CN108960929A - Consider the social networks marketing seed user choosing method that existing product influences - Google Patents
Consider the social networks marketing seed user choosing method that existing product influences Download PDFInfo
- Publication number
- CN108960929A CN108960929A CN201810778152.2A CN201810778152A CN108960929A CN 108960929 A CN108960929 A CN 108960929A CN 201810778152 A CN201810778152 A CN 201810778152A CN 108960929 A CN108960929 A CN 108960929A
- Authority
- CN
- China
- Prior art keywords
- user
- marketing
- social networks
- product
- seed
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 69
- 230000000694 effects Effects 0.000 claims description 17
- 238000004364 calculation method Methods 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 10
- 230000002829 reductive effect Effects 0.000 claims description 9
- 230000003321 amplification Effects 0.000 claims description 8
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 8
- 230000003993 interaction Effects 0.000 claims description 6
- 230000005764 inhibitory process Effects 0.000 claims description 5
- 230000008901 benefit Effects 0.000 claims description 4
- 238000005336 cracking Methods 0.000 claims description 4
- 239000011159 matrix material Substances 0.000 claims description 4
- 238000004590 computer program Methods 0.000 claims description 3
- 238000003860 storage Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 claims description 2
- 238000005457 optimization Methods 0.000 abstract description 6
- 230000002457 bidirectional effect Effects 0.000 abstract description 5
- 235000013350 formula milk Nutrition 0.000 description 13
- 238000004422 calculation algorithm Methods 0.000 description 12
- 238000002474 experimental method Methods 0.000 description 6
- 230000006854 communication Effects 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 238000012217 deletion Methods 0.000 description 3
- 230000037430 deletion Effects 0.000 description 3
- 244000097202 Rathbunia alamosensis Species 0.000 description 2
- 235000009776 Rathbunia alamosensis Nutrition 0.000 description 2
- 241000700605 Viruses Species 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 238000009472 formulation Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 244000046052 Phaseolus vulgaris Species 0.000 description 1
- 235000010627 Phaseolus vulgaris Nutrition 0.000 description 1
- 238000004873 anchoring Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 230000000452 restraining effect Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000003612 virological effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- Game Theory and Decision Science (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Primary Health Care (AREA)
- Tourism & Hospitality (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of social networks marketing seed user choosing methods that consideration existing product influences, it include the following steps: (1) that social networks is modeled as a width socialgram G=(V, E, W), wherein vertex set V indicates all social user's set in social networks, side collection E indicates social networks set between social user, and W indicates the weight set on all sides；Step 2, initialization seed user's set S are empty set, i.e.,Initialising subscriber set U is all user's set, i.e. U ← V in step 1；Step 3, the marketing influence for calculating all users in set U in step 2.The social networks marketing seed user choosing method that above-mentioned consideration existing product influences, consider influence of the existing product to marketing product promotion, the marketing influence of social user is measured in terms of social networks and existing product two, using return rate of marketing as optimization aim, gathered using bidirectional optimistic method selected seed user, to maximize marketing return rate.
Description
Technical field
The present invention relates to social networks and information communication sphere, more particularly to the social networks for considering that existing product influences
Marketing seed user choosing method.
Background technique
Online social networks (the Online Social such as Facebook, Twitter, Sina weibo and wechat
Networks, abbreviation OSNs), our daily life is constantly being incorporated, people are made friends by social networks, deliver sight
Point and sharing information.Moreover, since social networks reduces the threshold of communication, more and more users start social network
Network carries out on line, network marketing and shopping online as the important tool for participating in social activities so that social activities be moved to
Nowadays become very universal, it can be seen that, social networks is changing our life deeply.
Nowadays the userbase of social networks is huge, contains huge researching value and commercial value, only Facebook
Just reach 21.3 hundred million in the monthly active users in the end of the year 2017, fourth quarter business revenue volume is 13,000,000,000 dollars, wherein 98.5%
From advertising business；It can be seen that carrying out network marketing using social networks and promoting the pillar industry for having become current social networks
Business.Compared to traditional media advertisement, social networks is marketed, and covering crowd is wider, route of transmission is more various, communication mode is more efficient,
Network marketing mode most popular at present is virus marketing (Viral Marketing), it is passed from mouth to mouth using social user
(Word-of-mouth) propagation characteristic will go out with marketing product promotion, and communication process very efficient quick is similar to virus
It propagates, so gain the name.
Social networks marketing passes through mouth as seed user, then by them as shown in Figure 1, choosing a collection of initial user first
Mouth passes on from one to another mode and promotes to surrounding user (such as good friend or bean vermicelli), if surrounding user receives the marketing product, will continue
It promotes, marketing behavior is constantly propagated between user in cascaded fashion, achievees the purpose that advertisement promotion.Currently, social networks is marketed
The optimization problem for being usually modeled as belt restraining is solved, i.e. the marketing payoff maximization problem of budgetary restraints, and marketing needs
A part of initial user is chosen as seed user, marketing product promotion is gone out by its marketing influence, battalion herein
Pin influence power indicates the ability that user goes out marketing product promotion, and budget then is used to limit the expense of marketing, generally pays
Seed user.Above-mentioned optimization problem aims at: seed user is selected, to maximize marketing income.Existing method for solving
Main to design greedy algorithm using the monotonicity and submodularity of marketing revenue function, monotonicity means income with seed user
Increase and increase, but submodularity is being gradually reduced marginal benefit i.e. increment simultaneously.Greedy algorithm is by continuously attempting to
Finding out one every time can bring the user of maximum marginal benefit as seed user, until budget uses until exhausted, process such as Fig. 2
It is shown.
