Sorting distance is classified to classification in classification sequence, obtains the sub- sequence of preset number
Row；Wherein, the Sorting distance is the difference between the sequence number of two classifications, in same subsequence, between adjacent classification
Sequence number difference is equal；
Classification in same subsequence is divided into multiple classification combinations so that each classification combination is than previous classification group
Many classifications are closed, the category information paid close attention to respectively according to described each user got in advance counts the combination of each classification
The not duplicate customer overlay capacity of accumulation, and the combination of each classification is calculated relative to user's overlay capacity growth that previous classification is combined
Rate；
It is fitted according to each user's overlay capacity growth rate obtained in same subsequence and preset Function Format,
Obtain the fitting function under correspondence Sorting distance；
After information dispensing side have selected target keyword, the target keyword is mapped as to multiple targets in system
Classification, each target class purpose sequence number and single user's overlay capacity are determined according to the basic data；
Using the minimum target classification of sequence number as target fiducials classification, by the target fiducials classification respectively with other each targets
Classification constitutes objective cross, and calculates two target class purpose Sorting distances in each objective cross；
The corresponding fitting function of Sorting distance of each objective cross is utilized respectively, each objective cross is estimated relative to mesh
Mark accumulation user's increment of benchmark classification, and by the institute of the single user's overlay capacity of target fiducials classification and each objective cross
State accumulation user increment to be added, estimate out accumulation user's overlay capacity when all target classifications are combined；
When receiving the request for obtaining accumulation user's overlay capacity, it is supplied to information to throw the accumulation user's overlay capacity estimated
The side of putting.
A kind of suggestion device of user's overlay capacity information, including：
Basic data acquiring unit, for the category information paid close attention to respectively according to each user got in advance, is obtained
Basic data；The basic data includes the single user's overlay capacity of each classification, according to single user's overlay capacity by big
The classification sequence obtained after to small sequence, and sequence number of each classification in classification sequence；
Classification taxon, for according to each classification in classification sequence Sorting distance difference classification is classified,
Obtain preset number subsequence；Wherein, the Sorting distance is the difference between the sequence number of two classifications, in same subsequence
In, the sequence number difference between adjacent classification is equal；
Growth Rate Calculation unit, is combined for the classification in same subsequence to be divided into multiple classifications so that each class
Mesh combination combines many classifications than previous classification, and the classification paid close attention to respectively according to described each user got in advance is believed
Breath, counts the not duplicate customer overlay capacity of each classification combined accumulated, and calculates the combination of each classification relative to previous classification
User's overlay capacity growth rate of combination；
Fitting unit, for according to each user's overlay capacity growth rate obtained in same subsequence and preset function
Form is fitted, and obtains the fitting function under correspondence Sorting distance；
Target classification determining unit, for after information dispensing side have selected target keyword, by the target keyword
Multiple target classifications in system are mapped as, each target class purpose sequence number and single user are determined according to the basic data
Overlay capacity；
Objective cross determining unit, for using the minimum target classification of sequence number as target fiducials classification, by the target base
Quasi- classification constitutes objective cross with other each target classifications respectively, and calculates two target class purposes sequences in each objective cross
Distance；
Unit is estimated, the corresponding fitting function of the Sorting distance for being utilized respectively each objective cross estimates each mesh
Mark combines accumulation user's increment relative to target fiducials classification, and by the single user's overlay capacity of target fiducials classification and respectively
The accumulation user increment of individual objective cross is added, and estimates out accumulation user when all target classifications are combined
Overlay capacity；
Tip element, for when receiving the request for obtaining accumulation user's overlay capacity, the accumulation user estimated to be covered
Amount is supplied to information dispensing side.
The specific embodiment provided according to the application, this application discloses following technique effect：
By the embodiment of the present application, using the classification in the subsequence of limited quantity as representative, the combination of some classifications is calculated
The not duplicate customer overlay capacity of accumulation, and can also obtain corresponding fitting function under each Sorting distance, to be fitted pair
Answer under Sorting distance, certain classification combines the accumulation user's growth rate combined relative to previous classification, and then just can be according to each
Accumulation user's overlay capacity of each classification combination, is estimated out by information dispensing side in fitting function and subsequence under Sorting distance
Accumulation user's overlay capacity when the target classification of selection is combined, to be pointed out accordingly to information dispensing side.Can
See, by that with upper type, amount of calculation can be narrowed down in the range of the corresponding subsequence of Sorting distance of limited quantity so that meter
Calculation amount is limited in the range of computing system can realize.Meanwhile, entered using the not duplicate customer overlay capacity of classification combined accumulated
The fitting of line function and it is follow-up estimate, therefore, solve same user the problem of inhomogeneity repeats now.
Certainly, any product for implementing the application it is not absolutely required to while reaching all the above advantage.
Brief description of the drawings
, below will be to institute in embodiment in order to illustrate more clearly of the embodiment of the present application or technical scheme of the prior art
The accompanying drawing needed to use is briefly described, it should be apparent that, drawings in the following description are only some implementations of the application
Example, for those of ordinary skill in the art, on the premise of not paying creative work, can also be obtained according to these accompanying drawings
Obtain other accompanying drawings.
Fig. 1 is the flow chart for the method that the embodiment of the present application is provided；
Fig. 2 is the schematic diagram for the device that the embodiment of the present application is provided.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present application, the technical scheme in the embodiment of the present application is carried out clear, complete
Site preparation is described, it is clear that described embodiment is only some embodiments of the present application, rather than whole embodiments.It is based on
Embodiment in the application, the every other embodiment that those of ordinary skill in the art are obtained belongs to the application protection
Scope.
