CN104281594A - User coverage information prompting method and device - Google Patents

Abstract

The invention discloses a user coverage information prompting method and device. The method includes the steps of obtaining basic data, classifying categories according to different sequencing distances of the categories in a category sequence, dividing the categories in each sub-sequence into multiple category sets, calculating the user coverage increase rate of each category set relative to a prior category set, obtaining a fitting function under the corresponding sequencing distance according to the user coverage increase rate in each sub-sequence and a preset function format, mapping target keywords into multiple target categories in a system after an information issuing party selects the target keywords, calculating the sequencing distance of each target set, and estimating the accumulated user coverage of each target category set through the corresponding fitting function so that the accumulated user coverage can be provided for the information issuing party. By means of the user coverage information prompting method and device, the calculation amount can be reduced, and the problem that the same user repeatedly appears in different categories is solved.

Description

The reminding method of user's overlay capacity information and device

Technical field

The application relates to the information prompting technical field in the directed impression information process of point of interest, particularly relates to reminding method and the device of user's overlay capacity information.

Background technology

In some internet service platform, often have information to throw in the webpage of direction business platform and throw in some customizing messages, while being paid close attention to by user to this webpage, these customizing messages also can be concerned about, even clicked, the page importing to information input side oneself brings flow for it.At first, the customizing messages thrown in same webpage is generally fixing, but for large-scale business platform, its user (viewer) One's name is legion, for the different access user of same webpage, the point that user pays close attention to may be different.Such as, for certain E-commerce transaction platform, its business object provided can be divided into multiple classification from multiple dimension, as clothing, digital product class etc., some user may be interested in clothing information, and other users may logarithmic code products interested etc.Now, if throw in fixing customizing messages in same webpage (such as certain website homepage), then mean to only have certain customers can be interested in this information, for other users, be equivalent to the space of a whole page wasting this customizing messages place.

For this reason, the information proposing " point of interest directed " in some systems throws in mode, also by the oriented approach that the keyword of the information side of inputing selection and the possible point of interest of user match.Information input side can obtain the keyword of system recommendation by inputting the modes such as the descriptor of its customizing messages, can also screen these keywords, according to the selection result, keyword can be mapped to intrasystem classification by system, the keyword that information input side is selected is " cost performance is high ", then " cost performance is high " can be mapped as classification " digital product " by system, if and the keyword selected is " comfortable feel ", then " comfortable feel " can be mapped as classification " household articles " by system, Deng, like this can the point of interest of awareness information input side, this point of interest can represent with the classification that each selected keyword is corresponding.

Simultaneity factor will analyze the current browse webpage content of each information browse user and history focus, obtain the focus of each user, and this focus can represent by the classification in custom system equally.Such as, the focus of user's first comprises " clothes ", " digital product " etc.Like this, just the focus of the point of interest of information input side and user can be matched, the user crowd that the customizing messages of information input side matches is thrown in.That is, for same webpage, for different user ID, its customizing messages that can see may be different, but focus that is all basic and user matches, therefore, the space of a whole page that netpage user represents customizing messages can be made full use of, make same webpage can bring customer flow for different information input sides simultaneously.

In the information input mode of above-mentioned this point of interest orientation, after information input side have selected certain or some keywords, system can also estimate corresponding classification combination can great user's overlay capacity, and be shown to information input side, and then information input side can determine whether select these keywords to throw in accordingly.But for large-scale business platform, the classification quantity of the business object wherein comprised is generally very many, if the user's overlay capacity wanting accurate various possible classification combination corresponding, need huge calculated amount.Such as hypothesis has N number of classification, then need the user's overlay capacity calculating 2^N classification combination, for the N of thousands of magnitude even up to ten thousand, even if this calculated amount is all difficult to realize for existing large-scale distributed computing system.

In addition, when calculating user's overlay capacity of classification combination, simply overlay capacity corresponding separately for each classification can't be added, because same user may pay close attention to different classifications, namely such as user A is electrical type user, is again mother and baby's group of subscribers, again sports fan, if the interest classification that information input side is selected comprises this three kinds of classifications, that user A can only calculate a user, and can not calculate three times.That is flow estimates the problem that also will solve same user and repeat in different classifications.

In a word, the technical matters solved in the urgent need to those skilled in the art is just: how carrying out, in the process that flow estimates, reducing calculated amount, and solves the problem that same user repeats in different classifications.

Summary of the invention

This application provides reminding method and the device of user's overlay capacity information, can calculated amount be reduced, and solve the problem that same user repeats in different classifications.

This application provides following scheme:

A reminding method for user's overlay capacity information, comprising:

According to the category information that each user got in advance pays close attention to respectively, obtain basic data; Described basic data comprises the independent user's overlay capacity of each classification, according to the classification sequence obtained after the independent descending sequence of user's overlay capacity, and the sequence number of each classification in classification sequence;

According to the difference of each classification Sorting distance in classification sequence, classification is classified, obtain a preset number subsequence; Wherein, described Sorting distance is the difference between the sequence number of two classifications, and in same subsequence, the sequence number difference between adjacent classification is all equal;

Classification in same subsequence is divided into the combination of multiple classification, each classification is made to combine than the many classifications of previous classification combination, according to the category information that described each user got in advance pays close attention to respectively, add up the not duplicate customer overlay capacity of each classification combined accumulated, and calculate each classification and combine the user's overlay capacity rate of growth combined relative to previous classification;

Carry out matching according to each user's overlay capacity rate of growth obtained in same subsequence and preset Function Format, obtain the fitting function under corresponding Sorting distance;

After information input side have selected target keyword, described target keyword is mapped as the multiple target classifications in system, determines each target class object sequence number and independent user's overlay capacity according to described basic data;

Using target classification minimum for sequence number as target fiducials classification, by this target fiducials classification, each target classification forms objective cross with other respectively, and calculates two target class object Sorting distances in each objective cross;

Utilize the fitting function that the Sorting distance of each objective cross is corresponding respectively, estimate the accumulation user increment of each objective cross relative to target fiducials classification, and user's overlay capacity independent for target fiducials classification is added with described user's increment of accumulating of each objective cross, estimate out accumulation user overlay capacity when being combined by all target classifications;

When receiving the request obtaining accumulation user overlay capacity, the accumulation user overlay capacity estimated is supplied to information input side.