There are following technical problems for traditional technology:
Firstly, there are a kind of tradeoffs between income and budget, be difficult to formulate suitable budget before marketing, especially when pair
It is even more so when product and the market shortage enough understanding of marketing；Secondly, algorithm is mainly used for solving single product at present
Marketing, it is not intended that the influence of existing product, however, user has had very before receiving promotion in actual life
Multi-product, these products may play the role of promotion or inhibition to the popularization of marketing product, such as when user has purchased baby
It is easier to promote baby milk to it after feeding bottle, certainly if when he has possessed baby milk, sales promotion meeting is highly difficult；Again,
Existing method is when calculating the influence power in user's marketing process, the main topological attribute degree of progress from user in social networks
Amount did not accounted for influence caused by existing product both without the intimate degree between differentiation user yet, thus can not accurately reflect
User's marketing influence, it is limited to practical marketing directive significance with the seed user that this finds out；Finally, the complexity of existing method
It is excessively high, for nowadays extensive social networks, lack practical value.
Summary of the invention
Based on this, it is difficult to determine for budget existing for existing method, marketing influence measurement inaccuracy, lacks to existing
The problems such as the considerations of product and high method complexity, the present invention provide the social networks marketing seed for considering that existing product influences
User's choosing method, to marketing influence in terms of user social contact network topology attribute, intimate degree and existing product three
Careful measurement is carried out, income is substituted with return rate, proposes a kind of Efficient Solution algorithm towards extensive social networks, is solved
Promotion problem on large-scale social networks obtains highest income with the smallest expense.
A kind of social networks marketing seed user choosing method for considering existing product and influencing, comprising:
Social networks is modeled as a width socialgram G=(V, E, W) by step 1, and wherein vertex set V indicates social networks
In all social users' set, side collection E indicates social networks set between social user, and W indicates the weight set on all sides；
Step 2, initialization seed user's set S are empty set, i.e.,Initialising subscriber set U is institute in step 1
There are user's set, i.e. U ← V；
Step 3, the marketing influence for calculating all users in set U in step 2；
Step 4, to two set S and U in step 2 respectively pass through addition user and delete user be updated.
The social networks marketing seed user choosing method that above-mentioned consideration existing product influences, considers existing product to marketing
The influence of product promotion measures the marketing influence of social user in terms of social networks and existing product two, with battalion
Return rate is sold as optimization aim, is gathered using bidirectional optimistic method selected seed user, to maximize marketing return rate；This hair
Bright method includes four steps: building is based on interactive social networks model, the initialization of seed user set, marketing influence
It calculates and seed user set updates；The present invention optimizes marketing return rate rather than marketing income, avoids the problem of budget formulation,
Make marketing influence measurement more accurate the considerations of spending intimate between existing product and user, bidirectional optimistic method is only right
All users, which do, once judges whether it is seed user, and one seed user needs of the every selection of existing method are useful to institute
Family judgement is primary, and in comparison, method of the invention is decreased obviously complexity, answers suitable for the marketing of extensive social networks
With, and higher return rate can be obtained.
In other one embodiment, " social networks is modeled as a width socialgram G=(V, E, W), wherein vertex
Collecting V indicates that all social user's set, side collection E indicate social networks set between social user in social networks, and W indicates all sides
On weight set；" specifically include:
Users all in social networks are established into social side according to social networks each other, are denoted as euv, euvWith direction, table
Show the relationship from user u to v；Social networks nondirectional for social networks, can be considered as the special case of digraph, any
Two directed edge e are established between two users u and vuvAnd evu；
By the neighbours of any user u be defined as with the related side u, according to the direction on side, there are two neighbours for user u tool
Collection, in-degree neighbours collect Nu -=v ∈ V | evu∈ E } and out-degree neighbours collection Nu +=v ∈ V | euv∈E}；
Side euvOn weight wuvThe intimate degree for indicating user u to v is measured, all weights with the frequency of interaction of u to v
Constitute set W.