In the embodiment of the present application, in order to when estimating user's overlay capacity, amount of calculation is reduced, and solve same user
The problem of repeating in the different classifications that same classification is combined, employs and user's overlay capacity that classification is combined is segmented
The mode of fitting, the accumulation user's overlay capacity combined by fitting function to classification, and it is supplied to information dispensing side.It is right below
This is introduced in detail.
Referring to Fig. 1, the embodiment of the present application provides a kind of reminding method of user's overlay capacity information, and this method can include
Following steps：
S101：The category information that each user according to getting in advance pays close attention to respectively, obtains basic data；The basis
Data include the single user's overlay capacity of each classification, according to the class obtained after the single descending sequence of user's overlay capacity
Mesh sequence, and sequence number of each classification in classification sequence；
Wherein, can be to the use that is collected into preset time period when obtaining the category information that each user pays close attention to respectively
The historical operation behavior record at family is counted, and then counts each user interested in which classification respectively.For example, for
Certain user A, has found that webpage majority browsed the user A is all clothing, digital product according to its historical operation behavior record
Webpage where the business object of class, simultaneously it is also possible to find, in the browsed business object of user, finally produced purchase
The business object for buying or subscribing behavior is also clothing and digital product class, therefore, these comprehensive information, it is possible to obtain mostly
Get user's A attention rates highest several classifications.Other users also all carry out similar processing respectively, so, finally
Each user can be obtained interested in which classification respectively.Wherein, specifically obtained from the historical operation behavior record of user
User may refer to the realization in prior art, I will not elaborate to the implementation method of all kinds of purpose attention rates.Certainly, it is right
For each user, its classification paid close attention to may have a lot, in the embodiment of the present application, can only take what family was most paid close attention to
Several classifications are added to specific statistic processes, for example, each user only takes 6 classifications that it is most paid close attention to, certainly, if
Whole classifications of user's actual concern are just added to statistic processes by the classification of certain user concern less than 6.
Get each user respectively to the attention rate of which classification it is higher after, it is possible to obtain each classification independent
User's overlay capacity, wherein, the single user's overlay capacity of certain classification be focus include such purpose number of users.Namely
Say, the classification that can be respectively paid close attention to according to each user, count each classification and distinguish corresponding number of users, by the number of users
Amount is defined as the single user's overlay capacity of each classification.For example, it is assumed that one has three users, respectively A, B, C, wherein：
The classification of user A concerns includes：Clothing, digital product class and toiletries；
The classification of user B concerns includes：Clothing, toiletries and the outdoor class of motion；
The classification of user C concerns includes：Toiletries and digital product class.
Then for clothing, user A and user B are paid close attention to, therefore, and the single user's overlay capacity of the clothing is
2；For digital product class, user A and user C are paid close attention to, therefore, and the single user's overlay capacity of digital product class is
2；For toiletries, user A, B, C are paid close attention to, therefore, and the single user's overlay capacity of toiletries is 3, with such
Push away.Certainly, in actual system, number of users and classification quantity are all a lot, and the merely just simple principle to statistics is entered
Row is introduced.
After the single user's overlay capacity of each classification is obtained, it is possible to according to single user's overlay capacity to each classification
It is ranked up, generates a classification sequence, and the continuous sequence number of each classification imparting respectively in sequence.So, for one
For individual classification, both sides information can be got, one is single user's overlay capacity, and another is exactly in the sequence
Sequence number, can be using the information of these two aspects as the build-in attribute of classification, and the follow-up Function Fitting of progress and flow estimate calculating.
For example, the storage format of classification and attribute can be：
(Key：Classification) (value：Sort sequence number+single user overlay capacity)
Next just can be first with basic data achieved above, the Function Fitting being segmented.
S102：According to each classification, the difference of Sorting distance is classified to classification in classification sequence, obtains preset number
Individual subsequence；Wherein, the Sorting distance is the difference between the sequence number of two classifications, in same subsequence, adjacent classification
Between sequence number difference it is equal；
Be the equal of according to certain interval from classification sequence specifically when classifying according to Sorting distance to classification
Extracted, the classification extracted constitutes a new sequence, due to being extracted from the classification sequence in step S101
Come, therefore, it can referred to as one subsequence.Likewise, when being extracted according to others interval, other sons can be obtained
Sequence.
For example, the classification sequence sorted from big to small according to the single user's overlay capacity of each classification is：Cat_1、Cat_2、
Cat_3、......、Cat_N.Then, specifically when classifying to classification, it is possible to sequence is extracted since the classification of serial number 1
Number at intervals of length classification constitute subsequence.
Such as length=1, then extract { Cat_1, Cat_2, Cat_3......Cat_N } and be used as a subsequence；
Length=2, then extract { Cat_1, Cat_3, Cat_5......Cat_m-2, Cat_m...... } and be used as one
Subsequence；
Length=k is then categorized as { Cat_1, Cat_k+1, Cat_2k+1...... } as a subsequence.
So, multiple subsequences can finally be produced.
