CN114387064A - E-commerce platform potential customer recommendation method and system based on comprehensive similarity - Google Patents
E-commerce platform potential customer recommendation method and system based on comprehensive similarity Download PDFInfo
- Publication number
- CN114387064A CN114387064A CN202210039838.6A CN202210039838A CN114387064A CN 114387064 A CN114387064 A CN 114387064A CN 202210039838 A CN202210039838 A CN 202210039838A CN 114387064 A CN114387064 A CN 114387064A
- Authority
- CN
- China
- Prior art keywords
- client
- clients
- similarity
- group
- core
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 37
- 239000011159 matrix material Substances 0.000 claims abstract description 29
- 239000013598 vector Substances 0.000 claims abstract description 23
- 230000006870 function Effects 0.000 claims description 28
- 238000004364 calculation method Methods 0.000 claims description 23
- 238000005295 random walk Methods 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 10
- 230000007704 transition Effects 0.000 claims description 6
- 238000009826 distribution Methods 0.000 claims description 2
- 239000013604 expression vector Substances 0.000 claims description 2
- 238000005070 sampling Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 8
- 238000004590 computer program Methods 0.000 description 7
- 238000012545 processing Methods 0.000 description 4
- 238000003860 storage Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Images
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
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a method and a system for recommending potential customers of an e-commerce platform based on comprehensive similarity. The method comprises the following steps: selecting a motif, finding out all instances of the motif, constructing a high-order adjacency matrix of a client, then obtaining a client sequence, obtaining a representation vector of the client, and further calculating the contact compactness between the clients; calculating the similarity degree between the characteristics carried by the client; weighting the contact compactness and the characteristic similarity to obtain the comprehensive similarity of the client, and further obtaining a KNN (K nearest neighbor) graph; selecting the client with the largest core from the clients which are not distributed with the group as a core client; taking the selected core clients as initial groups, and sequentially adding the clients which enable the group fitness function increment to be maximum in the group neighbor clients into the groups; and repeating the core customer selection and the group expansion until all the customers in the e-commerce platform customer network belong to the group. The invention can effectively excavate the group existing in the E-commerce platform customer network and recommend the commodity with the largest purchasing frequency for the group.
Description
Technical Field
The invention relates to a method and a system for recommending potential customers of an e-commerce platform based on comprehensive similarity.
Background
Social members gradually form a certain stable relationship due to interaction in activities such as work, study, life, entertainment and the like, and further form a social network. With the rapid development of internet technology, people introduce the concept of early social networks into the internet, and an online social network facing social network services is created. Representative products of online social networks are WeChat, microblog, Taobao at home, and Facebook, Twitter, etc. at foreign countries. The explosion of online social networks has greatly changed people's lifestyles, e.g., online shopping has become the mainstream shopping, and over 80% of online customers often use online shopping. The network users can use the social network to make friends, play games, interact and cooperate with customers all over the world, and the distance between people is increased. By analyzing the social network, the method for recommending friends with similar characteristics and the same group for the user can enhance the stickiness of the user to software and improve the economic benefit. The recommendation system of the e-commerce platform is an important technology for improving the marketing level at present, and by dividing the customers, the recommendation system can effectively improve the recommendation success rate for recommending articles frequently purchased by the same group among the customers of the same group, so as to bring higher sales volume for the commodities. However, since the information of the e-commerce platform client network is very complex, it is difficult to directly analyze the information. With the progress of research, a plurality of potential customer recommendation methods for the multi-provider platform are generated. The method enables the structure of the network to be clearer by dividing the customers in the e-commerce customer network into groups with similar characteristics. However, the existing method still has several key problems: the first is that the information is not fully utilized, some methods ignore high-order structures in the network, and some methods can only use the type attribute; the second is that there is sometimes a mismatch between the topology and the client attributes in the network, and these methods cannot adaptively adjust the degree of contribution between the topology and the client attributes.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provides a method and a system for recommending potential customers of an e-commerce platform based on comprehensive similarity, which can effectively dig out groups existing in a customer network of the e-commerce platform and recommend commodities with the largest purchase frequency for the groups.