In other one embodiment, " step 3, the marketing influence for calculating all users in set U in step 2；"
It specifically includes:
Construct between all products the interactional matrix H in marketing process, arbitrary element p in HAB∈ [- 1,1], table
Show influence when having possessed product B to purchase product A is continued, pABWhen > 0, illustrate that existing procucts B has the popularization of marketing product A
Facilitation, pABWhen < 0, then inhibition is played, and pABWhen=0, illustrate that B does not influence the marketing of A；pABValue by pair
Product sales histories record statistics obtains；
For network marketing, marketing influence is defined as any user u for product A successful referral to its out-degree neighbours
The probability of v:
Wherein,Indicate effect of the user social contact relationship in marketing, Pv indicates the product collection that user v has possessed
It closes,Indicate effect of the existing procucts to marketing, α and β then indicate the weight of two kinds of effects, and meet alpha+beta=1, use
It promotes the sale of products total marketing influence of A at family are as follows:
In other one embodiment, " step 4, to two set S and U in step 2 respectively pass through addition user and
User is deleted to be updated." specifically include:
Step 401 judges whether set S and U are equal, if equal, output set S (or U) is used as the seed of marketing
Family set；
Step 402 is subordinated to user and collects and U but is not belonging to arbitrarily select a user u in seed user collection S, i.e. u ∈ S U,
According to the marketing influence of user in step 3, calculates separately addition u to S and delete the marginal returnses rate of u from U, be denoted as Δ+
And Δ-；
Step 403, with probability Δ+/(Δ++Δ-) u is added to set S, i.e. S ← S ∪ { u }, otherwise from user's set U
Middle deletion u, i.e. U ← U { u }, go to step 401；
In other one embodiment, in the step 402, adds u to S and delete the marginal returnses rate of u from U
Calculation method are as follows:
Market income σA(S) it is defined as the product of the number of users successfully marketed and the price of product A:
σA(S)=| πA(S)|pA (3)
Wherein, πA(S) the user's set successfully marketed is indicated, operator | | indicate set sizes, pAIndicate product A
Price；σA(S) there is submodularity and monotonicity, value increases with the growth of seed set S, and amplification is gradually reduced；
Market disbursement cA(S) it is defined as paying the expense of seed user and pays all products of successfully marketing
The sum of the expense of user:
Wherein T indicates users' set of all products of successfully marketing, in the set each user product is successfully promoted to
Its at least one out-degree neighbour, set sizes calculate as follows:
WhereinIndicate that user successfully promotes product A to its at least one out-degree neighbour's
Probability；cA(S) there is submodularity and monotonicity, value increases with the growth of seed set S, and amplification is gradually reduced；
Will marketing return rate be defined as unit costs bring income, with marketing income and market disbursement ratio come
It calculates, is defined as follows:
oA(S) there is submodularity but do not have monotonicity；
It calculates and user u is added to caused by seed user set S by the size variation of successfully marketing user set are as follows:
Formula (7) are substituted into formula (6), the marginal returnses rate for calculating addition u to S is the difference of addition front and back return rate:
oA(u | S)=oA(SU{u})-oA(S) (8)
To avoid the occurrence of negative value, marginal returnses rate Δ+Only take positive number:
Δ+=max { oA(u | S), 0 } (9)
Equally, calculating and deleting the marginal returnses rate of u from U is to delete the difference of front and back return rate:
oA(u | U { u })=oA(U)-oA(U\{u}) (10)
Marginal returnses rate Δ-Only take positive number:
Δ-=max { oA(u | U { u }), 0 } (11)
In other one embodiment, in the step 403, with probability Δ+/(Δ++Δ-) u is added to set S's
Method are as follows:
The random number r between one 0 to 1 is generated, Δ is calculated+/(Δ++Δ-), compare the two size, if Δ+/(Δ++
Δ-) < r, then u is added to set S, otherwise, deletes u from user's set U.
In other one embodiment, | πA(S) | calculation method are as follows: each user attempts to it in seed set S
Out-degree neighbours promote the sale of products, and the probability cracked a prospect is its marketing influence, are calculated by formula (1), its out-degree if cracking a prospect
Neighbours continue to promote according to the same manner, until successfully being promoted there is no new user, the use successfully marketed at this time
Amount is | πA(S)|。
A kind of computer equipment can be run on a memory and on a processor including memory, processor and storage
The step of computer program, the processor realizes any one the method when executing described program.
A kind of computer readable storage medium, is stored thereon with computer program, realization when which is executed by processor
The step of any one the method.
A kind of processor, the processor is for running program, wherein described program executes described in any item when running
Method.
Detailed description of the invention
Fig. 1 is the social networks marketing exemplary diagram in background technique.
Fig. 2 is the existing initial point selection algorithm execution flow chart in background technique.
Fig. 3 is that a kind of social networks marketing seed user for considering that existing product influences provided by the embodiments of the present application is chosen
The flow chart of method.
Fig. 4 is that the marketing profit return rate of algorithms of different under the application emulation experiment compares figure.
Fig. 5 is that the runing time of algorithms of different under the application the Realization of Simulation compares figure.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
Refering to Fig. 3, a kind of social networks marketing seed user choosing method for considering existing product and influencing, comprising:
Social networks is modeled as a width socialgram G=(V, E, W) by step 1, and wherein vertex set V indicates social networks
In all social users' set, side collection E indicates social networks set between social user, and W indicates the weight set on all sides；
Step 2, initialization seed user's set S are empty set, i.e.,Initialising subscriber set U is institute in step 1
There are user's set, i.e. U ← V；
Step 3, the marketing influence for calculating all users in set U in step 2；
Step 4, to two set S and U in step 2 respectively pass through addition user and delete user be updated.