Need exist for carrying out some following explanation：
First, for identical Sorting distance, if the starting classification of selection is different, the subsequence generated may
Can be different, for example, as length=2, if extracted since the classification of serial number 1, obtained subsequence is
Cat_1, Cat_3, Cat_5......Cat_m-2, Cat_m ..., wherein, m is odd number；But if being from serial number 2
Classification start extract, then the subsequence obtained for Cat_2, Cat_4, Cat_6 ..., Cat_n-2, Cat_
N ..., wherein, n is even number.Also, when length is bigger, the number for the different subsequences that can be produced is more.By
During follow-up Function Fitting, it is necessary to respectively obtain the fitting function under each Sorting distance, and under certain Sorting distance
Fitting function, be to be calculated to obtain according to all kinds of purpose user overlay capacity situations in the subsequence under the Sorting distance.Therefore,
A subsequence can be only extracted under amount of calculation, each Sorting distance in order to reduce, so, under each Sorting distance, it is only necessary to
Can be to calculate corresponding fitting function based on a subsequence.Wherein, the corresponding subsequence of each Sorting distance can
It is to be extracted since the classification of serial number 1, that is, the starting classification of each subsequence is classification sequence in basic data
The minimum classification of middle sequence number.Certainly, in the case where amount of calculation allows, it can also be risen respectively with different under same Sorting distance
Beginning classification extracts multiple subsequences, finally in digital simulation function, can equally be based respectively on each subsequence and be counted
Calculate, each subsequence under same Sorting distance is finally calculated to obtained value again and averaged.It is demonstrated experimentally that same Sorting distance
It is of substantially equal that each lower subsequence, which calculates obtained value, and this is further showed that, lower of same Sorting distance extracts one
Subsequence has reasonability, that is to say, that while more accurately can obtaining fitting function, also reduce amount of calculation.
Second, the classification in system is large number of, for example, typically there is thousands of or even up to ten thousand, in theory will, if classification
Quantity is N, then the Sorting distance between different classifications has N-1 kinds.But when reality is classified according to Sorting distance to classification, no
Need the fitting for all entering line function for all Sorting distances, generally, one maximum can be set for Sorting distance,
As long as carrying out Function Fitting to each Sorting distance below the maximum.For example, it is assumed that classification sum is 2000, most
Big Sorting distance can take 100, then carry out Function Fitting for each Sorting distance respectively.So, can further it reduce
Amount of calculation, and these fitting functions are general with regard to that can meet most demands in practical application.
S103：Classification in same subsequence is divided into multiple classification combinations so that each classification combination is than previous
The many classifications of classification combination, count the not duplicate customer overlay capacity of each classification combined accumulated, and calculate the combination of each classification
The user's overlay capacity growth rate combined relative to previous classification；
In same subsequence, (namely single user's overlay capacity is from big from small to large still according to sequence number for each classification
To small) order arrangement.Specific purpose user overlay capacity situations all kinds of according to same subsequence are under correspondence Sorting distance
Function when being fitted, the classification in same subsequence can be divided into multiple classifications first and combined.It is specific to divide class
, can be by classification on the basis of classification in subsequence (such as the minimum classification of sequence number), and according to classification in son when mesh is combined
Order in sequence adds a classification and constitutes next classification combination every time, so, and each classification combination is than previous classification
The all many classifications of combination.For example：
For subsequence：Cat_1, Cat_k+1, Cat_2k+1 ...), the combination of obtained classification can include：
{ Cat_1 }, { Cat_1, Cat_k+1 }, { Cat_1, Cat_k+1, Cat_2k+1 } ... ..
That is, first classification combination is made up of first classification in subsequence, second classification combination is just
The combination being made up of the first two classification in subsequence, the 3rd classification combination is exactly by first three classification group in subsequence
Into combination, that is, the combination of i-th classification is exactly the combination being made up of the preceding i classification in subsequence, by that analogy.
Obtain after multiple classification combinations, combined for the classification comprising two and two or more classification in same subsequence
For, the category information that can also be paid close attention to according to each user counted in step S101 counts classification combined accumulated
Not duplicate customer overlay capacity.Specifically, due to that can know each user respectively to which classification from the data being previously obtained
Attention rate it is higher, and the identification information such as the ID of each known user therefore, it can count same classification combination
Each interior classification distinguishes corresponding user's mark, so, and user's mark is collected, and removes user's mark of repetition, most
The user's identified number obtained eventually, it is possible to be defined as the not duplicate customer overlay capacity of the classification combined accumulated.
For example, it is still assumed that one has three users, respectively A, B, C, wherein：
The classification of user A concerns includes：Clothing, digital product class and toiletries；
The classification of user B concerns includes：Clothing, toiletries and the outdoor class of motion；
The classification of user C concerns includes：Toiletries and digital product class.
Assuming that certain classification is combined as { clothing, digital product class }, wherein, the user of clothing covering include user A and
User B (single user's overlay capacity is 2), the user of digital product class covering including user A and user C, (cover by single user
Lid amount is 2), now, the user that clothing and digital product class are covered is collected, and is removed after the user of repetition, is obtained
User include A, B, C, accordingly, it is possible to obtain the not duplicate customer overlay capacity of the classification combined accumulated be 3, wherein, user A
All occur in two classifications, but can only calculate once.
In a word, the user of accumulation can be counted in the manner described above for each classification combination in same subsequence
Overlay capacity.Afterwards, it is possible to calculate user overlay capacity of each classification combination relative to the combination of previous classification inside subsequence
Growth rate.If for example, to calculate classification combination the user's overlay capacity growth rate of { Cat_1, Cat_k+1 } relative to { Cat_1 },
It can then be calculated according to below equation (1)：
So, for same subsequence, it is possible to calculate multiple user's overlay capacity growth rates, such as：
PLength=k={ P1, P2, P3........} (2)
And then rule can be found from these growth rates, calculate the fitting function under correspondence Sorting distance.For it
He can also enter the fitting of line function in this manner by Sorting distance.