A suggestion device for user's overlay capacity information, comprising:

Basic data acquiring unit, for the category information paid close attention to respectively according to each user got in advance, obtains basic data; Described basic data comprises the independent user's overlay capacity of each classification, according to the classification sequence obtained after the independent descending sequence of user's overlay capacity, and the sequence number of each classification in classification sequence;

Classification taxon, for classifying to classification according to the difference of each classification Sorting distance in classification sequence, obtains a preset number subsequence; Wherein, described Sorting distance is the difference between the sequence number of two classifications, and in same subsequence, the sequence number difference between adjacent classification is all equal;

Growth Rate Calculation unit, for the classification in same subsequence being divided into the combination of multiple classification, each classification is made to combine than the many classifications of previous classification combination, according to the category information that described each user got in advance pays close attention to respectively, add up the not duplicate customer overlay capacity of each classification combined accumulated, and calculate each classification and combine the user's overlay capacity rate of growth combined relative to previous classification;

Fitting unit, for carrying out matching according to each user's overlay capacity rate of growth obtained in same subsequence and preset Function Format, obtains the fitting function under corresponding Sorting distance;

Target classification determining unit, for after information input side have selected target keyword, is mapped as the multiple target classifications in system, determines each target class object sequence number and independent user's overlay capacity according to described basic data by described target keyword;

Objective cross determining unit, for using target classification minimum for sequence number as target fiducials classification, by this target fiducials classification, each target classification forms objective cross with other respectively, and calculates two target class object Sorting distances in each objective cross;

Estimate unit, for utilizing the fitting function that the Sorting distance of each objective cross is corresponding respectively, estimate the accumulation user increment of each objective cross relative to target fiducials classification, and user's overlay capacity independent for target fiducials classification is added with described user's increment of accumulating of each objective cross, estimate out accumulation user overlay capacity when being combined by all target classifications;

Tip element, for when receiving the request obtaining accumulation user overlay capacity, is supplied to information input side by the accumulation user overlay capacity estimated.

According to the specific embodiment that the application provides, this application discloses following technique effect:

Pass through the embodiment of the present application, with the classification in the subsequence of limited quantity for representative, calculate the not duplicate customer overlay capacity of some classification combined accumulated, and fitting function corresponding under each Sorting distance can be obtained, in order under the corresponding Sorting distance of matching, certain classification combines the accumulation user rate of growth combined relative to last classification, and then just can according to the accumulation user overlay capacity of classification combination each in the fitting function under each Sorting distance and subsequence, estimate out accumulation user overlay capacity when being combined by the target classification that information input side is selected, so that input side points out accordingly to information.Visible, by with upper type, calculated amount can be narrowed down within the scope of subsequence corresponding to the Sorting distance of limited quantity, in calculated amount is limited in scope that computing system can realize.Meanwhile, utilize the not duplicate customer overlay capacity of classification combined accumulated to carry out the matching of function and follow-up estimating, therefore, solve the problem that same user repeats now in inhomogeneity.

Certainly, the arbitrary product implementing the application might not need to reach above-described all advantages simultaneously.

Accompanying drawing explanation

In order to be illustrated more clearly in the embodiment of the present application or technical scheme of the prior art, be briefly described to the accompanying drawing used required in embodiment below, apparently, accompanying drawing in the following describes is only some embodiments of the application, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.

Fig. 1 is the process flow diagram of the method that the embodiment of the present application provides;

Fig. 2 is the schematic diagram of the device that the embodiment of the present application provides.

Embodiment

Below in conjunction with the accompanying drawing in the embodiment of the present application, be clearly and completely described the technical scheme in the embodiment of the present application, obviously, described embodiment is only some embodiments of the present application, instead of whole embodiments.Based on the embodiment in the application, the every other embodiment that those of ordinary skill in the art obtain, all belongs to the scope of the application's protection.

In the embodiment of the present application, in order to can when estimating user's overlay capacity, reduce calculated amount, and solve the problem repeated in the different classifications that same user combines at same classification, have employed the mode of user's overlay capacity of classification combination being carried out to piecewise fitting, carry out the accumulation user overlay capacity to classification combination by fitting function, and be supplied to information input side.Below this is introduced in detail.

See Fig. 1, the embodiment of the present application provides a kind of reminding method of user's overlay capacity information, and the method can comprise the following steps:

S101: the category information paid close attention to respectively according to each user got in advance, obtains basic data; Described basic data comprises the independent user's overlay capacity of each classification, according to the classification sequence obtained after the independent descending sequence of user's overlay capacity, and the sequence number of each classification in classification sequence;

Wherein, when obtaining the category information that each user pays close attention to respectively, can add up the historical operation behavior record of the user collected in preset time period, then counting each user interested in which classification respectively.Such as, for certain user A, find according to its historical operation behavior record, the browsed webpage majority of this user A is all the webpage at business object place of clothing, digital product class, simultaneously, also may find, in the business object that user is browsed, mostly the business object finally producing purchase or the behavior of reservation is also clothing and digital product class, therefore, these information comprehensive, just can get several classifications that this user A attention rate is the highest.Other users also carry out similar process respectively, like this, finally just can obtain each user interested in which classification respectively.Wherein, specifically from the historical operation behavior record of user, obtain the implementation method of user to all kinds of object attention rate, see the realization in prior art, no longer can describe in detail here.Certainly, for each user, its classification paid close attention to may have a lot, in the embodiment of the present application, several classifications of paying close attention to most at family can only be taken to join concrete statistic processes, such as, each user only gets its 6 classifications paid close attention to most, certainly, if the classification that certain user pays close attention to is less than 6, just whole classifications of this user's actual concern are joined statistic processes.