In order to achieve the purpose, the technical scheme of the invention is as follows: a comprehensive similarity-based E-commerce platform potential customer recommendation method comprises the following steps:
step S1, firstly, selecting a motif, finding out all instances of the motif in the social network, and constructing a high-order adjacency matrix of a client; then, random walk is respectively carried out on a low-order adjacent matrix and a high-order adjacent matrix of the client to obtain a client sequence; finally, training a client sequence by using a Skip-gram model to obtain a representation vector of the client, and calculating the contact compactness between the clients according to the representation vector of the client;
step S2, calculating the similarity between the characteristics carried by the client;
step S3, weighting the contact compactness and the characteristic similarity of the client by using a self-adaptive weighting strategy to obtain the comprehensive similarity of the client; then, reserving K clients with the maximum similarity for each client according to the comprehensive similarity of the clients as neighbors, and generating a KNN graph with the comprehensive similarity as a weight;
step S4, calculating the core of the client, and selecting the client with the largest core from the clients which are not distributed with the group as the core client;
step S5, the core customers selected in the step S4 are used as initial groups, the customers which enable the group fitness function increment to be maximum in the group neighbor customers are added into the groups in sequence, and the process is repeated until the fitness function can not be increased continuously;
and S6, repeating the steps S5 and S6 until all the customers in the E-commerce platform customer network belong to the group, finding out certain type of commodities which are purchased most frequently in the same group, and recommending the certain type of commodities to other customers in the group.
The invention also provides an e-commerce platform potential customer recommendation system based on the comprehensive similarity, which comprises the following steps:
the client contact closeness calculation module is used for firstly selecting a motif, finding out all instances of the motif in a social network and constructing a high-order adjacency matrix of a client; then, random walk is respectively carried out on a low-order adjacent matrix and a high-order adjacent matrix of the client to obtain a client sequence; finally, training a client sequence by using a Skip-gram model to obtain a representation vector of the client, and calculating the contact compactness between the clients according to the representation vector;
the client characteristic similarity calculation module is used for calculating the similarity degree between the characteristics carried by the client;
the client comprehensive similarity KNN graph generation module is used for weighting the contact compactness and the characteristic similarity of the client by using a self-adaptive weighting strategy to obtain the comprehensive similarity of the client; then, reserving K clients with the maximum similarity for each client according to the comprehensive similarity of the clients as neighbors, and generating a KNN graph with the comprehensive similarity as a weight;
the core client selection module is used for calculating the core of the client and selecting the client with the largest core from the clients which are not distributed with groups as a core client;
the client group expansion module is used for taking the selected core clients as initial groups, sequentially adding the clients which enable the group fitness function to be maximum in group neighbor clients into the groups, and repeating the process until the fitness function cannot be continuously increased;
and the commodity recommending module is used for finding out certain commodities which are purchased most frequently in the same group and recommending the certain commodities to other customers in the group.
Compared with the prior art, the invention has the following beneficial effects:
(1) the invention provides a self-adaptive weighting strategy, which can self-adaptively adjust the contribution degree between the network topology structure and the client attribute according to the self clustering coefficient of the network, and avoid fussy parameter adjustment.
(2) The invention provides the comprehensive similarity of the clients, combines various information in the network, designs the calculation strategy of the core of the clients, can effectively find out the core clients in the network and improves the accuracy of group division.
Drawings
FIG. 1 is a flow chart of an embodiment of the present invention.
Detailed Description
The technical scheme of the invention is specifically explained below with reference to the accompanying drawings.