S104：Each user's overlay capacity growth rate and preset Function Format according to being obtained in same subsequence are carried out
Fitting, obtains the fitting function under correspondence Sorting distance；
Due to being sorted in same subsequence in the embodiment of the present application according to the single user's overlay capacity of each classification, because
This, a power function curve can be presented in the change for each user's overlay capacity growth rate in formula (2).So, entering
When line function is fitted, it is possible to which the function setup that will be simulated in advance is power-law scheme, different Sorting distances is corresponding to be intended
Closing function has identical form, when being fitted respectively to each Sorting distance, seeks to calculate in power function and is
Number and/or power exponent, that is to say, that for different Sorting distances, with different coefficient and/or power exponent, specifically
Value need determined according to each user's overlay capacity growth rate calculated in step S103.
For example, the form of power function can be as follows：
Wherein, length represents classification Sorting distance, αkRepresent the coefficient of fitting function, βkRepresent that the power of fitting function refers to
Number (for just), M is the sequence number that increased classification is combined in current classification combination relative to previous classification.The process of fitting is sought to
In the case where K takes a certain specific value, (α is calculatedk, βk) value, there is N number of different Sorting distance, it is possible to obtain N groups
(αk, βk) value.
So far, the process of Function Fitting terminates, the information such as keyword that can be just selected afterwards according to information dispensing side, right
Accumulation user's overlay capacity of classification combination is estimated.
S105：After information dispensing side have selected target keyword, the target keyword is mapped as many in system
Individual target classification, each target class purpose sequence number and single user's overlay capacity are determined according to the basic data；
When implementing, after information dispensing side have selected target keyword, system keyword can be just mapped as be
Target classification in system, and it is usually multiple.For example, the keyword of information dispensing side selection is " comfortable ventilating ", " arrived in time
Account ", " pure cotton " and " cost performance ", the 4 target classifications obtained after mapping are respectively " sport footwear ", " rechargeable card ", " clothes footwear
Cap ", " electronic product ".Next the accumulation user overlay capacity of the classification combination of whole target classification compositions can just be carried out
Estimate.Specifically when being estimated, the basic data got in step S101 still can be used, while can also use step
Fitting function under each Sorting distance obtained in S104.
S106：Using the minimum target classification of sequence number as target fiducials classification, by the target fiducials classification respectively with other
Each target classification constitutes objective cross, and calculates two target class purpose Sorting distances in each objective cross；
For example, the keyword selected according to information dispensing side, the target classification mapped out includes：Tetra- classifications of A, B, C, D,
In the classification sequence that step S101 is obtained, this corresponding sequence number of four target classifications and overlay capacity are respectively A (2, Num_A), B
(6, Num_B), C (10, Num_C), D (20, Num_D).Wherein, because the sequence of A classifications is most forward, therefore, using A classifications as base
Quasi- classification, then respectively obtains { A, B }, { A, C }, { A, D } such three objective cross, and each objective cross includes two mesh
Classification is marked, one of them is target fiducials classification, while two target class purpose sequences in each objective cross can be calculated
Distance.Wherein：
LengthA、B=6-2=4
LengthA、C=10-2=8
LengthA、D=20-2=18
S107：The corresponding fitting function of Sorting distance of each objective cross is utilized respectively, each objective cross phase is estimated
For accumulation user's increment of target fiducials classification, and by the single user's overlay capacity of target fiducials classification and each target group
The accumulation user increment closed is added, and estimates out accumulation user's overlay capacity when all target classifications are combined；
After multiple objective cross, and each self-corresponding Sorting distance are obtained, it is possible to be utilized respectively correspondence sequence away from
From fitting function, to estimate accumulation user increment of each objective cross relative to target fiducials classification.That is, can divide
Do not obtain increased non-repetitive user overlay capacity N1 of the objective cross { A, B } relative to { A }, objective cross { A, C } relative to
Increased non-repetitive user's overlay capacity N2 of { A }, and increase of the objective cross { A, D } relative to { A } can be respectively obtained
Non-repetitive user's overlay capacity N3, so, as long as the single user's overlay capacities of classification A are added with N1, N2, N3, it is possible to
Obtain the not duplicate customer overlay capacity of { A, B, C, D } this classification combined accumulated.
Wherein, because fitting function is the function on accumulating user's growth rate, therefore, specifically according to fitting function meter
Calculate certain objective cross relative to target fiducials classification accumulation user's increment when, can follow the steps below：
First, by another target in the corresponding Sorting distance of objective cross and objective cross outside target fiducials classification
The sequence number of classification is (for example, for objective cross { A, B }, classification A is benchmark classification, therefore the sequence number just refers to classification B sequence
Number), it is brought into the fitting function of the Sorting distance, the accumulation user for obtaining the objective cross relative to target fiducials classification increases
Long rate.For example, fitting function during { A, B } correspondence length=4, fitting function during { A, C } correspondence length=8, { A, D }
Correspondence length=18 fitting function, therefore accumulation user growth rate of the objective cross { A, B } relative to { A } can be by length
Fitting function when=4 is obtained, that is, by M=6, in fitting function when being brought into length=4：
Wherein, rint () is to round up.Due to having drawn (α in step S1044, β4) value, therefore, so that it may
To calculate a definite numeral, the numeral just represents accumulation user growth rate of the objective cross { A, B } relative to { A }.
It is how many equivalent to the growth ratio known relative to { A } after above-mentioned growth rate is obtained, if to calculate
Accumulation user increment of the objective cross { A, B } relative to { A }, also needs to count in theory the accumulation user covering of { A, B }
Amount.But in the embodiment of the present application, in order to reduce amount of calculation, before making full use of during training obtains fitting function
Statistical value through obtaining, the corresponding subsequences of { A, B } corresponding Sorting distance length=4 can be found first, if wherein
Just classification A, B are included, then equivalent to accumulation user's overlay capacity that { A, B } was counted during being trained, therefore, directly
The user's overlay capacity for connecing use { A } adds accumulation user's overlay capacity of { A, B }, multiplied by { A, the B } obtained with calculating relative to { A }
Accumulate user's growth rate, it is possible to obtain the accumulation user's increment of { A, B } relative to { A }.