After getting each user and be higher to the attention rate of which classification respectively, just can obtain user's overlay capacity that each classification is independent, wherein, the independent user's overlay capacity of certain classification is that focus comprises such object number of users.That is, the classification can paid close attention to respectively according to each user, counts the number of users that each classification is corresponding respectively, this number of users is defined as user's overlay capacity that each classification is independent.Such as, suppose that one has three users, be respectively A, B, C, wherein:

The classification that user A pays close attention to comprises: clothing, digital product class and toiletries;

The classification that user B pays close attention to comprises: clothing, toiletries and outdoor class of moving;

The classification that user C pays close attention to comprises: toiletries and digital product class.

Then for clothing, user A and user B has paid close attention to, and therefore, the independent user's overlay capacity of this clothing is 2; For digital product class, user A and user C has paid close attention to, and therefore, the independent user's overlay capacity of digital product class is 2; For toiletries, user A, B, C have paid close attention to, and therefore, the independent user's overlay capacity of toiletries is 3, by that analogy.Certainly, in actual system, number of users and classification quantity are all a lot, and just the simple principle to statistics is introduced here.

After obtaining the independent user's overlay capacity of each classification, just according to independent user's overlay capacity, each classification can be sorted, generate a classification sequence, and each classification be respectively in sequence gives continuous print sequence number.Like this, for a classification, the information of two aspects can be got, one is independent user's overlay capacity, another is exactly sequence number in the sequence, can the information of these two aspects as the build-in attribute of classification, carry out follow-up Function Fitting and flow estimates calculating.Such as, the storage format of classification and attribute can be:

(Key: classification) (value: user's overlay capacity of sequence sequence number+independent)

Next just first can utilize the basic data of above acquisition, carry out the Function Fitting of segmentation.

S102: classify to classification according to the difference of each classification Sorting distance in classification sequence, obtains a preset number subsequence; Wherein, described Sorting distance is the difference between the sequence number of two classifications, and in same subsequence, the sequence number difference between adjacent classification is all equal;

Specifically when classifying to classification according to Sorting distance, be the equal of extract according to certain interval from classification sequence, the classification extracted forms a new sequence, owing to being extract from the classification sequence step S101, therefore, a subsequence can be called.Same, when extracting according to other interval, other subsequence can be obtained.

Such as, the classification sequence sorted from big to small according to user's overlay capacity that each classification is independent is: Cat_1, Cat_2, Cat_3 ..., Cat_N.Then, specifically when classifying to classification, the classification that just can be 1 from sequence number, extract the classification composition subsequence that sequence number is spaced apart length.

As length=1, then { Cat_1, Cat_2, Cat_3......Cat_N} are as a subsequence in extraction;

Length=2, then { Cat_1, Cat_3, Cat_5......Cat_m-2, Cat_m......} are as a subsequence in extraction;

Length=k is then categorized as, and { Cat_1, Cat_k+1, Cat_2k+1......} are as a subsequence.

Like this, finally multiple subsequence can be produced.

Here need to carry out some explanation following:

First, for identical Sorting distance, if the initial classification selected is different, the subsequence then generated may be different, such as, as length=2, if extracted from the classification that sequence number is 1, the subsequence then obtained is { Cat_1, Cat_3, Cat_5......Cat_m-2, Cat_m ... }, and wherein, m is odd number; But if extract from the classification that sequence number is 2, then the subsequence obtained is { Cat_2, Cat_4, Cat_6 ..., Cat_n-2, Cat_n ... }, and wherein, n is even number.Further, when length is larger, the number of the different subsequences that can produce is more.Due in follow-up Function Fitting process, need to obtain respectively the fitting function under each Sorting distance, and the fitting function under certain Sorting distance, be calculate according to object user overlay capacity situation all kinds of in the subsequence under this Sorting distance.Therefore, in order to reduce calculated amount, under each Sorting distance, only can extract a subsequence, like this, under each Sorting distance, only needing just can calculate corresponding fitting function based on a subsequence.Wherein, the subsequence that each Sorting distance is corresponding can be extract from the classification that sequence number is 1, and namely the initial classification of each subsequence is the classification that sequence number is minimum in the classification sequence of basic data.Certainly, when calculated amount allows, also multiple subsequence can be extracted with different initial classifications respectively under same Sorting distance, final when digital simulation function, can calculate based on each subsequence respectively equally, finally again the value that each subsequence under same Sorting distance calculates be averaged.Experiment proves, the value that under same Sorting distance, each subsequence calculates is substantially equal, and this also shows further, lower of same Sorting distance extracts a subsequence and has rationality, that is, while fitting function can being obtained more accurately, calculated amount is also reduced.

The second, the classification One's name is legion in system, such as, generally have several thousand even up to ten thousand, incite somebody to action in theory, if classification quantity is N, then the Sorting distance between different classification has N-1 kind.But when reality is classified to classification according to Sorting distance, do not need the matching all carrying out function for all Sorting distances, generally, one maximal value can be set for Sorting distance, as long as carry out Function Fitting to each Sorting distance below this maximal value.Such as, suppose that classification adds up to 2000, maximum Sorting distance can get 100, then carries out Function Fitting for each Sorting distance respectively.Like this, calculated amount can be reduced further, and these fitting functions generally just can meet the most demands in practical application.