As shown in fig. 1, the present embodiment provides an e-commerce platform potential customer recommendation method based on comprehensive similarity, which is characterized by including the following steps:
step S1, firstly, selecting a motif, finding out all instances of the motif in the social network, and constructing a high-order adjacency matrix of a client; then, random walk is respectively carried out on a low-order adjacent matrix and a high-order adjacent matrix of the client to obtain a client sequence; finally, training a client sequence by using a Skip-gram model to obtain a representation vector of the client, and calculating the contact compactness between the clients according to the representation vector of the client;
step S2, calculating the similarity between the characteristics carried by the client;
step S3, weighting the contact compactness and the characteristic similarity of the client by using a self-adaptive weighting strategy to obtain the comprehensive similarity of the client; then, reserving K clients with the maximum similarity for each client according to the comprehensive similarity of the clients as neighbors, and generating a KNN graph with the comprehensive similarity as a weight;
step S4, calculating the core of the client, and selecting the client with the largest core from the clients which are not distributed with the group as the core client;
step S5, the core customers selected in the step S4 are used as initial groups, the customers which enable the group fitness function increment to be maximum in the group neighbor customers are added into the groups in sequence, and the process is repeated until the fitness function can not be increased continuously;
and S6, repeating the steps S5 and S6 until all the customers in the E-commerce platform customer network belong to the group, finding out certain type of commodities which are purchased most frequently in the same group, and recommending the certain type of commodities to other customers in the group.
In this example, the step S1 is implemented as follows:
step S11, selecting a motif, searching all instances of the motif in the network, and generating a high-order adjacency matrix according to the found instances of the motif; wherein, the calculation formula is shown as formula (1):
wherein M represents the type of motif,
representing the number of times that client u and client v appear in the type motif instance;
step S12, respectively performing random walk on the low-order adjacency matrix and the high-order adjacency matrix to obtain a client sequence; each client can execute a plurality of walks, and each walk generates a client sequence with fixed length; random walk process randomly and uniformly selects initial clients v from networkiThen a neighbor client v is selected by taking the transition probability of the client as the probability of random selectionjAdding the current sequence; then, continue to select vjThe neighbor client adds the sequence and repeats the process until the sequence length reaches a preset value; for the low-order adjacency matrix, the transition probabilities to the neighbor clients are all equal; for a high-order adjacency matrix, the transition probability to a neighbor client is equal to the edge weight of the neighbor client divided by the total edge weight of the client and all neighbors;
s13, putting all the random walk sequences of the client into a Skip-gram model for training to generate an expression vector of the client; training the client sequence using the Skip-gram model, there are the following objective functions for each client pair:
wherein
Is v isiIs representative of the center vector of (a),
is v isjIs used to represent the context vector of (a),
the method comprises the following steps of (1) obtaining an empirical distribution of random sampling of negative samples, wherein sigma is a sigmoid function, and K is the number of the negative samples;
step S14, calculating the contact closeness between the clients for each client, wherein the calculation formula of the contact closeness is shown in formula (3):
wherein v is1,v2Representing any two clients in the network,
a corresponding customer vector is represented that represents the customer vector,
representing customer vectors
Cosine similarity between them.
In this example, the step S2 is implemented as follows:
calculating the characteristic similarity of the client, wherein the calculation method is shown as formula (4):
wherein S isc(u, v) similarity of Category attributes, Sn(u, v) represents the similarity of the numerical attributes, and the calculation formulas are respectively as follows:
wherein, TujIndicating that client u owns the attribute j, η (T)uj，Tvj) Calculating the number of attributes, s, having the same type between client u and client v1Is the number of categorical attributes, s2σ is a hyperparameter for the number of numeric attributes.
In this example, the step S3 is implemented as follows:
step S31, calculating the comprehensive similarity between the clients, wherein the calculation formula is as follows:
CS(u，v)＝Cρ*Hos(u，v)+(1-C)ρ*S(u，v)#(8)
c is a network average clustering coefficient, Hos (u, v) is the connection and density of the client u and the client v, S (u, v) is the feature similarity of the client u and the client v, and rho is a hyper-parameter for adjusting the contribution degree;
and step S32, reserving K clients with the maximum comprehensive similarity as neighbors for each client, and generating a KNN graph with the comprehensive similarity as the edge weight.