But as it was noted above, in the embodiment of the present application, a subsequence may have only been extracted under same Sorting distance,
Therefore, it is more likely that occurring not including A, B situation in subsequence.For example, the corresponding subsequences of Sorting distance length=4 are：
accuNumj=Cat_1, Cat_5, Cat_9 ...
And A is Cat_2, B is Cat_6, therefore, is not appeared in the subsequence, at this point it is possible to from the subsequence
In, obtain and target fiducials category number (i.e. classification A sequence number 2) immediate first classification (corresponding Cat_ in this example embodiment
1), and with another kind of purpose sequence number in objective cross (i.e. classification B sequence number 6) immediate second classification is (in this example embodiment
Correspondence Cat_5), then, use is not repeated according to what the single user's overlay capacity of the first classification, the first classification and the second classification were accumulated
Family overlay capacity, and the objective cross estimate out the objective cross phase relative to accumulation user's growth rate of target fiducials classification
For accumulation user's increment of target fiducials classification.That is, calculating user's overlay capacity and { Cat_1, Cat_ of { Cat_1 }
5 } accumulation user's overlay capacity sum, Y when being then multiplied by M=6Length=4Value, you can draw objective cross { A, B } than classification A
The discreet value of increased non-repetitive user's overlay capacity.
Using same method, objective cross { A, C } non-repetitive user more increased than classification A can also be calculated respectively
The discreet value of overlay capacity, and objective cross { A, D } non-repetitive user's overlay capacity more increased than classification A discreet value.Finally
The pre- of the single user's overlay capacities of classification A and each objective cross non-repetitive user's overlay capacity more increased than classification A is calculated again
Valuation sum, it is possible to obtain the not duplicate customer overlay capacity of { A, B, C, D } this classification combined accumulated finally estimated out.
Summary calculating process, can be with abstract for below equation：
Wherein, n is target class purpose quantity, and OrderNumber is another outside target fiducials classification in each objective cross
One target class purpose sequence number.
S108：When receiving the request for obtaining accumulation user's overlay capacity, the accumulation user's overlay capacity estimated is supplied to
Information dispensing side.
After the not duplicate customer overlay capacity of { A, B, C, D } this classification combined accumulated is obtained, if information dispensing side
Need, then can be pointed out in interface.For example, when implementing, the interface of keyword can be selected in information dispensing side
The upper button that printed words such as " obtain user overlay capacity " are provided, after have selected keyword in information dispensing side, if clicking this
Button, it is possible to shown the not duplicate customer overlay capacity for the accumulation estimated out on the surface.So, information dispensing side
It can judge whether user's overlay capacity meets the demand of oneself accordingly, if it is satisfied, then can be thrown according to the keyword
Put, otherwise, other keywords can also be reselected.
In a word, by the embodiment of the present application, using the classification in the subsequence of limited quantity as representative, some classifications are calculated
The not duplicate customer overlay capacity of combined accumulated, and corresponding fitting function under each Sorting distance can also be obtained, to intend
Close under correspondence Sorting distance, certain classification combines the accumulation user's growth rate combined relative to previous classification, and then just can basis
Accumulation user's overlay capacity of each classification combination, estimates out and is thrown by information in fitting function and subsequence under each Sorting distance
Accumulation user's overlay capacity when the target classification of the side's of putting selection is combined, to be carried accordingly to information dispensing side
Show.It can be seen that, by that with upper type, amount of calculation can be narrowed down in the range of the corresponding subsequence of Sorting distance of limited quantity,
Amount of calculation is limited in the range of computing system realizes.Meanwhile, covered using the not duplicate customer of classification combined accumulated
Lid is measured into the fitting of line function and estimating subsequently, therefore, solves same user in asking that inhomogeneity repeats now
Topic.
It should be noted that in the embodiment of the present application, the executive agent of each step described in Fig. 1 can be certain business platform
Server, wherein, obtain basic data and the process that is fitted to the function under each Sorting distance, can online under it is complete
Into.
Reminding method with user's overlay capacity information that the embodiment of the present application is provided is corresponding, and the embodiment of the present application is also provided
A kind of suggestion device of user's overlay capacity information, referring to Fig. 2, the device can include：
Basic data acquiring unit 201, for the category information paid close attention to respectively according to each user got in advance, is obtained
Take basic data；The basic data include the single user's overlay capacity of each classification, according to single user's overlay capacity by
The classification sequence obtained after small sequence, and sequence number of each classification in classification sequence are arrived greatly；
Classification taxon 202, for the difference of Sorting distance to be divided classification in classification sequence according to each classification
Class, obtains preset number subsequence；Wherein, the Sorting distance is the difference between the sequence number of two classifications, in same son
In sequence, the sequence number difference between adjacent classification is equal；
Growth Rate Calculation unit 203, is combined so that each for the classification in same subsequence to be divided into multiple classifications
Classification combination combines many classifications, the classification paid close attention to respectively according to described each user got in advance than previous classification
Information, counts the not duplicate customer overlay capacity of each classification combined accumulated, and calculates the combination of each classification relative to previous class
User's overlay capacity growth rate of mesh combination；
Fitting unit 204, for according to each user's overlay capacity growth rate for being obtained in same subsequence and preset
Function Format is fitted, and obtains the fitting function under correspondence Sorting distance；
The corresponding unit of the above 201 to 204 is the unit needed for training process.