S103: the classification in same subsequence is divided into the combination of multiple classification, each classification is made to combine than the many classifications of previous classification combination, add up the not duplicate customer overlay capacity of each classification combined accumulated, and calculate each classification and combine the user's overlay capacity rate of growth combined relative to previous classification;

In same subsequence, each classification still according to sequence number from small to large (namely independent user's overlay capacity from big to small) order arrangement.Specifically when carrying out matching according to object user overlay capacity situation all kinds of in same subsequence to the function under corresponding Sorting distance, first the classification in same subsequence can be divided into the combination of multiple classification.Specifically when dividing classification combination, can be benchmark classification by a classification (classification that such as sequence number is minimum) in subsequence, and according to the order of classification in subsequence add at every turn a classification form next classification combination, like this, each classification combination is than the many classifications of previous classification combination.Such as:

For subsequence: Cat_1, Cat_k+1, Cat_2k+1 ...), the classification combination obtained can comprise: { Cat_1}, { Cat_1, Cat_k+1}, { Cat_1, Cat_k+1, Cat_2k+1}.....

That is, first classification combination is made up of the classification of first in subsequence, second classification combination is exactly the combination be made up of the first two classification in subsequence, 3rd classification combination is exactly the combination be made up of the classification of first three in subsequence, also be, the combination of i-th classification is exactly the combination be made up of i classification before in subsequence, by that analogy.

Obtain the combination of multiple classification in same subsequence after, for the classification combination comprising two and two or more classification, the category information can also paid close attention to according to each user counted in step S101, counts the not duplicate customer overlay capacity of classification combined accumulated.Concrete, owing to can know that from the data obtained in advance each user is higher to the attention rate of which classification respectively, and the identification information such as the ID of each user known, therefore, the user ID that in the combination of same classification, each classification is corresponding respectively can be counted, like this, user ID is gathered, and remove the user ID of repetition, the user ID number finally obtained, just can be defined as the not duplicate customer overlay capacity of this classification combined accumulated.

Such as, still suppose that one has three users, is respectively A, B, C, wherein:

The classification that user A pays close attention to comprises: clothing, digital product class and toiletries;

The classification that user B pays close attention to comprises: clothing, toiletries and outdoor class of moving;

The classification that user C pays close attention to comprises: toiletries and digital product class.

Suppose that certain classification is combined as { clothing, digital product class }, wherein, the user that clothing covers comprises user A and user B (independent user's overlay capacity is 2), the user that digital product class covers comprises user A and user C (independent user's overlay capacity is 2), now, the user that clothing and digital product class cover is gathered, and after removing the user of repetition, the user obtained comprises A, B, C, therefore, the not duplicate customer overlay capacity that just can obtain this classification combined accumulated is 3, wherein, user A occurs in two classifications, but only can calculate once.

In a word, user's overlay capacity of accumulation can be counted in the manner described above for each classification combination in same subsequence.Afterwards, just can calculate the inner each classification of subsequence and combine the user's overlay capacity rate of growth combined relative to last classification.Such as, if to calculate classification combination Cat_1, Cat_k+1} relative to user's overlay capacity rate of growth of Cat_1}, then can calculate according to following formula (1):

Like this, for same subsequence, multiple user's overlay capacity rate of growth just can be calculated, such as:

P _length＝k＝{P ₁，P ₂，P ₃........} (2)

And then rule can be found from these rate of growth, calculate the fitting function under corresponding Sorting distance.Also the matching of function can be carried out in this manner for other Sorting distances.

S104: carry out matching according to each user's overlay capacity rate of growth obtained in same subsequence and preset Function Format, obtain the fitting function under corresponding Sorting distance;

Owing to sorting according to user's overlay capacity that each classification is independent in subsequence same in the embodiment of the present application, therefore, the change for each user's overlay capacity rate of growth in formula (2) can present a power function curve.Like this, when carrying out Function Fitting, the function setup just can will simulated in advance is power-law scheme, the fitting function that different Sorting distances is corresponding has identical form, when carrying out matching to each Sorting distance respectively, will calculate the coefficient in power function and/or power exponent exactly, that is, for different Sorting distances, have different coefficients and/or power exponent, concrete value needs to determine according to each user's overlay capacity rate of growth calculated in step S103.

Such as, the form of power function can be as follows:

$Y_{\mathrm{length}=k}={\mathrm{\α}}_k{\left(\frac{\mathrm{rint}\left(\frac{M}{k}\right)}{k}\right)}^{-{\mathrm{\β}}_k}---\left(3\right)$

Wherein, length represents classification Sorting distance, α _krepresent the coefficient of fitting function, β _krepresent the power exponent (being just) of fitting function, M is the sequence number of the classification that the combination of current classification increases relative to previous classification combination.The process of matching is exactly when K gets a certain concrete value, will calculate (α _k, β _k) value, there is N number of different Sorting distance, just can obtain N group (α _k, β _k) value.

The information such as so far, the process of Function Fitting terminates, the keyword just can selected according to information input side afterwards, estimate the accumulation user overlay capacity of classification combination.

S105: after information input side have selected target keyword, is mapped as the multiple target classifications in system by described target keyword, determine each target class object sequence number and independent user's overlay capacity according to described basic data;

During specific implementation, after information input side have selected target keyword, keyword just can be mapped as the target classification in system by system, and is generally multiple.Such as, the keyword that information input side is selected is " comfortable ventilating ", " arriving account in time ", " cotton textiles " and " cost performance ", and 4 the target classifications obtained after mapping are respectively " sport footwear ", " rechargeable card ", " clothes shoes and hats ", " electronic product ".Next just can estimate the accumulation user overlay capacity of the classification combination of whole target classifications composition.Specifically when estimating, still can use the basic data got in step S101, also can use the fitting function under each Sorting distance obtained in step S104 simultaneously.

S106: using target classification minimum for sequence number as target fiducials classification, by this target fiducials classification, each target classification forms objective cross with other respectively, and calculates two target class object Sorting distances in each objective cross;

Such as, according to the keyword that information input side is selected, the target classification mapped out comprises: A, B, C, D tetra-classifications, in the classification sequence that step S101 obtains, the sequence number that these four target classifications are corresponding 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 category-A object is the most forward, therefore, with category-A order for benchmark classification, then obtain { A, B}, { A, C}, { A respectively, such three objective cross of D}, each objective cross comprises two target classifications, and one of them is target fiducials classification, can calculate two target class object Sorting distances in each objective cross simultaneously.Wherein:

Length _A、B＝6-2＝4

Length _A、C＝10-2＝8

Length _A、D＝20-2＝18

S107: utilize the fitting function that the Sorting distance of each objective cross is corresponding respectively, estimate the accumulation user increment of each objective cross relative to target fiducials classification, and user's overlay capacity independent for target fiducials classification is added with described user's increment of accumulating of each objective cross, estimate out accumulation user overlay capacity when being combined by all target classifications;

Obtain multiple objective cross, and after each self-corresponding Sorting distance, just can utilize the fitting function of corresponding Sorting distance respectively, estimate the accumulation user increment of each objective cross relative to target fiducials classification.Also be, { A, B} are relative to { non-repetitive user's overlay capacity N1 of the increase of A} can to obtain objective cross respectively, { A, C} are relative to { non-repetitive user's overlay capacity N2 of the increase of A} for objective cross, and { A, D} are relative to { non-repetitive user's overlay capacity N3 of the increase of A} can to obtain objective cross respectively, like this, as long as be added with N1, N2, N3 by user's overlay capacity independent for classification A, { the not duplicate customer overlay capacity of this classification combined accumulated of A, B, C, D} just can be obtained.

Wherein, because fitting function is the function about accumulation user rate of growth, therefore, specifically when calculating the accumulation user increment of certain objective cross relative to target fiducials classification according to fitting function, can carry out according to following steps:

First, by another target class object sequence number outside target fiducials classification in Sorting distance corresponding for objective cross and objective cross (such as, for objective cross { A, B}, classification A is benchmark classification, therefore this sequence number just refers to the sequence number of classification B), be brought in the fitting function of this Sorting distance, obtain the accumulation user rate of growth of this objective cross relative to target fiducials classification.Such as, { fitting function during A, B} correspondence length=4, { fitting function during A, C} correspondence length=8, { the fitting function of the corresponding length=18 of A, D}, therefore { A, B} are relative to { the accumulation user rate of growth of A} can be obtained by fitting function during length=4 for objective cross, also namely, by M=6, in fitting function when being brought into length=4:

$Y_{\mathrm{length}=4}={\mathrm{\α}}_4{\left(\frac{\mathrm{rint}\left(\frac{6}{4}\right)}{4}\right)}^{-{\mathrm{\β}}_4}---\left(4\right)$

Wherein, rint () is for rounding up.Owing to having drawn (α in step S104 ₄, β ₄) value, therefore, just can calculate a definite numeral, this numeral just represents objective cross, and { A, B} are relative to { the accumulation user rate of growth of A}.

After obtaining above-mentioned rate of growth, be equivalent to cicada relative to { growth ratio of A} is how many, if will calculate objective cross, { A, B} are relative to { the accumulation user increment of A} also needs to count { the accumulation user overlay capacity of A, B} in theory.But in the embodiment of the present application, in order to reduce calculated amount, obtain the statistical value obtained in the process of fitting function in training before making full use of, can first find { A, the subsequence that the Sorting distance length=4 that B} is corresponding is corresponding, if wherein just comprise classification A, B, then be equivalent to carrying out adding up { A in the process of training, the accumulation user overlay capacity of B}, therefore, { user's overlay capacity of A} adds { A to direct use, the accumulation user overlay capacity of B}, be multiplied by { the A calculated again, B} is relative to { the accumulation user rate of growth of A}, just { A can be obtained, B} is relative to { the accumulation user increment of A}.

But as mentioned before, in the embodiment of the present application, under same Sorting distance, only may extract a subsequence, therefore, probably occur the situation not comprising A, B in subsequence.Such as, the subsequence that Sorting distance length=4 is corresponding is:

accuNum _j＝{Cat_1、Cat_5，Cat_9，......}

And A is Cat_2, B is Cat_6, therefore, do not appear in this subsequence, now, can from this subsequence, obtain and target fiducials category number (i.e. the sequence number 2 of classification A) immediate first classification (in this example embodiment corresponding Cat_1), and with object sequence number another kind of in objective cross (i.e. the sequence number 6 of classification B) immediate second classification (in this example embodiment corresponding Cat_5), then, according to user's overlay capacity that the first classification is independent, the not duplicate customer overlay capacity that first classification and the second classification are accumulated, and this objective cross is relative to the accumulation user rate of growth of target fiducials classification, estimate out the accumulation user increment of this objective cross relative to target fiducials classification.Also namely, { user's overlay capacity of Cat_1} and { the accumulation user overlay capacity sum of Cat_1, Cat_5}, Y when being then multiplied by M=6 is calculated _length=4value, the objective cross { discreet value of non-repetitive user's overlay capacity that A, B} increase than classification A can be drawn.

Profit uses the same method, and can also calculate objective cross { discreet value of non-repetitive user's overlay capacity that A, C} increase than classification A, and the objective cross { discreet value of non-repetitive user's overlay capacity that A, D} increase than classification A respectively.The final discreet value sum calculating non-repetitive user's overlay capacity that the independent user's overlay capacity of classification A increases than classification A with each objective cross again, { the not duplicate customer overlay capacity of this classification combined accumulated of A, B, C, D} just can finally estimated out.

Comprehensive above-mentioned computation process can abstractly be following formula:

$\left\{\begin{array}{c}{\mathrm{accuNum}}_1+{\mathrm{\Σ}}_{i=2}^n(\left({\mathrm{accuNum}}_i\right)*Y_{\mathrm{length}})\\ Y_{\mathrm{length}}={\mathrm{\α}}_{\mathrm{length}}{\left(\frac{\mathrm{rint}\left(\frac{{\mathrm{OrderNumber}}_i}{\mathrm{length}}\right)}{\mathrm{length}}\right)}^{-{\mathrm{\β}}_{\mathrm{length}}}\\ \mathrm{accuNum}=\{\left({\mathrm{Cat}}_1\right),\left({\mathrm{Cat}}_{k+1}\right),\left({\mathrm{Cat}}_{2k+1}\right)......\}\end{array}\right.---\left(5\right)$

Wherein, n is target class object quantity, and OrderNumber is another target class object sequence number in each objective cross outside target fiducials classification.

S108: when receiving the request obtaining accumulation user overlay capacity, the accumulation user overlay capacity estimated is supplied to information input side.

{ after the not duplicate customer overlay capacity of this classification combined accumulated of A, B, C, D}, if information input side needs, then can point out in interface obtaining.Such as, during specific implementation, the button of printed words such as " obtaining user's overlay capacity " can be provided on the interface that keyword is selected by information input side, after information input side have selected keyword, if click this button, just the not duplicate customer overlay capacity of the accumulation estimated out can be shown on the surface.Like this, information input side can judge whether this user's overlay capacity meets the demand of oneself accordingly, if met, then can throw according to this keyword, otherwise, other keyword can also be reselected.

In a word, pass through the embodiment of the present application, with the classification in the subsequence of limited quantity for representative, calculate the not duplicate customer overlay capacity of some classification combined accumulated, and fitting function corresponding under each Sorting distance can be obtained, in order under the corresponding Sorting distance of matching, certain classification combines the accumulation user rate of growth combined relative to last classification, and then just can according to the accumulation user overlay capacity of classification combination each in the fitting function under each Sorting distance and subsequence, estimate out accumulation user overlay capacity when being combined by the target classification that information input side is selected, so that input side points out accordingly to information.Visible, by with upper type, calculated amount can be narrowed down within the scope of subsequence corresponding to the Sorting distance of limited quantity, in calculated amount is limited in scope that computing system can realize.Meanwhile, utilize the not duplicate customer overlay capacity of classification combined accumulated to carry out the matching of function and follow-up estimating, therefore, solve the problem that same user repeats now in inhomogeneity.

It should be noted that, in the embodiment of the present application, the executive agent of each step described in Fig. 1 can be the server of certain business platform, wherein, obtains basic data and the function under each Sorting distance is carried out to the process of matching, can online under complete.

Corresponding with the reminding method of user's overlay capacity information that the embodiment of the present application provides, the embodiment of the present application additionally provides a kind of suggestion device of user's overlay capacity information, and see Fig. 2, this device can comprise:

Basic data acquiring unit 201, for the category information paid close attention to respectively according to each user got in advance, obtains basic data; Described basic data comprises the independent user's overlay capacity of each classification, according to the classification sequence obtained after the independent descending sequence of user's overlay capacity, and the sequence number of each classification in classification sequence;

Classification taxon 202, for classifying to classification according to the difference of each classification Sorting distance in classification sequence, obtains a preset number subsequence; Wherein, described Sorting distance is the difference between the sequence number of two classifications, and in same subsequence, the sequence number difference between adjacent classification is all equal;

Growth Rate Calculation unit 203, for the classification in same subsequence being divided into the combination of multiple classification, each classification is made to combine than the many classifications of previous classification combination, according to the category information that described each user got in advance pays close attention to respectively, add up the not duplicate customer overlay capacity of each classification combined accumulated, and calculate each classification and combine the user's overlay capacity rate of growth combined relative to previous classification;

Fitting unit 204, for carrying out matching according to each user's overlay capacity rate of growth obtained in same subsequence and preset Function Format, obtains the fitting function under corresponding Sorting distance;

The unit of above 201 to 204 correspondences is units required in training process.

Target classification determining unit 205, for after information input side have selected target keyword, is mapped as the multiple target classifications in system, determines each target class object sequence number and independent user's overlay capacity according to described basic data by described target keyword;

Objective cross determining unit 206, for using target classification minimum for sequence number as target fiducials classification, by this target fiducials classification, each target classification forms objective cross with other respectively, and calculates two target class object Sorting distances in each objective cross;

Estimate unit 207, for utilizing the fitting function that the Sorting distance of each objective cross is corresponding respectively, estimate the accumulation user increment of each objective cross relative to target fiducials classification, and user's overlay capacity independent for target fiducials classification is added with described user's increment of accumulating of each objective cross, estimate out accumulation user overlay capacity when being combined by all target classifications;

Each unit of above 205 to 207 correspondences is the unit estimated needed for process.

Tip element 208, for when receiving the request obtaining accumulation user overlay capacity, is supplied to information input side by the accumulation user overlay capacity estimated.

Wherein, due in each subsequence, according to independent user's overlay capacity, classification is sorted, therefore, the change that each classification combination covers rate of growth relative to the user that last classification combines is rendered as power function curve, therefore, preset Function Format can be power-law scheme, described power-law scheme comprises coefficient to be determined and/or power exponent, and therefore, fitting unit 204 specifically may be used for:

According to each user's overlay capacity rate of growth obtained in same classification subsequence and preset power-law scheme, determine coefficient and/or power exponent under corresponding Sorting distance; Described coefficient and/or power exponent and this Sorting distance are brought in preset Function Format, obtain the fitting function under corresponding Sorting distance.

In order to reduce calculated amount further, under same Sorting distance, a subsequence only can be got.Further, the subsequence under each Sorting distance is all initial classification with the classification that sequence number in described classification sequence is minimum.

During specific implementation, estimating unit 207 can comprise:

Growth Rate Calculation subelement, for by another target class object sequence number outside target fiducials classification in Sorting distance corresponding for described objective cross and objective cross, be brought in the fitting function of this Sorting distance, obtain the accumulation user rate of growth of this objective cross relative to target fiducials classification;

Increment computation subunit, for from subsequence corresponding to this Sorting distance, obtain and immediate first classification of target fiducials category number, and with immediate second classification of another kind of object sequence number in objective cross, according to user's overlay capacity that described first classification is independent, and the first not duplicate customer overlay capacity accumulated of classification and the second classification, and this objective cross described is relative to the accumulation user rate of growth of target fiducials classification, estimate out the accumulation user increment of this objective cross relative to target fiducials classification.

Wherein, when the difference according to Sorting distance is classified to classification, the ratio between maximum Sorting distance and classification sum is less than preset threshold value.

During specific implementation, basic data acquiring unit 201 specifically may be used for:

In advance according to the historical operation behavior record of user within preset time period, obtain the category information of the highest preset number of this user's attention rate;

The information got from each user is gathered, counts the number of users that each classification is corresponding respectively, this number of users is defined as user's overlay capacity that each classification is independent.

When carrying out Function Fitting, described fitting unit 204, when adding up the not duplicate customer overlay capacity of each classification combined accumulated, specifically may be used for:

Count the user ID that in the combination of same classification, each classification is corresponding respectively;

Described user ID is gathered, and removes the user ID of repetition, by final user ID number, be defined as the not duplicate customer overlay capacity of this classification combined accumulated.

In a word, pass through the embodiment of the present application, with the classification in the subsequence of limited quantity for representative, calculate the not duplicate customer overlay capacity of some classification combined accumulated, and fitting function corresponding under each Sorting distance can be obtained, in order under the corresponding Sorting distance of matching, certain classification combines the accumulation user rate of growth combined relative to last classification, and then just can according to the accumulation user overlay capacity of classification combination each in the fitting function under each Sorting distance and subsequence, estimate out accumulation user overlay capacity when being combined by the target classification that information input side is selected, so that input side points out accordingly to information.Visible, by with upper type, calculated amount can be narrowed down within the scope of subsequence corresponding to the Sorting distance of limited quantity, in calculated amount is limited in scope that computing system can realize.Meanwhile, utilize the not duplicate customer overlay capacity of classification combined accumulated to carry out the matching of function and follow-up estimating, therefore, solve the problem that same user repeats now in inhomogeneity.

As seen through the above description of the embodiments, those skilled in the art can be well understood to the mode that the application can add required general hardware platform by software and realizes.Based on such understanding, the technical scheme of the application can embody with the form of software product the part that prior art contributes in essence in other words, this computer software product can be stored in storage medium, as ROM/RAM, magnetic disc, CD etc., comprising some instructions in order to make a computer equipment (can be personal computer, server, or the network equipment etc.) perform the method described in some part of each embodiment of the application or embodiment.

Each embodiment in this instructions all adopts the mode of going forward one by one to describe, between each embodiment identical similar part mutually see, what each embodiment stressed is the difference with other embodiments.Especially, for system or system embodiment, because it is substantially similar to embodiment of the method, so describe fairly simple, relevant part illustrates see the part of embodiment of the method.System described above and system embodiment are only schematic, the wherein said unit illustrated as separating component or can may not be and physically separates, parts as unit display can be or may not be physical location, namely can be positioned at a place, or also can be distributed in multiple network element.Some or all of module wherein can be selected according to the actual needs to realize the object of the present embodiment scheme.Those of ordinary skill in the art, when not paying creative work, are namely appreciated that and implement.

Above to the suggestion device of user's overlay capacity information that the application provides, be described in detail, apply specific case herein to set forth the principle of the application and embodiment, the explanation of above embodiment is just for helping method and the core concept thereof of understanding the application; Meanwhile, for one of ordinary skill in the art, according to the thought of the application, all will change in specific embodiments and applications.In sum, this description should not be construed as the restriction to the application.

Claims (10)

1. a reminding method for user's overlay capacity information, is characterized in that, comprising:

According to the category information that each user got in advance pays close attention to respectively, obtain basic data; Described basic data comprises the independent user's overlay capacity of each classification, according to the classification sequence obtained after the independent descending sequence of user's overlay capacity, and the sequence number of each classification in classification sequence;

According to the difference of each classification Sorting distance in classification sequence, classification is classified, obtain a preset number subsequence; Wherein, described Sorting distance is the difference between the sequence number of two classifications;

Classification in same subsequence is divided into the combination of multiple classification, each classification is made to combine than the many classifications of previous classification combination, according to the category information that described each user got in advance pays close attention to respectively, add up the not duplicate customer overlay capacity of each classification combined accumulated, and calculate each classification and combine the user's overlay capacity rate of growth combined relative to previous classification;

Carry out matching according to each user's overlay capacity rate of growth obtained in same subsequence and preset Function Format, obtain the fitting function under corresponding Sorting distance;

After information input side have selected target keyword, described target keyword is mapped as the multiple target classifications in system, determines each target class object sequence number and independent user's overlay capacity according to described basic data;

Select target benchmark classification in multiple target classifications that described mapping obtains, by this target fiducials classification, each target classification forms objective cross with other respectively, and calculates two target class object Sorting distances in each objective cross;

Utilize the fitting function that the Sorting distance of each objective cross is corresponding respectively, estimate the accumulation user increment of each objective cross relative to target fiducials classification, and user's overlay capacity independent for target fiducials classification is added with described user's increment of accumulating of each objective cross, estimate out accumulation user overlay capacity when being combined by all target classifications;

When receiving the request obtaining accumulation user overlay capacity, the accumulation user overlay capacity estimated is supplied to information input side.

2. method according to claim 1, it is characterized in that, described preset Function Format is power-law scheme, described power-law scheme comprises coefficient to be determined and/or power exponent, described each user's overlay capacity rate of growth according to obtaining in same classification subsequence and preset Function Format carry out matching, obtain the fitting function under corresponding Sorting distance, comprising:

According to each user's overlay capacity rate of growth obtained in same classification subsequence and preset power-law scheme, determine coefficient and/or power exponent under corresponding Sorting distance;

Described coefficient and/or power exponent and this Sorting distance are brought in preset Function Format, obtain the fitting function under corresponding Sorting distance.

3. method according to claim 1, is characterized in that, gets a subsequence under same Sorting distance.

4. method according to claim 3, is characterized in that, the subsequence under each Sorting distance is all initial classification with the classification that sequence number in described classification sequence is minimum.

5. the method according to any one of Claims 1-4, is characterized in that, describedly utilizes the fitting function that the Sorting distance of each objective cross is corresponding respectively, estimates the accumulation user increment of each objective cross relative to target fiducials classification, comprising:

By another target class object sequence number outside target fiducials classification in Sorting distance corresponding for described objective cross and objective cross, be brought in the fitting function of this Sorting distance, obtain the accumulation user rate of growth of this objective cross relative to target fiducials classification;

From the subsequence that this Sorting distance is corresponding, obtain and immediate first classification of target fiducials category number, and with immediate second classification of another kind of object sequence number in objective cross, according to user's overlay capacity that described first classification is independent, and the first not duplicate customer overlay capacity accumulated of classification and the second classification, and this objective cross described is relative to the accumulation user rate of growth of target fiducials classification, estimate out the accumulation user increment of this objective cross relative to target fiducials classification.

6. the method according to any one of Claims 1-4, is characterized in that, when the difference according to Sorting distance is classified to classification, 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, is characterized in that, according to the category information that each user got in advance pays close attention to respectively, obtains basic data, comprising:

In advance according to the historical operation behavior record of user within preset time period, obtain the category information of the highest preset number of this user's attention rate;

The information got from each user is gathered, counts the number of users that each classification is corresponding respectively, this number of users is defined as user's overlay capacity that each classification is independent.

8. method according to claim 7, is characterized in that, when carrying out Function Fitting, the not duplicate customer overlay capacity of each classification combined accumulated of described statistics, comprising:

Count the user ID that in the combination of same classification, each classification is corresponding respectively;

Described user ID is gathered, and removes the user ID of repetition, by final user ID number, be defined as the not duplicate customer overlay capacity of this classification combined accumulated.

9. a suggestion device for user's overlay capacity information, is characterized in that, comprising:

Basic data acquiring unit, for the category information paid close attention to respectively according to each user got in advance, obtains basic data; Described basic data comprises the independent user's overlay capacity of each classification, according to the classification sequence obtained after the independent descending sequence of user's overlay capacity, and the sequence number of each classification in classification sequence;

Classification taxon, for classifying to classification according to the difference of each classification Sorting distance in classification sequence, obtains a preset number subsequence; Wherein, described Sorting distance is the difference between the sequence number of two classifications, and in same subsequence, the sequence number difference between adjacent classification is all equal;

Growth Rate Calculation unit, for the classification in same subsequence being divided into the combination of multiple classification, each classification is made to combine than the many classifications of previous classification combination, according to the category information that described each user got in advance pays close attention to respectively, add up the not duplicate customer overlay capacity of each classification combined accumulated, and calculate each classification and combine the user's overlay capacity rate of growth combined relative to previous classification;

Fitting unit, for carrying out matching according to each user's overlay capacity rate of growth obtained in same subsequence and preset Function Format, obtains the fitting function under corresponding Sorting distance;

Target classification determining unit, for after information input side have selected target keyword, is mapped as the multiple target classifications in system, determines each target class object sequence number and independent user's overlay capacity according to described basic data by described target keyword;

Objective cross determining unit, for using target classification minimum for sequence number as target fiducials classification, by this target fiducials classification, each target classification forms objective cross with other respectively, and calculates two target class object Sorting distances in each objective cross;

Estimate unit, for utilizing the fitting function that the Sorting distance of each objective cross is corresponding respectively, estimate the accumulation user increment of each objective cross relative to target fiducials classification, and user's overlay capacity independent for target fiducials classification is added with described user's increment of accumulating of each objective cross, estimate out accumulation user overlay capacity when being combined by all target classifications;

Tip element, for when receiving the request obtaining accumulation user overlay capacity, is supplied to information input side by the accumulation user overlay capacity estimated.

10. device according to claim 9, is characterized in that, described in estimate unit and comprise:

Growth Rate Calculation subelement, for by another target class object sequence number outside target fiducials classification in Sorting distance corresponding for described objective cross and objective cross, be brought in the fitting function of this Sorting distance, obtain the accumulation user rate of growth of this objective cross relative to target fiducials classification;

Increment computation subunit, for from subsequence corresponding to this Sorting distance, obtain and immediate first classification of target fiducials category number, and with immediate second classification of another kind of object sequence number in objective cross, according to user's overlay capacity that described first classification is independent, and the first not duplicate customer overlay capacity accumulated of classification and the second classification, and this objective cross described is relative to the accumulation user rate of growth of target fiducials classification, estimate out the accumulation user increment of this objective cross relative to target fiducials classification.