In this example, the step S4 is implemented as follows:
step S41, calculating a customer centrality for each customer on the weighted KNN graph, wherein the calculation formula of the customer centrality is as follows:
where N (u) represents the neighbor set of customer u, and Seeds represents the set of seed customers that have been selected; d (u, v) represents the authorized Jaccard distance; the calculation method of the core of the client simultaneously considers the influence of the client in the network and the distance between the client and the selected core client;
and step S42, selecting the client with the largest core as the core client.
In this example, the step S5 is implemented as follows:
step S51, taking the core client as an initial group;
step S52, calculating the fitness function value increment of the group to the neighbor clients of the group, wherein the calculation formula is as follows:
wherein FcFitness function for Community c, Fc∪{v}Fitness function for post-community c for client v, FcThe specific formula is as follows:
wherein,
and
respectively representing the comprehensive similarity sum of the clients inside the group c and the comprehensive similarity sum of the external clients;
and step S53, selecting the neighbor client with the maximum fitness function increment to join the group.
In this example, the step S6 is implemented as follows:
step S61, counting the times of purchasing the same type of commodities by the customers in the same group;
step S62 is to recommend the product with the largest number of purchases for the customers of the same group.
The example also provides an e-commerce platform potential customer recommendation system based on comprehensive similarity, which includes:
the client contact closeness calculation module is used for firstly selecting a motif, finding out all instances of the motif in a social network and constructing a high-order adjacency matrix of a client; then, random walk is respectively carried out on a low-order adjacent matrix and a high-order adjacent matrix of the client to obtain a client sequence; finally, training a client sequence by using a Skip-gram model to obtain a representation vector of the client, and calculating the contact compactness between the clients according to the representation vector;
the client characteristic similarity calculation module is used for calculating the similarity degree between the characteristics carried by the client;
the client comprehensive similarity KNN graph generation module is used for weighting the contact compactness and the characteristic similarity of the client by using a self-adaptive weighting strategy to obtain the comprehensive similarity of the client; then, reserving K clients with the maximum similarity for each client according to the comprehensive similarity of the clients as neighbors, and generating a KNN graph with the comprehensive similarity as a weight;
the core client selection module is used for calculating the core of the client and selecting the client with the largest core from the clients which are not distributed with groups as a core client;
the client group expansion module is used for taking the selected core clients as initial groups, sequentially adding the clients which enable the group fitness function to be maximum in group neighbor clients into the groups, and repeating the process until the fitness function cannot be continuously increased;
and the commodity recommending module is used for finding out certain commodities which are purchased most frequently in the same group and recommending the certain commodities to other customers in the group.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The foregoing is directed to preferred embodiments of the present invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof, and the scope thereof is determined by the claims that follow. However, any simple modification, equivalent change and modification of the above embodiments according to the technical essence of the present invention are within the protection scope of the technical solution of the present invention.
Claims (8)
1. A method for recommending potential customers of an e-commerce platform based on comprehensive similarity is characterized by comprising the following steps:
step S1, firstly, selecting a motif, finding out all instances of the motif in the social network, and constructing a high-order adjacency matrix of a client; then, random walk is respectively carried out on a low-order adjacent matrix and a high-order adjacent matrix of the client to obtain a client sequence; finally, training a client sequence by using a Skip-gram model to obtain a representation vector of the client, and calculating the contact compactness between the clients according to the representation vector of the client;
step S2, calculating the similarity between the characteristics carried by the client;
step S3, weighting the contact compactness and the characteristic similarity of the client by using a self-adaptive weighting strategy to obtain the comprehensive similarity of the client; then, reserving K clients with the maximum similarity for each client according to the comprehensive similarity of the clients as neighbors, and generating a KNN graph with the comprehensive similarity as a weight;
step S4, calculating the core of the client, and selecting the client with the largest core from the clients which are not distributed with the group as the core client;
step S5, the core customers selected in the step S4 are used as initial groups, the customers which enable the group fitness function increment to be maximum in the group neighbor customers are added into the groups in sequence, and the process is repeated until the fitness function can not be increased continuously;
and S6, repeating the steps S5 and S6 until all the customers in the E-commerce platform customer network belong to the group, finding out certain type of commodities which are purchased most frequently in the same group, and recommending the certain type of commodities to other customers in the group.