Target classification determining unit 205, for after information dispensing side have selected target keyword, by the target critical
Word is mapped as multiple target classifications in system, determines each target class purpose sequence number according to the basic data and individually uses
Family overlay capacity；
Objective cross determining unit 206, for using the minimum target classification of sequence number as target fiducials classification, by the target
Benchmark classification constitutes objective cross with other each target classifications respectively, and calculates two target class purposes rows in each objective cross
Sequence distance；
Unit 207 is estimated, the corresponding fitting function of the Sorting distance for being utilized respectively each objective cross estimates each
Objective cross relative to target fiducials classification accumulation user's increment, and by the single user's overlay capacity of target fiducials classification with
The accumulation user increment of each objective cross is added, and is estimated out accumulation when all target classifications are combined and is used
Family overlay capacity；
The corresponding each unit of the above 205 to 207 is the unit needed for process of estimating.
Tip element 208, for when receiving the request for obtaining accumulation user's overlay capacity, the accumulation user estimated to be covered
Lid amount is supplied to information dispensing side.
Wherein, due in each subsequence, being sorted according to single user's overlay capacity to classification, therefore, often
Change of the individual classification combination relative to user's covering growth rate that previous classification is combined is rendered as power function curve, therefore, preset
Function Format can be power-law scheme, the power-law scheme includes coefficient to be determined and/or power exponent, therefore,
Fitting unit 204 specifically can be used for：
According to each user's overlay capacity growth rate obtained in same classification subsequence and preset power-law scheme, really
Make coefficient and/or power exponent under correspondence Sorting distance；The coefficient and/or power exponent and the Sorting distance are brought into pre-
In the Function Format put, the fitting function under correspondence Sorting distance is obtained.
A subsequence can be only taken under amount of calculation, same Sorting distance in order to further reduce.Also, each sequence away from
Subsequence under is using the minimum classification of sequence number in the classification sequence as starting classification.
When implementing, estimating unit 207 can include：
Growth Rate Calculation subelement, for by target base in the corresponding Sorting distance of the objective cross and objective cross
Another target class purpose sequence number outside quasi- classification, is brought into the fitting function of the Sorting distance, obtains the objective cross phase
For accumulation user's growth rate of target fiducials classification；
Increment computation subunit, for from the corresponding subsequence of the Sorting distance, obtaining and target fiducials classification sequence
Number immediate first classification, and with immediate second classification of another kind of purpose sequence number in objective cross, according to described
The single user's overlay capacity of one classification, and the first classification and the not duplicate customer overlay capacity of the second classification accumulation, and it is described
The objective cross estimates out the objective cross relative to target fiducials class relative to accumulation user's growth rate of target fiducials classification
Purpose accumulates user's increment.
Wherein, when classifying according to the difference of Sorting distance to classification, between maximum Sorting distance and classification sum
Ratio be less than preset threshold value.
When implementing, basic data acquiring unit 201 specifically can be used for：
Previously according to historical operation behavior record of the user within preset time period, user's attention rate highest is obtained pre-
Put the category information of number；
The information got at each user is collected, each classification is counted and distinguishes corresponding number of users,
The number of users is defined as the single user's overlay capacity of each classification.
When carrying out Function Fitting, the fitting unit 204 covers in the not duplicate customer for counting each classification combined accumulated
During lid amount, specifically it can be used for：
Count each classification in same classification combination and distinguish corresponding user's mark；
User mark is collected, and removes user's mark of repetition, by final user's identified number, it is determined that
For the not duplicate customer overlay capacity of the classification combined accumulated.
In a word, by the embodiment of the present application, using the classification in the subsequence of limited quantity as representative, some classifications are calculated
The not duplicate customer overlay capacity of combined accumulated, and corresponding fitting function under each Sorting distance can also be obtained, to intend
Close under correspondence Sorting distance, certain classification combines the accumulation user's growth rate combined relative to previous classification, and then just can basis
Accumulation user's overlay capacity of each classification combination, estimates out and is thrown by information in fitting function and subsequence under each Sorting distance
Accumulation user's overlay capacity when the target classification of the side's of putting selection is combined, to be carried accordingly to information dispensing side
Show.It can be seen that, by that with upper type, amount of calculation can be narrowed down in the range of the corresponding subsequence of Sorting distance of limited quantity,
Amount of calculation is limited in the range of computing system realizes.Meanwhile, covered using the not duplicate customer of classification combined accumulated
Lid is measured into the fitting of line function and estimating subsequently, therefore, solves same user in asking that inhomogeneity repeats now
Topic.
As seen through the above description of the embodiments, those skilled in the art can be understood that the application can
Realized by the mode of software plus required general hardware platform.Understood based on such, the technical scheme essence of the application
On the part that is contributed in other words to prior art can be embodied in the form of software product, the computer software product
It can be stored in storage medium, such as ROM/RAM, magnetic disc, CD, including some instructions are to cause a computer equipment
(can be personal computer, server, or network equipment etc.) performs some of each embodiment of the application or embodiment
Method described in part.
Each embodiment in this specification is described by the way of progressive, identical similar portion between each embodiment
Divide mutually referring to what each embodiment was stressed is the difference with other embodiment.Especially for system or
For system embodiment, because it is substantially similar to embodiment of the method, so describing fairly simple, related part is referring to method
The part explanation of embodiment.System and system embodiment described above is only schematical, wherein the conduct
The unit that separating component illustrates can be or may not be it is physically separate, the part shown as unit can be or
Person may not be physical location, you can with positioned at a place, or can also be distributed on multiple NEs.Can root
Some or all of module therein is factually selected to realize the purpose of this embodiment scheme the need for border.Ordinary skill
Personnel are without creative efforts, you can to understand and implement.