The social networks marketing seed user choosing method that above-mentioned consideration existing product influences, considers existing product to marketing
The influence of product promotion measures the marketing influence of social user in terms of social networks and existing product two, with battalion
Return rate is sold as optimization aim, is gathered using bidirectional optimistic method selected seed user, to maximize marketing return rate；This hair
Bright method includes four steps: building is based on interactive social networks model, the initialization of seed user set, marketing influence
It calculates and seed user set updates；The present invention optimizes marketing return rate rather than marketing income, avoids the problem of budget formulation,
Make marketing influence measurement more accurate the considerations of spending intimate between existing product and user, bidirectional optimistic method is only right
All users, which do, once judges whether it is seed user, and one seed user needs of the every selection of existing method are useful to institute
Family judgement is primary, and in comparison, method of the invention is decreased obviously complexity, answers suitable for the marketing of extensive social networks
With, and higher return rate can be obtained.
In other one embodiment, " social networks is modeled as a width socialgram G=(V, E, W), wherein vertex
Collecting V indicates that all social user's set, side collection E indicate social networks set between social user in social networks, and W indicates all sides
On weight set；" specifically include:
Users all in social networks are established into social side according to social networks each other, are denoted as euv, euvWith direction, table
Show the relationship from user u to v；Social networks nondirectional for social networks, can be considered as the special case of digraph, any
Two directed edge e are established between two users u and vuvAnd evu；
By the neighbours of any user u be defined as with the related side u, according to the direction on side, there are two neighbours for user u tool
Collection, in-degree neighbours collect Nu -=v ∈ V | evu∈ E } and out-degree neighbours collection Nu +=v ∈ V | euv∈E}；
Side euvOn weight wuvThe intimate degree for indicating user u to v is measured, all weights with the frequency of interaction of u to v
Constitute set W.
In other one embodiment, " step 3, the marketing influence for calculating all users in set U in step 2；"
It specifically includes:
Construct between all products the interactional matrix H in marketing process, arbitrary element p in HAB∈ [- 1,1], table
Show influence when having possessed product B to purchase product A is continued, pABWhen > 0, illustrate that existing procucts B has the popularization of marketing product A
Facilitation, pABWhen < 0, then inhibition is played, and pABWhen=0, illustrate that B does not influence the marketing of A；pABValue by pair
Product sales histories record statistics obtains；
For network marketing, marketing influence is defined as any user u for product A successful referral to its out-degree neighbours
The probability of v:
Wherein,Indicate effect of the user social contact relationship in marketing, Pv indicates the product collection that user v has possessed
It closes,Indicate effect of the existing procucts to marketing, α and β then indicate the weight of two kinds of effects, and meet alpha+beta=1, use
It promotes the sale of products total marketing influence of A at family are as follows:
In other one embodiment, " step 4, to two set S and U in step 2 respectively pass through addition user and
User is deleted to be updated." specifically include:
Step 401 judges whether set S and U are equal, if equal, output set S (or U) is used as the seed of marketing
Family set；
Step 402 is subordinated to user and collects and U but is not belonging to arbitrarily select a user u in seed user collection S, i.e. u ∈ S U,
According to the marketing influence of user in step 3, calculates separately addition u to S and delete the marginal returnses rate of u from U, be denoted as Δ+
And Δ-；
Step 403, with probability Δ+/(Δ++Δ-) u is added to set S, i.e. S ← S ∪ { u }, otherwise from user's set U
Middle deletion u, i.e. U ← U { u }, go to step 401；
In other one embodiment, in the step 402, adds u to S and delete the marginal returnses rate of u from U
Calculation method are as follows:
Market income σA(S) it is defined as the product of the number of users successfully marketed and the price of product A:
σA(S)=| πA(S)|pA (3)
Wherein, πA(S) the user's set successfully marketed is indicated, operator | | indicate set sizes, pAIndicate product A
Price；σA(S) there is submodularity and monotonicity, value increases with the growth of seed set S, and amplification is gradually reduced；
Market disbursement cA(S) it is defined as paying the expense of seed user and pays all products of successfully marketing
The sum of the expense of user:
Wherein T indicates users' set of all products of successfully marketing, in the set each user product is successfully promoted to
Its at least one out-degree neighbour, set sizes calculate as follows:
WhereinIndicate that user successfully promotes product A to its at least one out-degree neighbour's
Probability；cA(S) there is submodularity and monotonicity, value increases with the growth of seed set S, and amplification is gradually reduced；
Will marketing return rate be defined as unit costs bring income, with marketing income and market disbursement ratio come
It calculates, is defined as follows:
oA(S) there is submodularity but do not have monotonicity；
It calculates and user u is added to caused by seed user set S by the size variation of successfully marketing user set are as follows:
Formula (7) are substituted into formula (6), the marginal returnses rate for calculating addition u to S is the difference of addition front and back return rate:
oA(u | S)=oA(S∪{u})-oA(S) (8)
To avoid the occurrence of negative value, marginal returnses rate Δ+Only take positive number:
Δ+=max { oA(u | S), 0 } (9)
Equally, calculating and deleting the marginal returnses rate of u from U is to delete the difference of front and back return rate:
oA(u | U { u })=oA(U)-oA(U\{u}) (10)
Marginal returnses rate Δ-Only take positive number:
Δ-=max { oA(u | U { u }), o } (11)
In other one embodiment, in the step 403, with probability Δ+/(Δ++Δ-) u is added to set S's
Method are as follows:
The random number r between one 0 to 1 is generated, Δ is calculated+/(Δ++Δ-), compare the two size, if Δ+/(Δ++
Δ-) < r, then u is added to set S, otherwise, deletes u from user's set U.
In other one embodiment, | πA(S) | calculation method are as follows: each user attempts to it in seed set S
Out-degree neighbours promote the sale of products, and the probability cracked a prospect is its marketing influence, are calculated by formula (1), its out-degree if cracking a prospect
Neighbours continue to promote according to the same manner, until successfully being promoted there is no new user, the use successfully marketed at this time
Amount is | πA(S)|。
A concrete application scene of the invention is described below:
A kind of social networks marketing seed user choosing method for considering existing product and influencing, target is to select suitable kind
Child user maximizes the profit return rate of marketing.The flow chart of the method for the invention is as shown in Figure 3, comprising the following steps:
Social networks is modeled as a width socialgram G=(V, E, W) by step 1, and wherein vertex set V indicates social networks
In all social users' set, side collection E indicates social networks set between social user, and W indicates the weight set on all sides；
Step 2, initialization seed user's set S are empty set, i.e.,Initialising subscriber set U is institute in step 1
There are user's set, i.e. U ← V；
Step 3, the marketing influence for calculating all users in set U in step 2；
Step 4, to two set S and U in step 2 respectively pass through addition user and delete user be updated.
Further, in the step 1 socialgram G construction method are as follows:
Firstly, users all in social networks are established social side according to social networks each other, it is denoted as euv, euvWith side
To, relationship of the expression from user u to v, such as concern relation, so socialgram G is digraph.To keep unified, for social activity
The nondirectional social networks of relationship, can be considered as the special case of digraph, established between any two user u and v two it is oriented
Side euvAnd evu；
Secondly, by the neighbours of any user u be defined as with the related side u, according to the direction on side, there are two user u tools
Neighbours' collection, in-degree neighbours collect Nu -=v ∈ V | evu∈ E } and out-degree neighbours collection Nu +=v ∈ V | euv∈E}；
Again, side euvOn weight wuvThe intimate degree for indicating user u to v is measured, institute with the frequency of interaction of u to v
There is weight to constitute set W.
Further, in the step 3 user's marketing influence calculation method are as follows:
Firstly, the interactional matrix H in marketing process is constructed between all products, arbitrary element p in HAB∈[-1,
1], influence when having possessed product B to purchase product A is continued, p are indicatedABWhen > 0, illustrate that existing procucts B pushes away marketing product A
Extensively there are facilitation, pABWhen < 0, then inhibition is played, and pABWhen=0, illustrate that B does not influence the marketing of A.pABValue it is logical
It crosses and statistics acquisition is recorded to product sales histories.
Secondly, marketing influence is defined as any user u and goes out product A successful referral to it for network marketing
Spend the probability of neighbours v:
Wherein,Indicate effect of the user social contact relationship in marketing, Pv indicates the product collection that user v has possessed
It closes,Indicate effect of the existing procucts to marketing, α and β then indicate the weight of two kinds of effects, and meet alpha+beta=1.With
It promotes the sale of products total marketing influence of A at family are as follows:
Further, in the step 4 two set S and U renewal process are as follows:
Step 401 judges whether set S and U are equal, if equal, output set S (or U) is used as the seed of marketing
Family set；
Step 402 is subordinated to user and collects and U but is not belonging to arbitrarily select a user u in seed user collection S, i.e. u ∈ S U,
According to the marketing influence of user in step 3, calculates separately addition u to S and delete the marginal returnses rate of u from U, be denoted as Δ+
And Δ-；
Step 403, with probability Δ+/ u is added to set S, i.e. S ← S ∪ { u } by (Δ ++ Δ -), otherwise from user's set U
Middle deletion u, i.e. U ← U { u }, go to step 401；
Further, it in the step 402, adds u to S and deletes the calculation method of the marginal returnses rate of u from U are as follows:
Firstly, marketing income σA(S) it is defined as the product of the number of users successfully marketed and the price of product A:
σA(S)=| πA(S)|pA (3)
Wherein πA(S) the user's set successfully marketed is indicated, operator | | indicate set sizes, pAIndicate product A's
Price.|πA(S) | calculation method are as follows: each user attempts to promote the sale of products to its out-degree neighbours in seed set S, cracks a prospect
Probability be its marketing influence, by formula (1) calculate, if cracking a prospect its out-degree neighbours continue promoted according to the same manner,
Until successfully being promoted there is no new user, the number of users successfully marketed at this time is | πA(S)|。σA(S) there is son
Mould and monotonicity, value increases with the growth of seed set S, and amplification is gradually reduced.
It all is successfully sought secondly, marketing disbursement cA (S) is defined as paying the expense of seed user and pays
Sell the sum of the expense of user of product:
Wherein γ withIndicate that weight, T indicate user's set of all products of successfully marketing, each user will in the set
Product, which is successfully promoted, gives its at least one out-degree neighbour, and set sizes calculate as follows:
WhereinIndicate that user successfully promotes product A to its at least one out-degree neighbour's
Probability.CA (S) has submodularity and monotonicity, and value increases with the growth of seed set S, and amplification is gradually reduced.
Again, marketing return rate is defined as unit costs bring income, with marketing income and disbursement of marketing
Ratio calculates, and is defined as follows:
OA (S) has submodularity but does not have monotonicity.
User u is added to caused by seed user set S by the size variation of successfully marketing user set finally, calculating
Are as follows:
Formula (7) are substituted into formula (6), the marginal returnses rate for calculating addition u to S is the difference of addition front and back return rate:
oA(u | S)=oA(S∪{u})-oA(S) (8)
To avoid the occurrence of negative value, marginal returnses rate Δ+Only take positive number:
Δ+=max { oA(u | S), 0 } (9)
Equally, calculating and deleting the marginal returnses rate of u from U is to delete the difference of front and back return rate:
oA(u | U { u })=oA(U)-oA(U\{u}) (10)
Marginal returnses rate Δ-Only take positive number:
Δ-=max { oA(u | U { u }), 0 } (11)
Further, in the step 403, with probability Δ+/(Δ++Δ-) u is added to the method for set S are as follows:
The random number r between one 0 to 1 is generated, Δ is calculated+/(Δ++Δ-), compare the two size, if Δ+/(Δ++
Δ-) < r, then u is added to set S, otherwise, deletes u from user's set U.
In order to verify the validity of this method, a specific embodiment is provided by emulation experiment.Following setting is done in experiment:
Social network data collection by web crawler to Sina weibo grab obtain, data set include 21.3 general-purpose families and
9960000 concern relation sides, and interaction frequency records between user in 3 months by a definite date, and interaction does not occur in this time
Concern side is deleted from data set, only retains effectively concern side.Existing product number is set as | P |=10000, and each user
The average value for possessing the quantity of existing product is 300.Relationship is influenced between any pair of product to be indicated with number between one [- 1,1], if
The numerical value is not zero, then illustrates have an impact between product, and pin product of anchoring a tent is A, indicates the influential production of the distribution on A with set R
Product set, set sizes are | R |.Herein, weight is set as α=β=1 in formula (1), indicates relationship and existing product pair between user
The calculating of marketing influence plays phase same-action, and weight is set as γ=10 in formula (4),Expression is paid each seed and is used
The expense unit price at family be higher than pay successfully market product user expense unit price because marketing in seed user effect compared with
The user of success marketing product is bigger.CELFGreedy algorithm is the improvement to classical greedy algorithm, for solving budgetary restraints
Initial point selection problem, the marketing influence calculated in last round of initial point selection can be used in the process of implementation epicycle calculating,
To reduce classic algorithm complexity, the speed of service is improved.CELFGreedy algorithm is realized in an experiment, and by its
It is compared with the mentioned method of the present invention, it, can not be direct since CELFGreedy algorithm is only applicable to constrained optimization problem solving
Optimize for profit return rate, herein, budget be defined as seed user number, account for the x% of total number of users, x take from 0.001 to
A series of 100 values (such as x=0.001,0.005,0.01,0.05,0.1,0.5,1,5,10,20,50,100) calculate all pre-
Calculate lower marketing income, then find out profit return rate maximum value, compared with the result of the method for the present invention, experimental result such as Fig. 4 with
Shown in Fig. 5.On the influential product scale of marketing in existing product, with ratio | R |/| P | it indicates, Fig. 4 compared two methods
The marketing profit return rate obtained under different scales, with the growth of scale, on the influential existing product of marketing product by
Cumulative more, the quantity that each user possesses these products is consequently increased, and the specific gravity that existing product influences in marketing influence increases
Add, this is conducive to find out more potential seed users, and therefore, profit return rate also increases, but growth trend is with | R
|/| P | become larger and gradually slows down.For profit return rate, the two way method proposed by the invention ratio side CELFGreedy in experiment
Method is averagely higher by 1.3 times.Fig. 5 compared runing time of the two methods under different scales, the results showed that have to marketing product
The existing product scale of influence will not influence too much the runing time of method, because the difference of scale not will lead to battalion
Sell the variation of influence power calculating process and initial point selection process complexity.From the point of view of runing time, the present invention is mentioned in experiment
Two way method ratio CELFGreedy method out is fast 200 times average.
Each technical characteristic of embodiment described above can be combined arbitrarily, for simplicity of description, not to above-mentioned reality
It applies all possible combination of each technical characteristic in example to be all described, as long as however, the combination of these technical characteristics is not deposited
In contradiction, all should be considered as described in this specification.
The embodiments described above only express several embodiments of the present invention, and the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the inventive concept of the premise, various modifications and improvements can be made, these belong to protection of the invention
Range.Therefore, the scope of protection of the patent of the invention shall be subject to the appended claims.
Claims (10)
- The seed user choosing method 1. a kind of social networks for considering that existing product influences is marketed characterized by comprisingThe social networks is modeled as a width socialgram G=(V, E, W) by step 1, and wherein vertex set V indicates social networks In all social users' set, side collection E indicates social networks set between social user, and W indicates the weight set on all sides；Step 2, initialization seed user's set S are empty set, i.e.,Initialising subscriber set U is all users in step 1 Set, i.e. U ← V；Step 3, the marketing influence for calculating all users in set U in step 2；Step 4, to two set S and U in step 2 respectively pass through addition user and delete user be updated.
- The seed user choosing method 2. the social networks according to claim 1 for considering that existing product influences is marketed, it is special Sign is, " social networks is modeled as a width socialgram G=(V, E, W), wherein vertex set V indicates own in social networks Social user's set, side collection E indicate social networks set between social user, and W indicates the weight set on all sides；" specifically wrap It includes:Users all in social networks are established into social side according to social networks each other, are denoted as euv, euvWith direction, indicate from The relationship of family u to v；Social networks nondirectional for social networks, can be considered as the special case of digraph, use in any two Two directed edge e are established between family u and vuvAnd evu；By the neighbours of any user u be defined as with the related side u, according to the direction on side, there are two neighbours to collect for user u tool, enters It spends neighbours and collects Nu -=v ∈ V | evu∈ E } and out-degree neighbours collection Nu +=v ∈ V | euv∈E}；Side euvOn weight wuvThe intimate degree for indicating user u to v is measured with the frequency of interaction of u to v, and all weights are constituted Set W.
- The seed user choosing method 3. the social networks according to claim 1 for considering that existing product influences is marketed, it is special Sign is, " step 3, the marketing influence for calculating all users in set U in step 2；" specifically include:Construct between all products the interactional matrix H in marketing process, arbitrary element p in HAB∈ [- 1,1] is indicated Influence when possessing product B to purchase product A is continued, pABWhen > 0, illustrate that existing procucts B has promotion to the popularization of marketing product A Effect, pABWhen < 0, then inhibition is played, and pABWhen=0, illustrate that B does not influence the marketing of A；pABValue pass through to product Sales histories record statistics obtains；For network marketing, marketing influence is defined as any user u for product A successful referral to its out-degree neighbours v's Probability:Wherein,Indicating effect of the user social contact relationship in marketing, Pv indicates the product set that user v has possessed,Indicate effect of the existing procucts to marketing, α and β then indicate the weight of two kinds of effects, and meet alpha+beta=1, and user pushes away Sell total marketing influence of product A are as follows:
- The seed user choosing method 4. the social networks according to claim 1 for considering that existing product influences is marketed, it is special Sign is, " step 4 passes through addition user respectively to two set S and U in step 2 and deletes user and is updated." specific Include:Step 401 judges whether set S and U are equal, if equal, seed user collection of the output set S (or U) as marketing It closes；Step 402 is subordinated to user and collects and U but is not belonging to arbitrarily select a user u in seed user collection S, i.e. u ∈ S U, according to The marketing influence of user in step 3 calculates separately addition u to S and deletes the marginal returnses rate of u from U, is denoted as Δ+And Δ-；Step 403, with probability Δ+/(Δ++Δ-) u is added to set S, i.e. S ← S ∪ { u }, otherwise deleted from user's set U Except u, i.e. U ← U { u }, go to step 401.
- The seed user choosing method 5. the social networks according to claim 4 for considering that existing product influences is marketed, it is special Sign is, in the step 402, adds u to S and deletes the calculation method of the marginal returnses rate of u from U are as follows:Market income σA(S) it is defined as the product of the number of users successfully marketed and the price of product A:σA(S)=| πA(S)|pA (3)Wherein, πA(S) the user's set successfully marketed is indicated, operator | | indicate set sizes, pAIndicate the valence of product A Lattice；σA(S) there is submodularity and monotonicity, value increases with the growth of seed set S, and amplification is gradually reduced；Market disbursement cA(S) it is defined as paying the expense of seed user and pays the use of all products of successfully marketing The sum of the expense at family:Wherein T indicates user's set of all products of successfully marketing, and each user successfully promotes product to it extremely in the set Few out-degree neighbours, set sizes calculate as follows:WhereinIndicate that user successfully promotes product A to the general of its at least one out-degree neighbour Rate；cA(S) there is submodularity and monotonicity, value increases with the growth of seed set S, and amplification is gradually reduced；Marketing return rate is defined as unit costs bring income, is counted with marketing income and the ratio of marketing disbursement It calculates, is defined as follows:oA(S) there is submodularity but do not have monotonicity；It calculates and user u is added to caused by seed user set S by the size variation of successfully marketing user set are as follows:Formula (7) are substituted into formula (6), the marginal returnses rate for calculating addition u to S is the difference of addition front and back return rate:oA(u | S)=oA(S∪{u})-oA(S) (8)To avoid the occurrence of negative value, marginal returnses rate Δ+Only take positive number:Δ+=max { oA(u | S), 0 } (9)Equally, calculating and deleting the marginal returnses rate of u from U is to delete the difference of front and back return rate:oA(u | U { u })=oA(U)-oA(U\{u}) (10)Marginal returnses rate Δ-Only take positive number:Δ-=max { oA(u | U { u }), 0 } (11).
- The seed user choosing method 6. the social networks according to claim 5 for considering that existing product influences is marketed, it is special Sign is, in the step 403, with probability Δ+/(Δ++Δ-) u is added to the method for set S are as follows:The random number r between one 0 to 1 is generated, Δ is calculated+/(Δ++Δ-), compare the two size, if Δ+/(Δ++Δ-) < U is then added to set S by r, otherwise, deletes u from user's set U.
- The seed user choosing method 7. the social networks according to claim 5 for considering that existing product influences is marketed, it is special Sign is, | πA(S) | calculation method are as follows: each user attempts to promote the sale of products to its out-degree neighbours in seed set S, promote at The probability of function is its marketing influence, is calculated by formula (1), its out-degree neighbours continue to push away according to the same manner if cracking a prospect Pin, until successfully being promoted there is no new user, the number of users successfully marketed at this time is | πA(S)|。
- 8. a kind of computer equipment including memory, processor and stores the meter that can be run on a memory and on a processor Calculation machine program, which is characterized in that the processor realizes any one of claims 1 to 7 the method when executing described program Step.
- 9. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is held by processor The step of any one of claims 1 to 7 the method is realized when row.
- 10. a kind of processor, which is characterized in that the processor is for running program, wherein right of execution when described program is run Benefit requires 1 to 7 described in any item methods.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810778152.2A CN108960929A (en) | 2018-07-16 | 2018-07-16 | Consider the social networks marketing seed user choosing method that existing product influences |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810778152.2A CN108960929A (en) | 2018-07-16 | 2018-07-16 | Consider the social networks marketing seed user choosing method that existing product influences |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108960929A true CN108960929A (en) | 2018-12-07 |
Family
ID=64481670
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810778152.2A Pending CN108960929A (en) | 2018-07-16 | 2018-07-16 | Consider the social networks marketing seed user choosing method that existing product influences |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108960929A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109492831A (en) * | 2018-12-20 | 2019-03-19 | 平安科技（深圳）有限公司 | Recommended method, device, medium and the electronic equipment in marketing path |
| CN109711871A (en) * | 2018-12-13 | 2019-05-03 | 北京达佳互联信息技术有限公司 | A kind of potential customers determine method, apparatus, server and readable storage medium storing program for executing |
| CN110223125A (en) * | 2019-06-18 | 2019-09-10 | 东北大学 | User location acquisition methods under the income algorithm of node location core side |
| CN110750721A (en) * | 2019-10-21 | 2020-02-04 | 秒针信息技术有限公司 | Information pushing method and device, electronic equipment and readable storage medium |
-
2018
- 2018-07-16 CN CN201810778152.2A patent/CN108960929A/en active Pending
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109711871A (en) * | 2018-12-13 | 2019-05-03 | 北京达佳互联信息技术有限公司 | A kind of potential customers determine method, apparatus, server and readable storage medium storing program for executing |
| CN109492831A (en) * | 2018-12-20 | 2019-03-19 | 平安科技（深圳）有限公司 | Recommended method, device, medium and the electronic equipment in marketing path |
| CN109492831B (en) * | 2018-12-20 | 2023-12-29 | 平安科技（深圳）有限公司 | Marketing path recommendation method and device, medium and electronic equipment |
| CN110223125A (en) * | 2019-06-18 | 2019-09-10 | 东北大学 | User location acquisition methods under the income algorithm of node location core side |
| CN110223125B (en) * | 2019-06-18 | 2021-06-11 | 东北大学 | User position obtaining method under node position kernel-edge profit algorithm |
| CN110750721A (en) * | 2019-10-21 | 2020-02-04 | 秒针信息技术有限公司 | Information pushing method and device, electronic equipment and readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108960929A (en) | Consider the social networks marketing seed user choosing method that existing product influences | |
| Cai et al. | Real-time bidding by reinforcement learning in display advertising | |
| Zhang et al. | Incentive mechanism for proximity-based mobile crowd service systems | |
| JP5529873B2 (en) | Method, system and computer-readable recording medium for billing for online advertising | |
| CN108460618A (en) | A kind of resource allocation method and device, electronic equipment | |
| US20140188584A1 (en) | Product pricing system on electronic commerce using the internet | |
| CN108492146A (en) | Preferential value calculating method, server-side and client based on user-association behavior | |
| CN111582947A (en) | Coupon processing method and related device | |
| CN106875205A (en) | A kind of object selection method and device | |
| WO2020244468A1 (en) | Financial product recommendation method and apparatus, and electronic device and computer storage medium | |
| CN109242608A (en) | A kind of e-commerce system based on big data and cloud computing | |
| WO2022121957A1 (en) | Order output method and apparatus, computer device, and computer-readable storage medium | |
| US20120078730A1 (en) | Automatic Internet Search Advertising Campaign Variable Optimization for Aiding Advertising Agency Efficiencies | |
| Nabipay et al. | Flat versus metered rates, bundling, and bandwidth hogs | |
| Wu | Matching value and market design in online advertising networks: An empirical analysis | |
| Zhang et al. | Online auction-based incentive mechanism design for horizontal federated learning with budget constraint | |
| TW202034254A (en) | Business area management system and method | |
| Chen et al. | Gacha game analysis and design | |
| CN109886751A (en) | A kind of monthly concentration Bidding Mechanism method for solving based on Cooperative Evolutionary Algorithm | |
| CN105654342A (en) | Method for selecting initial users enabling social network cooperative influence maximization | |
| JP6143930B1 (en) | Marketing support method, program, computer storage medium, and marketing support system | |
| Cai et al. | Double auction for a data trading market with preferences and conflicts of interest | |
| KR102513416B1 (en) | Simultaneous multiple power trading method in a microgrid composed of M producers and N consumers | |
| Benzell et al. | Regulating digital platform monopolies: The case of facebook | |
| CN101751611A (en) | Channel remuneration settlement method, device and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181207 |
|
| RJ01 | Rejection of invention patent application after publication |