2. The e-commerce platform potential customer recommendation method based on comprehensive similarity as claimed in claim 1, wherein the step S1 is implemented as follows:
step S11, selecting a motif, searching all instances of the motif in the network, and generating a high-order adjacency matrix according to the found instances of the motif; wherein, the calculation formula is shown as formula (1):
wherein M represents the type of motif,
representing the number of times that client u and client v appear in the type motif instance;
step S12, respectively performing random walk on the low-order adjacency matrix and the high-order adjacency matrix to obtain a client sequence; each client can execute a plurality of walks, and each walk generates a client sequence with fixed length; random walk process randomly and uniformly selects initial clients v from networkiThen a neighbor client v is selected by taking the transition probability of the client as the probability of random selectionjAdding the current sequence; then, continue to select vjThe neighbor client adds the sequence and repeats the process until the sequence length reaches a preset value; for the low-order adjacency matrix, the transition probabilities to the neighbor clients are all equal; for a high-order adjacency matrix, the transition probability to a neighbor client is equal to the edge weight of the neighbor client divided by the total edge weight of the client and all neighbors;
s13, putting all the random walk sequences of the client into a Skip-gram model for training to generate an expression vector of the client; training the client sequence using the Skip-gram model, there are the following objective functions for each client pair:
wherein
Is v isiIs representative of the center vector of (a),
is v isjIs used to represent the context vector of (a),
the method comprises the following steps of (1) obtaining an empirical distribution of random sampling of negative samples, wherein sigma is a sigmoid function, and K is the number of the negative samples;
step S14, calculating the contact closeness between the clients for each client, wherein the calculation formula of the contact closeness is shown in formula (3):
wherein v is1,v2Representing any two clients in the network,
a corresponding customer vector is represented that represents the customer vector,
representing customer vectors
Cosine similarity between them.
3. The e-commerce platform potential customer recommendation method based on comprehensive similarity as claimed in claim 2, wherein the step S2 is implemented as follows:
calculating the characteristic similarity of the client, wherein the calculation method is shown as formula (4):
wherein S isc(u, v) similarity of Category attributes, Sn(u, v) represents the similarity of the numerical attributes, and the calculation formulas are respectively as follows:
wherein, TujIndicating that client u owns the attribute j, η (T)uj，Tvj) Calculating the number of attributes, s, having the same type between client u and client v1Is the number of categorical attributes, s2σ is a hyperparameter for the number of numeric attributes.
4. The e-commerce platform potential customer recommendation method based on comprehensive similarity as claimed in claim 3, wherein the step S3 is implemented as follows:
step S31, calculating the comprehensive similarity between the clients, wherein the calculation formula is as follows:
CS(u，v)＝Cρ*Hos(u，v)+(1-C)ρ*S(u，v)#(8)
c is a network average clustering coefficient, Hos (u, v) is the connection and density of the client u and the client v, S (u, v) is the feature similarity of the client u and the client v, and rho is a hyper-parameter for adjusting the contribution degree;
and step S32, reserving K clients with the maximum comprehensive similarity as neighbors for each client, and generating a KNN graph with the comprehensive similarity as the edge weight.
5. The e-commerce platform potential customer recommendation method based on comprehensive similarity according to claim 4, wherein the step S4 is implemented as follows:
step S41, calculating a customer centrality for each customer on the weighted KNN graph, wherein the calculation formula of the customer centrality is as follows:
where N (u) represents the neighbor set of customer u, and Seeds represents the set of seed customers that have been selected; d (u, v) represents the authorized Jaccard distance; the calculation method of the core of the client simultaneously considers the influence of the client in the network and the distance between the client and the selected core client;
and step S42, selecting the client with the largest core as the core client.