Above to the suggestion device of user's overlay capacity information provided herein, it is described in detail, herein should
The principle and embodiment of the application are set forth with specific case, the explanation of above example is only intended to help and managed
Solve the present processes and its core concept；Simultaneously for those of ordinary skill in the art, according to the thought of the application,
It will change in embodiment and application.In summary, this specification content should not be construed as to this Shen
Limitation please.
Claims (10)
1. a kind of reminding method of user's overlay capacity information, it is characterised in that including：
The category information that each user according to getting in advance pays close attention to respectively, obtains basic data；Wrapped in the basic data
Include the single user's overlay capacity of each classification, according to the classification sequence obtained after the single descending sequence of user's overlay capacity,
And sequence number of each classification in classification sequence；
According to each classification, the difference of Sorting distance is classified to classification in classification sequence, obtains preset number subsequence；
Wherein, the Sorting distance is the difference between the sequence number of two classifications；
Classification in same subsequence is divided into multiple classification combinations so that each classification combination combines many than previous classification
One classification, the category information paid close attention to respectively according to described each user got in advance, counts each classification combined accumulated
Not duplicate customer overlay capacity, and calculate each classification combination relative to previous classification combine user's overlay capacity growth rate；
It is fitted, is obtained according to each user's overlay capacity growth rate obtained in same subsequence and preset Function Format
Fitting function under correspondence Sorting distance；
After information dispensing side have selected target keyword, the target keyword is mapped as to multiple target class in system
Mesh, each target class purpose sequence number and single user's overlay capacity are determined according to the basic data；
The selection target benchmark classification in the obtained multiple target classifications of mapping, by the target fiducials classification respectively with other each mesh
Classification composition objective cross is marked, and calculates two target class purpose Sorting distances in each objective cross；
The corresponding fitting function of Sorting distance of each objective cross is utilized respectively, each objective cross is estimated relative to target base
Accumulation user's increment of quasi- classification, and the single user's overlay capacity of target fiducials classification and the described of each objective cross are tired out
Product user increment is added, and estimates out accumulation user's overlay capacity when all target classifications are combined；
When receiving the request for obtaining accumulation user's overlay capacity, it is supplied to information to deliver the accumulation user's overlay capacity estimated
Side.
2. according to the method described in claim 1, it is characterised in that the preset Function Format is power-law scheme, described
Power-law scheme includes coefficient to be determined and/or power exponent, described to be covered according to each user obtained in same subsequence
Lid amount growth rate and preset Function Format are fitted, and obtain the fitting function under correspondence Sorting distance, including：
According to each user's overlay capacity growth rate obtained in same subsequence and preset power-law scheme, correspondence is determined
Coefficient and/or power exponent under Sorting distance；
The coefficient and/or power exponent and the Sorting distance are brought into preset Function Format, obtain correspondence sequence away from
Fitting function under.
3. according to the method described in claim 1, it is characterised in that a subsequence is taken under same Sorting distance.
4. method according to claim 3, it is characterised in that the subsequence under each Sorting distance is with the classification sequence
The minimum classification of sequence number is starting classification in row.
5. the method according to any one of Claims 1-4, it is characterised in that each objective cross that is utilized respectively
The corresponding fitting function of Sorting distance, estimates accumulation user increment of each objective cross relative to target fiducials classification, bag
Include：
By another target classification in the corresponding Sorting distance of the objective cross and objective cross outside target fiducials classification
Sequence number, be brought into the fitting function of the Sorting distance, obtain the objective cross relative to target fiducials classification accumulation use
Family growth rate；
From the corresponding subsequence of the Sorting distance, obtain and immediate first classification of target fiducials category number, Yi Jiyu
Another kind of immediate second classification of purpose sequence number in objective cross, according to the single user's overlay capacity of first classification, with
And first classification and the accumulation of the second classification not duplicate customer overlay capacity, and the objective cross is relative to target fiducials classification
Accumulation user's growth rate, estimate out accumulation user increment of the objective cross relative to target fiducials classification.
6. the method according to any one of Claims 1-4, it is characterised in that in the difference according to Sorting distance to classification
When being classified, the ratio between maximum Sorting distance and classification sum is less than preset threshold value.
7. the method according to any one of Claims 1-4, it is characterised in that according to each user got in advance point
The category information do not paid close attention to, obtains basic data, including：
Previously according to historical operation behavior record of the user within preset time period, user's attention rate highest preset number is obtained
Purpose category information；
The information got at each user is collected, each classification is counted and distinguishes corresponding number of users, by this
Number of users is defined as the single user's overlay capacity of each classification.
8. method according to claim 7, it is characterised in that when carrying out Function Fitting, each classification group of the statistics
The not duplicate customer overlay capacity of accumulation is closed, including：
Count each classification in same classification combination and distinguish corresponding user's mark；
User mark is collected, and removes user's mark of repetition, by final user's identified number, is defined as this
The not duplicate customer overlay capacity of classification combined accumulated.
