CN108898459A - A kind of Method of Commodity Recommendation and device - Google Patents
A kind of Method of Commodity Recommendation and device Download PDFInfo
- Publication number
- CN108898459A CN108898459A CN201810665009.2A CN201810665009A CN108898459A CN 108898459 A CN108898459 A CN 108898459A CN 201810665009 A CN201810665009 A CN 201810665009A CN 108898459 A CN108898459 A CN 108898459A
- Authority
- CN
- China
- Prior art keywords
- commodity
- end article
- purchase
- random walk
- history
- 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
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
Landscapes
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the invention provides a kind of Method of Commodity Recommendation and devices, are applied to computer science.Method of Commodity Recommendation provided by the present invention avoids the problem of recommending cold start-up and much-sought-after item excessively to recommend new user, new things, keep commercial product recommending more rationally more acurrate, consumer purchases goods Discussing Convenience is promoted, increases supermarket's business revenue, is also put for supermarket shelves commodity and effective reference is provided.This includes the following steps:Obtain history shopping record;It is done shopping and is recorded according to the history, established commodity and purchase relational network together;In the commodity in purchase relational network, using Random Walk Algorithm, the random walk sequence corresponding with commodity all in purchase relational network is calculated separately;According to the corresponding random walk sequence of the commodity, the degree of association between the end article and other commodity is determined;According to the degree of association between the end article and other commodity, the commodity for needing to recommend are determined.The present invention is applied to commercial product recommending.
Description
Technical field
The present invention relates to computer science more particularly to a kind of novel Method of Commodity Recommendation and device.
Background technique
It is increasingly mature with e-commerce with the popularity of the internet, people be increasingly utilized e-commerce platform into
The acquisition of row merchandise news and the purchase of commodity.It is desirable to when buying commodity, e-commerce platform can be recommended to user
Extensive stock information, such as may interested commodity etc. to user recommended user.By commercial product recommending, user can be shortened and sought
The path of product required for looking for promotes user experience.
Proposed algorithm is more to commonly use instantly and popular one of algorithm.Current main proposed algorithm mainly includes being based on
The recommendation of content, is based on correlation rule recommendation at collaborative filtering recommending.In content-based recommendation algorithm, pass through user first
Learn the interest of user to the evaluating characteristic of object, then calculates the matching journey between user interest and item characteristic attribute
Degree.The project for finally recommending matching degree high to user;Proposed algorithm based on collaborative filtering uses nearest neighbor algorithm meter first
It calculates and likes immediate neighbor user with target user, mesh is then predicted according to the weighted value that nearest neighbor evaluates commodity
User is marked to the fancy grade of particular commodity, commercial product recommending is finally carried out according to the fancy grade of user;Based on correlation rule
Recommend based on correlation rule, using commodity have been purchased as regular head, recommended is as rule body, by calculating two kinds of commodity
Same purchase probability target user is recommended.
The above common proposed algorithm all has some limitations:The hobby of content-based recommendation algorithm requirement user
It allows for being expressed with content characteristic, the feature of extraction, which should guarantee accuracy again, has certain practical significance, and not
The estimate of situation of other users can explicitly be obtained;Proposed algorithm core based on collaborative filtering is based on historical data, to new
User and new article have the problem of " cold start-up ", the effect of recommendation dependent on user's history preference data number and it is accurate
Property, and in most realization, user's history preference is to be stored with sparse matrix, and the calculating on sparse matrix has
A little obvious problems, including may the wrong preference of a minority can have a great impact to the accuracy of recommendation etc.;It is based on
The proposed algorithm calculation amount of correlation rule is larger, while there is also cold start-up and sparsity problems, and are easy to appear popular item
The problem of mesh is excessively recommended.Three kinds of algorithms all cannot be accurately rationally customer recommendation commodity in summary, and make things convenient for customers purchase
Object promotes supermarket's business revenue.
Summary of the invention
To overcome above-mentioned defect in the prior art, the technical problems to be solved by the invention are:A kind of commodity are provided to push away
Method and device is recommended, puts and reference is provided for electric business supermarket commercial product recommending or supermarket shelves cargo, customer is promoted and buys convenience
Degree promotes supermarket's business revenue.
In order to achieve the above objectives, the present invention adopts the following technical scheme that:
In a first aspect, the embodiment of the present application provides a kind of Method of Commodity Recommendation, including:Obtain history shopping record;Root
It does shopping and records according to the history, establish commodity and purchase relational network together;Wherein, the commodity include N number of quotient in relational network with purchasing
Product, N >=2;It is done shopping and is recorded according to the history, however, it is determined that two kinds of commodity were bought simultaneously, then were carried out described two commodity
Connection;It include end article in N number of commodity;In the commodity in purchase relational network, Random Walk Algorithm, difference are utilized
Calculate the corresponding random walk sequence of N number of commodity;According to the corresponding random walk sequence of the N number of commodity, determine described in
The degree of association between end article and other commodity;According to the degree of association between the end article and other commodity, determining is needed
The commodity to be recommended.
Optionally, this method further includes:According to the corresponding random walk sequence of the N number of commodity, calculated using Word2vec
Method calculates separately the real vector of N number of commodity;According to the real vector of N number of commodity, calculate the end article with
Cosine similarity in N number of commodity between other commodity.
Optionally, described to need the commodity recommended to include:Cosine similarity in N number of commodity with the end article
Greater than the commodity of predetermined threshold.The predetermined threshold can preferably limit commodity similar with end article, the predetermined threshold
Value is set as ε.
Optionally, the cosine similarity calculation formula is:
Wherein n indicates the vector dimension of the real vector of commodity,Indicate the real vector of end article,
The real vector for indicating other commodity in addition to end article in N number of commodity, and if only if predetermined threshold ε≤similarityxWhen,
RetainCorresponding product name, remaining removal.
Optionally, the history shopping record, including:It obtains in preset time period, purchase when different consumer purchases goods
Object record.
In Method of Commodity Recommendation provided by the present invention, by obtaining history shopping record, building commodity are same to purchase network of personal connections
Network obtains each commodity random walk sequence using Random Walk Algorithm, according to the corresponding random walk sequence of commodity, meter
The degree of association between end article and other commodity is calculated, according to the degree of association, the commodity for needing to recommend are determined, by the high quotient of similarity
Product recommend client.This Method of Commodity Recommendation efficiently avoids the problem of " cold start-up " of article, makes new user in purchase quotient
Reasonably recommended when product, and solve the problems, such as that popular project is excessively recommended, improved the accuracy of commercial product recommending, be
Electric business supermarket commercial product recommending or supermarket shelves put offer reference, promote customer and buy convenience degree, promote supermarket's business revenue.
Second aspect, the embodiment of the present application provide a kind of device for recommending the commodity, including:Module is obtained, is gone through for obtaining
History shopping record;Module is constructed, for establishing commodity with purchase relationship after the acquisition module obtains the history shopping record
Network;Wherein, the commodity include N number of commodity, N >=2 in relational network with purchasing;It is done shopping and is recorded according to the history, however, it is determined that
Two kinds of commodity were bought simultaneously, then were attached described two commodity;It include end article in N number of commodity;Sequence
Generation module purchases relational network the commodity are same for after the building module constructs the commodity with purchase relational network
In, using Random Walk Algorithm, respectively obtain the corresponding random walk sequence of N number of commodity;Determining module, for described
After sequence generating module obtains the random walk sequence, the degree of association between the end article and other commodity is determined;It pushes away
Module is recommended, for determining and needing after the degree of association that the determining module calculates between the end article and other commodity
The commodity of recommendation.
Optionally, the determining module specifically includes:Vector calculation unit and similarity calculated;The vector calculates
Unit, for calculating separately N number of quotient using Word2vec algorithm according to the corresponding random walk sequence of the N number of commodity
The real vector of product;The similarity calculated, for respectively obtaining the reality of N number of commodity in the vector calculation unit
After number vector, the cosine similarity in the end article and N number of commodity between other commodity is calculated.
Optionally, described to need the commodity recommended to include:Cosine similarity in N number of commodity with the end article
Greater than the commodity of predetermined threshold.
Optionally, the cosine similarity calculation formula of the similarity calculated is:
Wherein n indicates the vector dimension of the real vector of commodity,Indicate the real vector of end article,
The real vector for indicating other commodity in addition to end article in N number of commodity, and if only if predetermined threshold ε≤similarityxWhen,
RetainCorresponding product name, remaining removal.
Optionally, the history shopping record for obtaining module and obtaining, including:It obtains in preset time period, it is different
Shopping record when consumer purchases goods.
The device for recommending the commodity bring as provided by above-mentioned second aspect has the technical effect that based on above-mentioned first aspect
What provided Method of Commodity Recommendation was realized, therefore the device for recommending the commodity bring technical effect is brought with the Method of Commodity Recommendation
Technical effect it is identical, which is not described herein again.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
Other attached drawings are obtained according to these attached drawings.
Fig. 1 is the step schematic diagram of Method of Commodity Recommendation provided by the embodiment of the present invention;
Fig. 2 is the flow diagram of Method of Commodity Recommendation provided by the embodiment of the present invention;
A kind of Fig. 3 customer provided by the embodiment of the present invention buys the same purchase relation schematic diagram of commodity;
Fig. 4 another kind customer provided by the embodiment of the present invention buys the same purchase relational structure schematic diagram of commodity;
Fig. 5 is commodity with purchase relational network building schematic diagram;
Fig. 6 is device for recommending the commodity construction module schematic diagram of the present invention;
Fig. 7 is the structural schematic diagram of device for recommending the commodity determining module of the present invention.
Specific embodiment
In order to make the foregoing objectives, features and advantages of the present invention clearer and more comprehensible, implement below in conjunction with the present invention
Attached drawing in example, technical scheme in the embodiment of the invention is clearly and completely described.Obviously, described embodiment
Only a part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, the common skill in this field
Art personnel all other embodiment obtained under the premise of not having, making creative work belongs to what the present invention protected
Range.
It is described as follows for the technical term that embodiments herein is used:
Word2vec algorithm of the present invention is the open a software work for being used to train term vector of Google company
Tool, each of sentence word is fast and effeciently mapped to k according to given corpus, by the training pattern after optimization by it
In dimension space there is true value to obtain vector, core architecture includes CBOW and Skip-gram.CBOW is speculated from original statement
Target words;And Skip-gram is exactly the opposite, is to deduce original statement from target words.
The degree of association of the present invention is the correlation degree characterized between two things, mathematically refers to that two functions are similar
Degree.Similarity degree is the quantitative criteria for measuring similarity between article, and the highest commodity of similarity are customer's most probable one
The commodity of purchase are played, the method for calculating similarity has very much, and there are commonly based on the similarity of cosine (Cosine-based)
It calculates and is based on being associated with the similarity calculation of (Correlation-based).
Embodiment one:
The embodiment of the present invention provides a kind of novel Method of Commodity Recommendation, applied under the scene recommended commodity, is somebody's turn to do
Method includes the following steps:
As shown in Figure 1, being a kind of step schematic diagram of Method of Commodity Recommendation provided by the embodiments of the present application.In the schematic diagram
Including obtain customer purchase product name, building commodity with purchase relational network, the random walk sequence for obtaining each commodity,
Word2vec algorithm, the real vector for obtaining each commodity, cosine similarity calculate, tie the higher commodity of similarity
It ties up sale or recommends to user.Wherein, obtain customer purchase product name refer to, in the period of certain in, gone through by supermarket
History data server obtains the institute that all product names of customer A purchase, all product names of customer B purchase, customer C are bought
There is product name until all product names that customer M is bought;Building commodity refer to purchase relational network, are cared for according to described M
The shopping record of visitor, using all product names of M customer purchase as point, to appear in the quotient in same shopping list
Line between product is side, and composition commodity are the same as purchase relational network figure;The random walk sequence for obtaining each commodity refers to, according to
The commodity of building desire to make money or profit with purchase relational network and obtain all random walk sequences of commodity 1 with Random Walk Algorithm, commodity 2 own
Random walk sequence, all random walk sequence of commodity 3 is until commodity N all random walk sequence;Word2vec is calculated
Method is a opening for training the software tool of term vector;The real vector of each commodity refers to, according to each quotient
All random walk sequences of product calculate the real vector of commodity 1, the real vector of commodity 2, quotient using Word2vec algorithm
Real vector of the real vector of product 3 up to commodity N;Cosine similarity calculating refers to, according to the reality of each obtained commodity
Number vector calculates the cosine value in end article and N number of commodity between other commodity using cosine formula;Similarity is higher
Commodity carry out bundle sale or to user recommend refer to, calculate in end article and N number of commodity between other commodity
Cosine value be closer to the commodity of numerical value 1, the commodity and end article are subjected to bundle sale or to the use of purchase end article
Recommended at family.
Specifically it is described in detail with Fig. 2 example;
S101, history shopping record is obtained.
Illustratively, can be by obtaining in preset time period, the mode of shopping record when different consumer purchases goods,
Obtain history shopping record.For example, extracting the shopping record of all customers in some supermarket's half a year.
S102, building commodity purchase together relational network.
It is done shopping and is recorded according to the history, established commodity and purchase relational network together;Wherein, the commodity are the same as in purchase relational network
Including N number of commodity, N >=2;It is done shopping and is recorded according to the history, however, it is determined that two kinds of commodity were bought simultaneously, then will be described two
Commodity are attached;It include end article in N number of commodity.
Illustratively, if first customer has purchased commodity 1, commodity 2, commodity 3 and commodity in same shopping list
7, using four kinds of product names as vertex, commodity are attached two-by-two, i.e., commodity 1 are connect with commodity 2, commodity 1 and commodity 3 connect
It connects, commodity 1 are connect with commodity 7, commodity 2 are connect with commodity 3, commodity 2 are connect with commodity 7, commodity 3 connect with commodity 7 and obtain nothing
To commodity with purchase relational network schematic diagram, as shown in Figure 3;Second customer has purchased commodity 3, quotient in same shopping list
Product 6 and commodity 7, using three kinds of product names as vertex, commodity are attached two-by-two, i.e., commodity 3 connect with commodity 6, commodity 3
It is connect with commodity 7, commodity 6 connect with commodity 7 and obtain undirected commodity with purchase relational network schematic diagram, as shown in Figure 4.Similarly,
Three customers have purchased commodity 4 and commodity 6 in same shopping list, using described two product names as vertex, quotient two-by-two
Product are attached, i.e., commodity 4 are connect with commodity 6 obtains undirected commodity with purchase relational network schematic diagram;4th customer is same
It opens in shopping list and has purchased commodity 2 and commodity 4, using described two product names as vertex, commodity are attached two-by-two, i.e. quotient
Product 2 are connect with commodity 4 obtains undirected commodity with purchase relational network schematic diagram;5th customer buys in same shopping list
Commodity 4 and commodity 5, using described two product names as vertex, commodity are attached two-by-two, i.e., commodity 4 connect with commodity 5 and obtain
Undirected commodity are obtained with purchase relational network schematic diagram.(the undirected commodity of above-mentioned third to the 5th customer are the same as purchase relational network and the
One with the undirected commodity of second customer with purchase the generating mode of relational network it is identical, therefore be not shown) by this five
The commodity that position customer is bought carry out the combination of side collection, final to obtain with purchase commodity relation network diagram, i.e. Fig. 5.
S103, random walk sequence is obtained.
In the commodity in purchase relational network, using Random Walk Algorithm, calculate separately N number of commodity it is corresponding with
Machine migration sequence.
Illustratively, by taking Fig. 5 as an example, all nodes in the same purchase commodity relation network [commodity 1, commodity 2,
Commodity 3, commodity 4, commodity 5, commodity 6, commodity 7] in using equiprobable possibility select a node as random walk rise
Initial point, wherein the probability that each node is selected is 1/7, it is commodity 1 by the start node that random selection obtains random walk,
Traverse node sequence VS=(1);It is start node with commodity 1, in its neighbor node [commodity 2, commodity 3, commodity 7] equal probability
Destination node is selected, the selected probability of each neighbor node is 1/3, after random selection, destination node 2, traversal section
Point sequence VS=(1,2);For each step of random walk, destination node is all the neighbor node equal probability from present node
A randomly selected point, while destination node being added in traverse node sequence.
S104, the degree of association between end article and other commodity is determined.
According to the corresponding random walk sequence of the N number of commodity, the pass between the end article and other commodity is determined
Connection degree.
Specifically, step S104, may include:
The random walk sequence that S1041, basis obtain uses the Word2vec algorithm in natural language processing method to distinguish
Calculate the real vector of N number of commodity;
S1042, according to the real vector of N number of commodity, calculate other quotient in the end article and N number of commodity
Cosine similarity between product.
Similarity is to measure the quantitative criteria of similarity degree between article, and the highest commodity of similarity are customer's most probable one
The commodity of purchase are played, the method for calculating similarity has very much, and there are commonly based on the similarity of cosine (Cosine-based)
It calculates and is based on being associated with the similarity calculation of (Correlation-based).
Specifically, the present invention measures the similitude between commodity, cosine similarity using cosine similarity calculation method
Also known as cosine similarity, measures the similarity of two vectors by calculating the cosine value of angle, and cosine value value range is
[- 1,1], angle is smaller, and cosine value tends to consistent closer to the direction of 1, two vector, and similarity is also higher, calculates target quotient
Cosine similarity in product and N number of commodity between other commodity:
Wherein n indicates the vector dimension of the real vector of commodity,Indicate the real vector of end article,
The real vector for indicating other commodity in addition to end article in N number of commodity, and if only if predetermined threshold ε≤similarityxWhen,
RetainCorresponding product name, remaining removal.It learns by a large amount of experimental data as predetermined threshold ε=0.75,
The relevance of the commodity and end article that remain is stronger.
The Word2vec algorithm is the open a software tool for being used to train term vector of Google company, its basis
Each of sentence word is fast and effeciently mapped in k dimension space by given corpus by the training pattern after optimization
Have true value obtain vector, core architecture includes CBOW and Skip-gram.CBOW is to speculate target word from original statement
Word;And Skip-gram is exactly the opposite, is to deduce original statement from target words.CBOW is proper to small-sized data, and
Skip-gram is performed better than in large-scale corpus.The present invention uses Skip-gram model, this model is predicted with current word
S103 is obtained random walk sequence as corpus by the probability appeared below, the present invention.
S105, the commodity for needing to recommend are determined according to the degree of association between commodity.
Specifically, using the cosine similarity between other commodity in end article and N number of commodity, to determine target
When the degree of association between commodity and other commodity, the commodity for needing to recommend include:
It is greater than the commodity of predetermined threshold in N number of commodity with the cosine similarity of the end article.
Specifically, and if only if ε≤similarityx, that is, the cosine value between end article and other commodity that calculates
When more than or equal to 0.75, the commodity and end article correlation are strong, can carry out using the commodity as Recommendations to user
Recommend.
It, can also while will be multiple when it includes multiple for being greater than the commodity of predetermined threshold with the cosine similarity of end article
Commodity are recommended to user, can also be pushed away in multiple commodity according to pre-defined rule selection one or more of which commodity
It recommends.
The high commodity of similarity are subjected to bundle sale:Web superstore has been added to shopping cart according to user or has bought
Commodity, to the user recommend with the high commodity of the commodity similarity, and the network linking of the commodity is recommended user;Entity
Supermarket arranges the commodity on shelf according to each commodity similarity, and the high commodity of similarity are placed in adjacent shelf.So not
It only improves user and buys convenience degree, also promote the business revenue of supermarket.
Embodiment two:
The embodiment of the present application provides a kind of commercial product recommending equipment, the method for executing above-mentioned commercial product recommending.Fig. 6 is shown
A kind of possible modular structure schematic diagram of involved commercial product recommending equipment.Specifically commercial product recommending equipment 10 includes:It obtains
Module 101, building module 102, sequence generating module 103, determining module 104, recommending module 105.Wherein
Module 101 is obtained, for obtaining history shopping record.
Specifically can be by obtaining in preset time period, the mode of shopping record when different consumer purchases goods obtains
History shopping record.For example, extracting the shopping record of all customers in some supermarket's half a year.
Module 102 is constructed, for establishing commodity with purchase relationship after the acquisition module obtains the history shopping record
Network;Wherein, the commodity include N number of commodity, N >=2 in relational network with purchasing;It is done shopping and is recorded according to the history, however, it is determined that
Two kinds of commodity were bought simultaneously, then were attached described two commodity;It include end article in N number of commodity.
Sequence generating module 103 is used for after the building module constructs the commodity with purchase relational network, in the quotient
Product, using Random Walk Algorithm, respectively obtain the corresponding random walk sequence of N number of commodity in purchase relational network.
Determining module 104, for determining the target after the sequence generating module obtains the random walk sequence
The degree of association between commodity and other commodity.
As shown in fig. 7, computing module 104 can specifically include:Vector calculation unit 1041, similarity calculated
1042.Wherein:
Vector calculation unit 1041, for being calculated using Word2vec according to the corresponding random walk sequence of the N number of commodity
Method calculates separately the real vector of N number of commodity.
Similarity calculated 1042, for respectively obtained in the vector calculation unit real numbers of N number of commodity to
After amount, the cosine similarity in the end article and N number of commodity between other commodity is calculated.
The cosine similarity formula is:
Wherein n indicates the vector dimension of the real vector of commodity,Indicate the real vector of end article,
Indicate the real vector of other commodity in addition to end article in N number of commodity.
Recommending module 105, for determining other commodity in the end article and N number of commodity in the determining module
Between after the degree of association, determine the commodity for needing to recommend.
It is greater than the commodity of predetermined threshold in N number of commodity with the cosine similarity of the end article.
Specifically, and if only if ε≤similarityx, that is, the cosine value between end article and other commodity that calculates
When more than or equal to 0.75, the commodity and end article correlation are strong, can carry out using the commodity as Recommendations to user
Recommend.
It, can also while will be multiple when it includes multiple for being greater than the commodity of predetermined threshold with the cosine similarity of end article
Commodity are recommended to user, can also be pushed away in multiple commodity according to pre-defined rule selection one or more of which commodity
It recommends.
Above-mentioned integrated module both can take the form of hardware realization, can also be real using the form of software function module
It is existing.It should be noted that be schematical, only a kind of logical function partition to the division of module in the embodiment of the present application,
There may be another division manner in actual implementation.For example, each functional module of each function division can be corresponded to, it can also be by two
A or more than two functions are integrated in a processing module.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the program can be stored in a computer-readable storage medium
In, the program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein, the storage medium can be magnetic
Dish, CD, read-only memory (Read-Only Memory, ROM) or random access memory (Random Access
Memory, RAM) etc..
The foregoing is merely a specific embodiment of the invention, but scope of protection of the present invention is not limited thereto, any
In the technical scope disclosed by the present invention, any changes or substitutions that can be easily thought of by those familiar with the art, all answers
It is included within the scope of the present invention.Therefore, protection scope of the present invention should be with the scope of protection of the claims
It is quasi-.
Claims (10)
1. a kind of Method of Commodity Recommendation, which is characterized in that include the following steps:
Obtain history shopping record;
It is done shopping and is recorded according to the history, established commodity and purchase relational network together;Wherein, include in the same purchase relational network of the commodity
N number of commodity, N >=2;It is done shopping and is recorded according to the history, however, it is determined that two kinds of commodity were bought simultaneously, then by described two commodity
It is attached;It include end article in N number of commodity;
In the commodity in purchase relational network, using Random Walk Algorithm, the corresponding random trip of N number of commodity is calculated separately
Walk sequence;
According to the corresponding random walk sequence of the N number of commodity, the degree of association between the end article and other commodity is determined;
According to the degree of association between the end article and other commodity, the commodity for needing to recommend are determined.
2. Method of Commodity Recommendation according to claim 1, which is characterized in that
According to the corresponding random walk sequence of the N number of commodity, the degree of association between the end article and other commodity is determined,
Including:
N number of commodity are calculated separately using Word2vec algorithm according to the corresponding random walk sequence of the N number of commodity
Real vector;
According to the real vector of N number of commodity, calculate remaining between other commodity in the end article and N number of commodity
String similarity.
3. Method of Commodity Recommendation according to claim 2, which is characterized in that
It is described to need the commodity recommended to include:It is greater than predetermined threshold with the cosine similarity of the end article in N number of commodity
The commodity of value.
4. Method of Commodity Recommendation according to claim 2, which is characterized in that the cosine similarity calculation formula is:
Wherein n indicates the vector dimension of the real vector of commodity,Indicate the real vector of end article,It indicates
Another real vector of commodity.
5. Method of Commodity Recommendation according to claim 1-4, which is characterized in that the history shopping record, packet
It includes:It obtains in preset time period, shopping record when different consumer purchases goods.
6. a kind of device for recommending the commodity, which is characterized in that including:
Module is obtained, for obtaining history shopping record;
Building module, for establishing after the acquisition module obtains the history shopping record, commodity are same to purchase relational network;Its
In, the commodity include N number of commodity, N >=2 in relational network with purchasing;It is done shopping and is recorded according to the history, however, it is determined that two kinds of commodity
It was bought, was then attached described two commodity simultaneously;It include end article in N number of commodity;
Sequence generating module is used for after the building module constructs the commodity with purchase relational network, in the commodity with purchase
In relational network, using Random Walk Algorithm, the corresponding random walk sequence of N number of commodity is respectively obtained;
Determining module, for after the sequence generating module obtains the random walk sequence, determine the end article with
The degree of association between other commodity;
Recommending module, for after the degree of association that the determining module determines between the end article and other commodity, really
The commodity for needing to recommend calmly.
7. the device for recommending the commodity according to claim 6, which is characterized in that the determining module specifically includes:To meter
Calculate unit and similarity calculated;
The vector calculation unit, for according to the corresponding random walk sequence of the N number of commodity, using Word2vec algorithm,
Calculate separately the real vector of N number of commodity;
The similarity calculated, for after the real vector that the vector calculation unit respectively obtains N number of commodity,
Calculate the cosine similarity in the end article and N number of commodity between other commodity.
8. the device for recommending the commodity according to claim 7, which is characterized in that
It is described to need the commodity recommended to include:It is greater than predetermined threshold with the cosine similarity of the end article in N number of commodity
The commodity of value.
9. the device for recommending the commodity according to claim 7, which is characterized in that the cosine similarity meter of the similarity unit
Calculating formula is:
Wherein n indicates the vector dimension of the real vector of commodity,Indicate the real vector of end article,It indicates
Another real vector of commodity.
10. the device for recommending the commodity a method according to any one of claims 6-8, which is characterized in that the acquisition module
The history obtained, which is done shopping, to be recorded, including:It obtains in preset time period, shopping record when different consumer purchases goods.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810665009.2A CN108898459B (en) | 2018-06-25 | 2018-06-25 | Commodity recommendation method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810665009.2A CN108898459B (en) | 2018-06-25 | 2018-06-25 | Commodity recommendation method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108898459A true CN108898459A (en) | 2018-11-27 |
CN108898459B CN108898459B (en) | 2020-12-01 |
Family
ID=64346321
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810665009.2A Active CN108898459B (en) | 2018-06-25 | 2018-06-25 | Commodity recommendation method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108898459B (en) |
Cited By (25)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109885758A (en) * | 2019-01-16 | 2019-06-14 | 西北工业大学 | A kind of recommended method of the novel random walk based on bigraph (bipartite graph) |
CN110084670A (en) * | 2019-04-15 | 2019-08-02 | 东北大学 | A kind of commodity on shelf combined recommendation method based on LDA-MLP |
CN110210905A (en) * | 2019-05-31 | 2019-09-06 | 拉扎斯网络科技(上海)有限公司 | Characteristic similarity calculation method, device, electronic equipment and computer storage medium |
CN110210900A (en) * | 2019-05-23 | 2019-09-06 | 中国银行股份有限公司 | Method, apparatus and equipment are determined based on the reference product of transaction association |
CN110264277A (en) * | 2019-06-25 | 2019-09-20 | 杭州网易再顾科技有限公司 | Data processing method and device, medium and the calculating equipment executed by calculating equipment |
CN110458637A (en) * | 2019-06-19 | 2019-11-15 | 中国平安财产保险股份有限公司 | Product method for pushing and its relevant device neural network based |
CN110782287A (en) * | 2019-10-25 | 2020-02-11 | 北京沃东天骏信息技术有限公司 | Entity similarity calculation method and device, article recommendation system, medium and equipment |
CN111177543A (en) * | 2019-12-23 | 2020-05-19 | 杭州火小二科技有限公司 | Object recommendation method and device, storage medium and electronic device |
CN111178951A (en) * | 2019-12-20 | 2020-05-19 | 深圳大学 | Commodity recommendation method and device |
CN111292164A (en) * | 2020-01-21 | 2020-06-16 | 上海风秩科技有限公司 | Commodity recommendation method and device, electronic equipment and readable storage medium |
CN111340566A (en) * | 2020-03-23 | 2020-06-26 | 京东数字科技控股有限公司 | Commodity classification method and device, electronic equipment and storage medium |
CN111415216A (en) * | 2020-02-11 | 2020-07-14 | 广州探途网络技术有限公司 | Commodity recommendation method and device, server and storage medium |
CN111523918A (en) * | 2019-02-02 | 2020-08-11 | 北京极智嘉科技有限公司 | Commodity clustering method, commodity clustering device, commodity clustering equipment and storage medium |
CN111724221A (en) * | 2019-03-19 | 2020-09-29 | 北京京东尚科信息技术有限公司 | Method, system, electronic device and storage medium for determining commodity matching information |
CN111815403A (en) * | 2020-06-19 | 2020-10-23 | 北京石油化工学院 | Commodity recommendation method and device and terminal equipment |
CN111882404A (en) * | 2020-08-06 | 2020-11-03 | 上海风秩科技有限公司 | Association push processing method and device |
CN112100221A (en) * | 2019-06-17 | 2020-12-18 | 腾讯科技(北京)有限公司 | Information recommendation method and device, recommendation server and storage medium |
CN112669128A (en) * | 2021-01-28 | 2021-04-16 | 天元大数据信用管理有限公司 | Real-time clustering recommendation method, device and medium |
CN112925993A (en) * | 2021-04-08 | 2021-06-08 | 国网电子商务有限公司 | Collaborative filtering recommendation method and device, storage medium and equipment |
CN112989030A (en) * | 2021-03-15 | 2021-06-18 | 中国美术学院 | Visualization method and system for similar product relationship |
CN113763014A (en) * | 2021-01-05 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | Article co-occurrence relation determining method and device and judgment model obtaining method and device |
CN113781152A (en) * | 2021-02-08 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Object recommendation method and device |
CN114169975A (en) * | 2021-12-17 | 2022-03-11 | 福州大学 | Shopping network commodity recommendation method and system based on random walk heterogeneous attention |
CN115544242A (en) * | 2022-12-01 | 2022-12-30 | 深圳市智加云栖科技有限公司 | Big data based similar commodity model selection recommendation method |
CN116861323A (en) * | 2023-07-24 | 2023-10-10 | 深圳丰享信息技术有限公司 | Method and device for solving long tail effect in recommendation |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101206751A (en) * | 2007-12-25 | 2008-06-25 | 北京科文书业信息技术有限公司 | Customer recommendation system based on data digging and method thereof |
CN101251859A (en) * | 2008-03-17 | 2008-08-27 | 北京科文书业信息技术有限公司 | Commercial articles clustering system and method based on shopping behaviors |
CN102087730A (en) * | 2009-12-08 | 2011-06-08 | 深圳市腾讯计算机系统有限公司 | Method and device for constructing product user network |
CN103208073A (en) * | 2012-01-17 | 2013-07-17 | 阿里巴巴集团控股有限公司 | Method and device for obtaining recommend commodity information and providing commodity information |
CN103700005A (en) * | 2013-12-17 | 2014-04-02 | 南京信息工程大学 | Association-rule recommending method based on self-adaptive multiple minimum supports |
CN105427125A (en) * | 2015-10-29 | 2016-03-23 | 电子科技大学 | Goods clustering method based on goods network connection graph |
KR101612482B1 (en) * | 2015-04-06 | 2016-04-14 | 주식회사 포워드벤처스 | System and method for providing shopping service |
CN107248095A (en) * | 2017-04-14 | 2017-10-13 | 北京小度信息科技有限公司 | Recommend method and device |
CN107730336A (en) * | 2016-08-12 | 2018-02-23 | 苏宁云商集团股份有限公司 | Commodity method for pushing and device in a kind of online transaction |
-
2018
- 2018-06-25 CN CN201810665009.2A patent/CN108898459B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101206751A (en) * | 2007-12-25 | 2008-06-25 | 北京科文书业信息技术有限公司 | Customer recommendation system based on data digging and method thereof |
CN101251859A (en) * | 2008-03-17 | 2008-08-27 | 北京科文书业信息技术有限公司 | Commercial articles clustering system and method based on shopping behaviors |
CN102087730A (en) * | 2009-12-08 | 2011-06-08 | 深圳市腾讯计算机系统有限公司 | Method and device for constructing product user network |
CN103208073A (en) * | 2012-01-17 | 2013-07-17 | 阿里巴巴集团控股有限公司 | Method and device for obtaining recommend commodity information and providing commodity information |
CN103700005A (en) * | 2013-12-17 | 2014-04-02 | 南京信息工程大学 | Association-rule recommending method based on self-adaptive multiple minimum supports |
KR101612482B1 (en) * | 2015-04-06 | 2016-04-14 | 주식회사 포워드벤처스 | System and method for providing shopping service |
CN105427125A (en) * | 2015-10-29 | 2016-03-23 | 电子科技大学 | Goods clustering method based on goods network connection graph |
CN107730336A (en) * | 2016-08-12 | 2018-02-23 | 苏宁云商集团股份有限公司 | Commodity method for pushing and device in a kind of online transaction |
CN107248095A (en) * | 2017-04-14 | 2017-10-13 | 北京小度信息科技有限公司 | Recommend method and device |
Non-Patent Citations (1)
Title |
---|
何有世 等: ""基于用户购买记录的改进协同过滤推荐"", 《计算机工程与设计》 * |
Cited By (37)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109885758B (en) * | 2019-01-16 | 2022-07-26 | 西北工业大学 | Random walk recommendation method based on bipartite graph |
CN109885758A (en) * | 2019-01-16 | 2019-06-14 | 西北工业大学 | A kind of recommended method of the novel random walk based on bigraph (bipartite graph) |
CN111523918B (en) * | 2019-02-02 | 2023-09-19 | 北京极智嘉科技股份有限公司 | Commodity clustering method, device, equipment and storage medium |
CN111523918A (en) * | 2019-02-02 | 2020-08-11 | 北京极智嘉科技有限公司 | Commodity clustering method, commodity clustering device, commodity clustering equipment and storage medium |
CN111724221A (en) * | 2019-03-19 | 2020-09-29 | 北京京东尚科信息技术有限公司 | Method, system, electronic device and storage medium for determining commodity matching information |
CN110084670B (en) * | 2019-04-15 | 2022-03-25 | 东北大学 | Shelf commodity combination recommendation method based on LDA-MLP |
CN110084670A (en) * | 2019-04-15 | 2019-08-02 | 东北大学 | A kind of commodity on shelf combined recommendation method based on LDA-MLP |
CN110210900A (en) * | 2019-05-23 | 2019-09-06 | 中国银行股份有限公司 | Method, apparatus and equipment are determined based on the reference product of transaction association |
CN110210900B (en) * | 2019-05-23 | 2021-05-25 | 中国银行股份有限公司 | Reference product determination method, device and equipment based on transaction relevance |
CN110210905A (en) * | 2019-05-31 | 2019-09-06 | 拉扎斯网络科技(上海)有限公司 | Characteristic similarity calculation method, device, electronic equipment and computer storage medium |
CN112100221B (en) * | 2019-06-17 | 2024-02-13 | 深圳市雅阅科技有限公司 | Information recommendation method and device, recommendation server and storage medium |
CN112100221A (en) * | 2019-06-17 | 2020-12-18 | 腾讯科技(北京)有限公司 | Information recommendation method and device, recommendation server and storage medium |
CN110458637A (en) * | 2019-06-19 | 2019-11-15 | 中国平安财产保险股份有限公司 | Product method for pushing and its relevant device neural network based |
CN110264277B (en) * | 2019-06-25 | 2022-08-16 | 杭州网易再顾科技有限公司 | Data processing method and device executed by computing equipment, medium and computing equipment |
CN110264277A (en) * | 2019-06-25 | 2019-09-20 | 杭州网易再顾科技有限公司 | Data processing method and device, medium and the calculating equipment executed by calculating equipment |
CN110782287A (en) * | 2019-10-25 | 2020-02-11 | 北京沃东天骏信息技术有限公司 | Entity similarity calculation method and device, article recommendation system, medium and equipment |
CN111178951B (en) * | 2019-12-20 | 2023-05-16 | 深圳大学 | Commodity recommendation method and device |
CN111178951A (en) * | 2019-12-20 | 2020-05-19 | 深圳大学 | Commodity recommendation method and device |
CN111177543A (en) * | 2019-12-23 | 2020-05-19 | 杭州火小二科技有限公司 | Object recommendation method and device, storage medium and electronic device |
CN111292164A (en) * | 2020-01-21 | 2020-06-16 | 上海风秩科技有限公司 | Commodity recommendation method and device, electronic equipment and readable storage medium |
CN111415216A (en) * | 2020-02-11 | 2020-07-14 | 广州探途网络技术有限公司 | Commodity recommendation method and device, server and storage medium |
CN111415216B (en) * | 2020-02-11 | 2023-11-07 | 广州探途网络技术有限公司 | Commodity recommendation method, commodity recommendation device, server and storage medium |
CN111340566A (en) * | 2020-03-23 | 2020-06-26 | 京东数字科技控股有限公司 | Commodity classification method and device, electronic equipment and storage medium |
CN111340566B (en) * | 2020-03-23 | 2023-12-08 | 京东科技控股股份有限公司 | Commodity classification method and device, electronic equipment and storage medium |
CN111815403B (en) * | 2020-06-19 | 2024-05-10 | 北京石油化工学院 | Commodity recommendation method and device and terminal equipment |
CN111815403A (en) * | 2020-06-19 | 2020-10-23 | 北京石油化工学院 | Commodity recommendation method and device and terminal equipment |
CN111882404A (en) * | 2020-08-06 | 2020-11-03 | 上海风秩科技有限公司 | Association push processing method and device |
CN113763014A (en) * | 2021-01-05 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | Article co-occurrence relation determining method and device and judgment model obtaining method and device |
CN112669128A (en) * | 2021-01-28 | 2021-04-16 | 天元大数据信用管理有限公司 | Real-time clustering recommendation method, device and medium |
CN113781152A (en) * | 2021-02-08 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Object recommendation method and device |
CN112989030A (en) * | 2021-03-15 | 2021-06-18 | 中国美术学院 | Visualization method and system for similar product relationship |
CN112925993A (en) * | 2021-04-08 | 2021-06-08 | 国网电子商务有限公司 | Collaborative filtering recommendation method and device, storage medium and equipment |
CN112925993B (en) * | 2021-04-08 | 2022-02-08 | 国网电子商务有限公司 | Collaborative filtering recommendation method and device, storage medium and equipment |
CN114169975A (en) * | 2021-12-17 | 2022-03-11 | 福州大学 | Shopping network commodity recommendation method and system based on random walk heterogeneous attention |
CN115544242A (en) * | 2022-12-01 | 2022-12-30 | 深圳市智加云栖科技有限公司 | Big data based similar commodity model selection recommendation method |
CN116861323A (en) * | 2023-07-24 | 2023-10-10 | 深圳丰享信息技术有限公司 | Method and device for solving long tail effect in recommendation |
CN116861323B (en) * | 2023-07-24 | 2024-02-23 | 深圳丰享信息技术有限公司 | Method and device for solving long tail effect in recommendation |
Also Published As
Publication number | Publication date |
---|---|
CN108898459B (en) | 2020-12-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108898459A (en) | A kind of Method of Commodity Recommendation and device | |
CN111369318B (en) | Recommendation method and system based on commodity knowledge graph feature learning | |
CN102456203B (en) | Determine method and the relevant apparatus of candidate products chained list | |
Kim et al. | Recommender system based on click stream data using association rule mining | |
US8631044B2 (en) | Machine optimization devices, methods, and systems | |
CN101751448B (en) | Commendation method of personalized resource information based on scene information | |
Liao et al. | Mining customer knowledge to implement online shopping and home delivery for hypermarkets | |
CN109658206A (en) | Information recommendation method and device | |
CN107230098A (en) | Method and system is recommended in a kind of timesharing of business object | |
CN106447463A (en) | Commodity recommendation method based on Markov decision-making process model | |
Elalfi et al. | Extracting rules from trained neural network using GA for managing E-business | |
CN106485559A (en) | Cheating recognition methods and device for on-line shop | |
CN110807691B (en) | Cross-commodity-class commodity recommendation method and device | |
CN110309410A (en) | A kind of information recommended method, platform and computer readable storage medium | |
Jibril et al. | Association rule mining approach: evaluating pre-purchase risk intentions in the online second-hand goods market | |
CN109615431A (en) | The system and method for data assets perception and pricing function are realized under big data background | |
Ge | A user-based collaborative filtering recommendation algorithm based on folksonomy smoothing | |
CN110781399A (en) | Cross-platform information pushing method and device | |
CN110738521B (en) | Client selling method and device for multi-merchant brand | |
Jain et al. | E-commerce product recommendation based on product specification and similarity | |
Meel et al. | Inverse document frequency-weighted Word2Vec model to recommend apparels | |
Fitrianah et al. | Analysis of Consumer Purchase Patterns on Handphone Accessories Sales Using FP-Growth Algorithm | |
Fu et al. | Percolation on shopping and cashback electronic commerce networks | |
CN110019154A (en) | Data processing, data quality accessment, recommended products determine method and relevant device | |
Hu et al. | Utilizing users' tipping points in E-commerce Recommender systems |
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 |