6. The e-commerce platform potential customer recommendation method based on comprehensive similarity as claimed in claim 5, wherein the step S5 is implemented as follows:
step S51, taking the core client as an initial group;
step S52, calculating the fitness function value increment of the group to the neighbor clients of the group, wherein the calculation formula is as follows:
wherein FcFitness function for Community c, Fc∪{v}Fitness function for post-community c for client v, FcThe specific formula is as follows:
wherein,
and
respectively representing the comprehensive similarity sum of the clients inside the group c and the comprehensive similarity sum of the external clients;
and step S53, selecting the neighbor client with the maximum fitness function increment to join the group.
7. The e-commerce platform potential customer recommendation method based on comprehensive similarity as claimed in claim 1, wherein the step S6 is implemented as follows:
step S61, counting the times of purchasing the same type of commodities by the customers in the same group;
step S62 is to recommend the product with the largest number of purchases for the customers of the same group.
8. An e-commerce platform potential customer recommendation system based on comprehensive similarity is characterized by comprising:
the client contact closeness calculation module is used for firstly selecting a motif, finding out all instances of the motif in a social network and constructing a high-order adjacency matrix of a client; then, random walk is respectively carried out on a low-order adjacent matrix and a high-order adjacent matrix of the client to obtain a client sequence; finally, training a client sequence by using a Skip-gram model to obtain a representation vector of the client, and calculating the contact compactness between the clients according to the representation vector;
the client characteristic similarity calculation module is used for calculating the similarity degree between the characteristics carried by the client;
the client comprehensive similarity KNN graph generation module is used for weighting the contact compactness and the characteristic similarity of the client by using a self-adaptive weighting strategy to obtain the comprehensive similarity of the client; then, reserving K clients with the maximum similarity for each client according to the comprehensive similarity of the clients as neighbors, and generating a KNN graph with the comprehensive similarity as a weight;
the core client selection module is used for calculating the core of the client and selecting the client with the largest core from the clients which are not distributed with groups as a core client;
the client group expansion module is used for taking the selected core clients as initial groups, sequentially adding the clients which enable the group fitness function to be maximum in group neighbor clients into the groups, and repeating the process until the fitness function cannot be continuously increased;
and the commodity recommending module is used for finding out certain commodities which are purchased most frequently in the same group and recommending the certain commodities to other customers in the group.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210039838.6A CN114387064B (en) | 2022-01-13 | 2022-01-13 | Electronic commerce platform potential customer recommendation method and system based on comprehensive similarity |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210039838.6A CN114387064B (en) | 2022-01-13 | 2022-01-13 | Electronic commerce platform potential customer recommendation method and system based on comprehensive similarity |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114387064A true CN114387064A (en) | 2022-04-22 |
| CN114387064B CN114387064B (en) | 2024-07-19 |
Family
ID=81201179
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210039838.6A Active CN114387064B (en) | 2022-01-13 | 2022-01-13 | Electronic commerce platform potential customer recommendation method and system based on comprehensive similarity |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114387064B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150134416A1 (en) * | 2013-11-11 | 2015-05-14 | International Business Machines Corporation | Initial marketing campaign targets |
| CN110222272A (en) * | 2019-04-18 | 2019-09-10 | 广东工业大学 | A kind of potential customers excavate and recommended method |
| CN110517114A (en) * | 2019-08-21 | 2019-11-29 | 广州云徙科技有限公司 | A kind of information-pushing method and system based on community discovery algorithm |