9. a kind of suggestion device of user's overlay capacity information, it is characterised in that including：
Basic data acquiring unit, for the category information paid close attention to respectively according to each user got in advance, obtains basis
Data；The basic data includes the single user's overlay capacity of each classification, descending according to single user's overlay capacity
The classification sequence obtained after sequence, and sequence number of each classification in classification sequence；
Classification taxon, for according to each classification, the difference of Sorting distance to be classified to classification in classification sequence, is obtained
Preset number subsequence；Wherein, the Sorting distance is the difference between the sequence number of two classifications, in same subsequence,
Sequence number difference between adjacent classification is equal；
Growth Rate Calculation unit, is combined for the classification in same subsequence to be divided into multiple classifications so that each classification group
The previous many classifications of classification combination of composition and division in a proportion, the category information paid close attention to respectively according to described each user got in advance,
The not duplicate customer overlay capacity of each classification combined accumulated is counted, and calculates the combination of each classification relative to the combination of previous classification
User's overlay capacity growth rate；
Fitting unit, for according to each user's overlay capacity growth rate obtained in same subsequence and preset Function Format
It is fitted, obtains the fitting function under correspondence Sorting distance；
Target classification determining unit, for after information dispensing side have selected target keyword, the target keyword to be mapped
For multiple target classifications in system, determine that each target class purpose sequence number and single user cover according to the basic data
Amount；
Objective cross determining unit, for using the minimum target classification of sequence number as target fiducials classification, by the target fiducials class
Mesh respectively with other each target classifications composition objective cross, and calculate in each objective cross two target class purposes sequences away from
From；
Unit is estimated, the corresponding fitting function of the Sorting distance for being utilized respectively each objective cross estimates each target group
Accumulation user's increment relative to target fiducials classification is closed, and by the single user's overlay capacity of target fiducials classification and each mesh
The accumulation user increment of mark combination is added, and estimates out accumulation user covering when all target classifications are combined
Amount；
Tip element, for when receiving the request for obtaining accumulation user's overlay capacity, the accumulation user's overlay capacity estimated to be carried
Supply information dispensing side.
10. device according to claim 9, it is characterised in that the unit of estimating includes：
Growth Rate Calculation subelement, for by target fiducials class in the corresponding Sorting distance of the objective cross and objective cross
Another target class purpose sequence number outside mesh, is brought into the fitting function of the Sorting distance, obtain the objective cross relative to
Accumulation user's growth rate of target fiducials classification；
Increment computation subunit, for from the corresponding subsequence of the Sorting distance, obtaining with target fiducials category number most
The first close classification, and with immediate second classification of another kind of purpose sequence number in objective cross, according to the first kind
The single user's overlay capacity of mesh, and the first classification and the not duplicate customer overlay capacity of the second classification accumulation, and the target
Accumulation user's growth rate relative to target fiducials classification is combined, the objective cross is estimated out tired relative to target fiducials classification
Product user's increment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310279704.2A CN104281594B (en) | 2013-07-04 | 2013-07-04 | The reminding method and device of user's overlay capacity information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310279704.2A CN104281594B (en) | 2013-07-04 | 2013-07-04 | The reminding method and device of user's overlay capacity information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104281594A CN104281594A (en) | 2015-01-14 |
| CN104281594B true CN104281594B (en) | 2017-09-01 |
Family
ID=52256473
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310279704.2A Active CN104281594B (en) | 2013-07-04 | 2013-07-04 | The reminding method and device of user's overlay capacity information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104281594B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102682005A (en) * | 2011-03-10 | 2012-09-19 | 阿里巴巴集团控股有限公司 | Method and device for determining preference categories |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7603348B2 (en) * | 2007-01-26 | 2009-10-13 | Yahoo! Inc. | System for classifying a search query |
-
2013
- 2013-07-04 CN CN201310279704.2A patent/CN104281594B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102682005A (en) * | 2011-03-10 | 2012-09-19 | 阿里巴巴集团控股有限公司 | Method and device for determining preference categories |
Non-Patent Citations (1)
| Title |
|---|
| 基于用户行为挖掘的数据流管理技术研究;李军;《中国博士学位论文全文数据库信息科技辑》;20130215(第2期);第I139-13页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104281594A (en) | 2015-01-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Shinde et al. | Hybrid personalized recommender system using centering-bunching based clustering algorithm | |
| CN105701216B (en) | A kind of information-pushing method and device | |
| CN107330050A (en) | A kind of personalized recommendation method and system | |
| CN106997358A (en) | Information recommendation method and device | |
| CN105723402A (en) | Systems and methods for determining influencers in a social data network | |
| CN103455522B (en) | Recommendation method and system of application extension tools | |
| CN108170692A (en) | A kind of focus incident information processing method and device | |
| CN108256119A (en) | A kind of construction method of resource recommendation model and the resource recommendation method based on the model | |
| CN109064293A (en) | Method of Commodity Recommendation, device, computer equipment and storage medium | |
| CN103729360A (en) | Interest label recommendation method and system | |
| CN107526810B (en) | Method and device for establishing click rate estimation model and display method and device | |
| CN101984437A (en) | Music resource individual recommendation method and system thereof | |
| CN106326391A (en) | Method and device for recommending multimedia resources | |
| CN105260414B (en) | User behavior similarity calculation method and device | |
| CN105654198B (en) | Brand advertisement effect optimization method capable of realizing optimal threshold value selection | |
| CN104376083B (en) | It is a kind of that method is recommended based on concern relation and the figure of multi-user's behavior | |
| CN104077417A (en) | Figure tag recommendation method and system in social network | |
| CN110197404A (en) | The personalized long-tail Method of Commodity Recommendation and system of popularity deviation can be reduced | |
| CN109977296A (en) | A kind of information-pushing method, device, equipment and storage medium | |
| CN108764949A (en) | A kind of information-pushing method and equipment | |
| CN106354867A (en) | Multimedia resource recommendation method and device | |
| CN106919995A (en) | A kind of method and device for judging user group's loss orientation | |
| CN107645667A (en) | Video recommendation method, system and server apparatus | |
| CN106156351A (en) | Multimedia resource recommendation information generates method and device | |
| CN114756764A (en) | Enterprise-based content information stream recommendation method and device, electronic equipment and storage medium |
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 |