| CN110930184A (en) * | 2019-11-14 | 2020-03-27 | 杭州天宽科技有限公司 | Potential customer mining and customer type selection method based on mixed recommendation algorithm |
| CN113159918A (en) * | 2021-04-09 | 2021-07-23 | 福州大学 | Bank client group mining method based on federal group penetration |
| CN113724042A (en) * | 2021-08-23 | 2021-11-30 | 中国建设银行股份有限公司 | Commodity recommendation method, commodity recommendation device, commodity recommendation medium and commodity recommendation equipment |
-
2022
- 2022-01-13 CN CN202210039838.6A patent/CN114387064B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150134416A1 (en) * | 2013-11-11 | 2015-05-14 | International Business Machines Corporation | Initial marketing campaign targets |
| CN110222272A (en) * | 2019-04-18 | 2019-09-10 | 广东工业大学 | A kind of potential customers excavate and recommended method |
| CN110517114A (en) * | 2019-08-21 | 2019-11-29 | 广州云徙科技有限公司 | A kind of information-pushing method and system based on community discovery algorithm |
| CN110930184A (en) * | 2019-11-14 | 2020-03-27 | 杭州天宽科技有限公司 | Potential customer mining and customer type selection method based on mixed recommendation algorithm |
| CN113159918A (en) * | 2021-04-09 | 2021-07-23 | 福州大学 | Bank client group mining method based on federal group penetration |
| CN113724042A (en) * | 2021-08-23 | 2021-11-30 | 中国建设银行股份有限公司 | Commodity recommendation method, commodity recommendation device, commodity recommendation medium and commodity recommendation equipment |
Non-Patent Citations (1)
| Title |
|---|
| 李杨: "基于客户喜好的双向个性化推荐算法", 计算机应用研究, vol. 38, no. 9, 15 September 2021 (2021-09-15), pages 2701 - 2709 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114387064B (en) | 2024-07-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Naruchitparames et al. | Friend recommendations in social networks using genetic algorithms and network topology | |
| Wang et al. | A mobile recommendation system based on logistic regression and gradient boosting decision trees | |
| Chamberlain et al. | Scalable hyperbolic recommender systems | |
| Ahn et al. | Facilitating cross-selling in a mobile telecom market to develop customer classification model based on hybrid data mining techniques | |
| Jiang et al. | An efficient evolutionary user interest community discovery model in dynamic social networks for internet of people | |
| CN107256241B (en) | Movie recommendation method for improving multi-target genetic algorithm based on grid and difference replacement | |
| CN111681067A (en) | Long-tail commodity recommendation method and system based on graph attention network | |
| CN106844637B (en) | Movie recommendation method for improving multi-target genetic algorithm based on orthogonal and clustering pruning | |
| CN109034960B (en) | Multi-attribute inference method based on user node embedding | |
| CN107657034A (en) | A kind of event social networks proposed algorithm of social information enhancing | |
| Navgaran et al. | Evolutionary based matrix factorization method for collaborative filtering systems | |
| CN113378470A (en) | Time sequence network-oriented influence maximization method and system | |
| CN109902823A (en) | A kind of model training method and equipment based on generation confrontation network | |
| Burtini et al. | Improving online marketing experiments with drifting multi-armed bandits | |
| CN113724096A (en) | Group knowledge sharing method based on public commodity evolution game model | |
| CN114387064B (en) | Electronic commerce platform potential customer recommendation method and system based on comprehensive similarity | |
| Hu et al. | Establishing Grey Criteria Similarity Measures for Multi-criteria Recommender Systems. | |
| CN113610608B (en) | User preference recommendation method and device, electronic equipment and storage medium | |
| CN110348879A (en) | For determining the method and device of user behavior value | |
| Tunali et al. | Multi-objective evolutionary product bundling: a case study | |
| CN109299368A (en) | A kind of method and system for the intelligent personalized recommendation of environmental information resource AI | |
| CN112559864B (en) | Bilinear graph network recommendation method and system based on knowledge graph enhancement | |
| CN115293815A (en) | Cross-platform e-commerce user alignment method based on user commodity interest | |
| Hu | Neighborhood-based Collaborative Filtering Using Grey Relational Analysis. | |
| Shakya et al. | Opposition-based genetic algorithm for community detection in social networks |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |