CN105025091A - Shop recommendation method based on position of mobile user - Google Patents

Shop recommendation method based on position of mobile user Download PDF

Info

Publication number
CN105025091A
CN105025091A CN201510363643.7A CN201510363643A CN105025091A CN 105025091 A CN105025091 A CN 105025091A CN 201510363643 A CN201510363643 A CN 201510363643A CN 105025091 A CN105025091 A CN 105025091A
Authority
CN
China
Prior art keywords
user
matrix
retail shop
shop
targeted customer
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201510363643.7A
Other languages
Chinese (zh)
Inventor
叶宁
卢华超
黄海平
沙超
王汝传
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing Post and Telecommunication University
Nanjing University of Posts and Telecommunications
Original Assignee
Nanjing Post and Telecommunication University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Nanjing Post and Telecommunication University filed Critical Nanjing Post and Telecommunication University
Priority to CN201510363643.7A priority Critical patent/CN105025091A/en
Publication of CN105025091A publication Critical patent/CN105025091A/en
Pending legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/50—Network services
    • H04L67/52—Network services specially adapted for the location of the user terminal
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information
    • H04W4/021—Services related to particular areas, e.g. point of interest [POI] services, venue services or geofences

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a shop recommendation method based on a position of a mobile user. The method comprises sending position information of the user to a server by a mobile phone used by the user, finding all shops related to the position by the server according to the position information sent by the user, finding all users related to the shops, and finally generating a user-shop scoring matrix; and the method further comprises initially filling the obtained user-shop scoring matrix, performing similarity calculation on a target user vector and other user vectors in the obtained matrix, generating neighbor users, performing scoring prediction on an unscored shop of the target use according to the neighbor users, and performing recommendation for the user according to a prediction result. After prediction scores of the unscored shops of the target user are obtained, the recommendation is performed for the user; and all the prediction scores are ranked, and then the shops, which have the high prediction scores and are the former N items, are recommended to the user.

Description

A kind of retail shop's recommend method based on location of mobile users
Technical field
The present invention relates to a kind of retail shop's recommend method based on location of mobile users, belong to communication technical field.
Background technology
Along with the continuous rapid growth of mobile phone quantity and the speed of mobile network and the raising of stability, the mobile phone application in conjunction with mobile Internet feature must become the very important demand of user.Position-based service (that is: LBS) is exactly not retrievable service of mobile Internet, and it can according to the position of user for user provides extremely personalized service.Under such major premise, emerge the application that a large amount of position-baseds is served, more then occur the service for life application of a collection of position-based service.Current commending system is own through penetrating into every field, as ecommerce, film and video, individualized music radio station, social networks and personalized advertisement etc., the application of position-based service presents the growth of explosion type, domestic group of e-commerce platform U.S., popular comment, Taobao etc., the micro-letter of social application, rice are chatted, footpath between fields, footpath between fields, kk look for friend etc., all have the characteristic of LBS.Obviously, the application of position-based service has become an important component part of people mobile network life.
Although the user that develops into of mobile radio communication provides a more colourful mobile network service platform, achieve user and network information resource is obtained anywhere or anytime and pushes, make as user provides ubiquitous mobile network service to become possibility.Especially move the rise of social network, for user in the network information service, share, provide great help in comment etc.But making rapid progress of COS and service content, limited mobile network resource and hardware resource, for mobile subscriber brings serious mobile message overload problem.From the mobile network environment of vastness, how to find the real interested information resources of user, enrich and meet the individual demand of mobile subscriber to information, becoming the technical barrier that in mobile communications network, personalized service field is urgently to be resolved hurrily gradually.
Commending system, as one of the solution of individual info service, all causes in industrial quarters and academia and pays close attention to widely.Wherein collaborative filtering recommending technology is one of technology be most widely used, although there is the problem such as " cold start-up ", " Deta sparseness ".But compared with traditional search engine, it not only focuses on relation between Search Results and sequence, but also emphasis considers that the personalization preferences model of user is on the impact of Search Results.In addition, the successful introducing of general fit calculation theory, conventional recommendation systems is made no longer only to pay close attention to " user-project " binary crelation, but the context information (as: time, position, surrounding people, temperament and interest, active state, network condition etc.) residing for user is together taken into account, make system automatically can find and utilize various contextual information, promote the recommendation quality of commending system, meet the individual information needs that user changes with contextual information change.And the present invention can solve problem above well.
Summary of the invention
The object of the invention there are provided a kind of retail shop's recommend method based on location of mobile users, this method solves the deficiency of above-mentioned existing issue, when Sparse, effectively can improve the predictablity rate of commending system.The method is based on mobile subscriber's real time position and associate(d) matrix decomposes the recommend method with collaborative filtering, realizes excavating the different characteristics of user, and realizes the personalized recommendation to user according to the interest of user.
The present invention solves the technical scheme that its technical problem takes: the present invention is a kind of mixing recommend method, the method adopts and dynamically generates rating matrix according to customer position information, singular value decomposition is carried out to after the preliminary completion of rating matrix, again through the dimensionality reduction of matrix, reduction obtains final filled matrix, the basis of final filled matrix adopts traditional Collaborative Filtering Recommendation Algorithm find out the retail shop recommended user, and recommendation results is fed back to user.
Method flow:
Step 1: user sends the positional information of user to server with mobile phone, server finds out all shops relevant to position according to the positional information that user sends, and finds out all users relevant to shop, finally generates the rating matrix in user-shop.
Step 11: mobile subscriber obtains real-time cellphone GPS locating information, and utilize data in mobile phone service to be uploaded onto the server by customer location.Be described as follows:
If the positional information that each user obtains is designated as location (x, y), wherein x is the longitude of user position, and y is the latitude of user position.
Step 12: server receives the positional information location of user (x, y), with this position for the center of circle, be the retail shop S in half path search peripheral extent with 2km location(s 1, s 2, s 3... .s n), wherein n is the quantity of the retail shop searched.
Step 13: find out all evaluation user U according to the history evaluation of this n retail shop location(u 1, u 2, u 3... u m), wherein m is the quantity of user.
Step 14: generate user-retail shop rating matrix R (m × n), 0 is filled to the scoring of the retail shop that user marks.
Wherein, the quantity of n representative of consumer, m represents the quantity of retail shop.Element r in matrix R i, jrepresent that user i is to the scoring of service item j, all users are to the integer representation of the scoring of retail shop between 0 ~ 5, and 0 represents that user does not mark, the height of the fancy grade of 1 ~ 5 expression user.
Step 2: user-shop rating matrix that step 1 obtains tentatively is filled up, namely the element in rating matrix being 0 is replaced, singular value decomposition is carried out to the matrix after filling up, utilize matrix after singular value decomposition, dimensionality reduction, reduction to carry out filling up again to the value in former rating matrix being 0, flow process as shown in Figure 4.
Step 21: in compute matrix R, each user is to the average score value of the retail shop of having marked, and this mean value is filled in the missing values of each user and obtains new matrix R '.
Step 22: utilize svd algorithm, is decomposed into U, S, V by the matrix R ' obtained tthree matrixes.
R′＝USV T
Step 23: diagonal matrix S is carried out abbreviation, is all reset to 0 by the singular value being less than 1 in all diagonal matrix, because diagonal matrix singular value is descending arrangement, is therefore that the row and column of 0 is deleted by numerical value all in s-matrix, generates new S kdiagonal matrix (be equivalent to retain front K capable and front K row);
Step 24: according to the S simplified k, corresponding is U by U, V matrix reduction k, matrix, U kfor the two-dimensional matrix of m × k, for the two-dimensional matrix of k × n;
Step 25: by three matrix U after simplification k, S k, and be multiplied, and be filled into the item of not marking of original matrix, obtain last filled matrix R ".
Step 3: to obtaining matrix in step 2, get targeted customer wherein and other user carries out Similarity Measure, produce neighbour user, then carry out score in predicting according to the retail shop of not marking of neighbour user to targeted customer, to recommend user according to predicting the outcome, flow process as shown in Figure 5.
Step 31: calculate the similitude between targeted customer and other users respectively, this method adopts Pearson relevance formula to carry out Similarity measures, and formula is as follows:
s i m ( u , v ) = Σ i ∈ S ( r u , i - r u ‾ ) ( r v , i - r v ‾ ) Σ i ∈ S ( r u , i - r u ‾ ) 2 Σ i ∈ S ( r v , i - r v ‾ ) 2
Wherein, sim (u, v) represents the similarity between user u and user v, r u, irepresent that user u is to the scoring of retail shop i, with represent the set of the average that user u and user v marks to all items, all retail shops corresponding in S representing matrix.
Step 32: the similarity between other users in targeted customer and matrix is done descending, adopt K arest neighbors (kNN, k-NearestNeighbor) algorithm searches user's arest neighbors, find K the user the highest with targeted customer's similarity as its arest neighbors, obtain neighbour user's collection of targeted customer.
Step 33: after the arest neighbors user collection obtaining targeted customer, adopt weighted average method to mark to the retail shop of not marking of targeted customer in original matrix, formula is as follows:
P u , i = r u ‾ + Σ x ∈ N B ( r x , i - r x ‾ ) · s i m ( u , x ) Σ x ∈ N B | s i m ( u , x ) |
Wherein, P u, ibe that user u marks to the prediction of retail shop i, u is targeted customer, and NB is neighbour user's collection of targeted customer u.
Step 4: obtain targeted customer do not mark retail shop prediction scoring after just can recommend user, this method TOP-N method, exactly by all predictions scoring sort, then by prediction mark the highest before N item recommend user.
The retail shop that the present invention is applied to based on location of mobile users is recommended.
Beneficial effect:
1, the inventive method proposes and utilizes the real-time position information of mobile subscriber to generate user-retail shop's rating matrix, effectively reduces the dimension of rating matrix, for follow-up calculating reduces time and space complexity.
2, the present invention adopts singular value decomposition algorithm to combine with collaborative filtering recommending, efficiently solves the Sparse Problem of rating matrix, and improves the predictablity rate of commending system.
Accompanying drawing explanation
Fig. 1 is information communication mechanism schematic diagram of the present invention.
Fig. 2 is the flow chart of method of the present invention.
Fig. 3 is the schematic flow sheet of step 1 of the present invention.
Fig. 4 is the schematic flow sheet of step 2 of the present invention.
Fig. 5 is the schematic flow sheet of step 3 of the present invention.
Embodiment
Below in conjunction with Figure of description, the invention is described in further detail.
Embodiment 1
As shown in Figure 2-5, the present invention suppose there is following application example: target mobile user u 1position gps coordinate is: (116.41757,39.917305), has 6 retail shops around this position, to these 6 retail shops carried out evaluate have 4 users.
(1) mobile subscriber's upload location information, system obtains customer position information and generates user-retail shop's rating matrix, and detailed process is as follows:
Step 1: mobile subscriber utilizes mobile phone to obtain real time GPS locating information location (116.41757, 39.917305), and utilize data in mobile phone service to be uploaded onto the server by customer location.
Step 2: server receives the positional information of user, with this position for the center of circle, the retail shop around search within the scope of 2km, now n=6, S location(s 1, s 2, s 3, s 4, s 5, s 6).
Step 3: find out all evaluation users according to the history evaluation of this 6 family retail shop, and find out the user of all scorings, now m=4, U location(u 1, u 2, u 3, u 4).
Step 4: generate user-retail shop rating matrix R (m × n), now R is 4 × 6 dimension matrixes.0 is filled to the scoring of the retail shop that user marks.
R = 0 1 3 0 5 2 2 0 2 1 5 3 0 0 3 2 0 1 4 5 0 2 0 0
(2) user-shop rating matrix that step 1 obtains tentatively is filled up, namely the element in rating matrix being 0 is replaced, singular value decomposition is carried out to the matrix after filling up, utilizes the matrix after singular value decomposition, dimensionality reduction, reduction to carry out filling up again to the value in former rating matrix being 0.
Step 21: in compute matrix R, each user is to the average score value of the retail shop of having marked, and is filled into by this score value in the missing values of each user.Now
R ′ = 3 1 3 3 5 2 2 3 2 1 5 3 2 2 3 2 2 1 4 5 4 2 4 4
Step 22: utilize svd algorithm, is decomposed into U, S, V by the matrix R ' obtained tthree matrixes.
S = 14.56 0 0 0 0 0 0 3.08 0 0 0 0 0 0 2.45 0 0 0 0 0 0 0.71 0 0
U = - 0.49 - 0.8 0.08 - 0.35 - 0.47 0.03 - 0.76 0.44 - 0.33 - 0.04 0.62 0.71 - 0.65 0.6 0.18 - 0.43
V T = - 0.39 - 0.4 - 0.41 - 0.27 - 0.56 - 0.37 - 0.01 0.72 - 0.02 - 0.4 - 0.49 0.28 0.28 - 0.02 0.53 0.44 - 0.59 - 0.31 - 0.62 0.36 0.37 - 0.05 0.27 - 0.52 - 0.17 - 0.56 - 0.56 0.72 - 0.05 - 0.03 - 0.59 0.3 0.3 0.23 - 0.18 0.65
Step 23: diagonal matrix S is carried out abbreviation, is all reset to 0 by the singular value being less than 1 in all diagonal matrix, because diagonal matrix singular value is descending arrangement, is therefore that the row and column of 0 is deleted by numerical value all in s-matrix, generates new S 3diagonal matrix (that is: being equivalent to retain front 3 row and front 3 row);
Step 24: according to the S simplified 3, corresponding to U, V tmatrix reduction is U 3, matrix, U 3be the two-dimensional matrix of 4 × 3, it is the two-dimensional matrix of 3 × 6;
Step 25: by three matrix U after simplification 3, S 3, and be multiplied, and be filled into the item of not marking of original matrix, obtain last filled matrix
R ′ ′ = 4 1 3 4 5 2 2 1 2 1 5 3 1 1 3 2 1 1 4 5 2 2 2 5
(3) to obtaining matrix in step 2, the targeted customer's vector got wherein carries out Similarity Measure with other user vector, producing neighbour user, then carry out score in predicting according to the retail shop of not marking of neighbour user to targeted customer, according to predicting the outcome user being recommended.
Step 31: calculate the similitude between targeted customer and other users respectively, this method adopts Pearson relevance formula to carry out Similarity measures.As calculated:
sim(u 1，u 2)≈0.511 sim(u 1，u 3)≈0.081 sim(u 1，u 4)≈-0.752
Step 32: the similarity between other users in targeted customer and matrix is done descending, adopt K arest neighbors (kNN, k-NearestNeighbor) algorithm searches user's arest neighbors, find 2 users the highest with targeted customer's similarity as its arest neighbors, obtain neighbour user's collection (u of targeted customer 2, u 3).
Step 33: after the arest neighbors user collection obtaining targeted customer, adopts weighted average method to mark to the retail shop of not marking of targeted customer in original matrix, as calculated P 1,1≈ 3, P isosorbide-5-Nitrae≈ 2.
Step 4: just can recommend user after the prediction scoring of retail shop of not marked, this method TOP-N method, exactly by all predictions scoring sort, then by prediction scoring the highest before N item recommend user, if now N=1, then now system to user u 1recommend S 1retail shop.
Embodiment 2
As shown in Figure 1, communication information Mechanism Model comprises: mobile subscriber terminal, communication link, server.
Based on retail shop's recommend method of location of mobile users flow process as shown in Figure 2, comprising:
Step 1: user sends the positional information of user to server with mobile phone, server finds out all shops relevant to position according to the positional information that user sends, and find out all users relevant to shop, finally generate the rating matrix in user-shop, flow process is as shown in Figure 3.
Step 11: mobile subscriber obtains real-time cellphone GPS locating information, and utilize data in mobile phone service to be uploaded onto the server by customer location.Be described as follows:
If the positional information that each user obtains is designated as location (x, y), wherein x is the longitude of user position, and y is the latitude of user position.
Step 12: server receives the positional information location of user (x, y), with this position for the center of circle, be the retail shop S in half path search peripheral extent with 2km location(s 1, s 2, s 3... .s n), wherein n is the quantity of the retail shop searched.
Step 13: find out all evaluation user U according to the history evaluation of this n retail shop location(u 1, u 2, u 3... u m), wherein m is the quantity of user.
Step 14: generate user-retail shop rating matrix R (m × n), 0 is filled to the scoring of the retail shop that user marks.
Wherein, the quantity of n representative of consumer, m represents the quantity of retail shop.Element r in matrix R i, jrepresent that user i is to the scoring of service item j, all users are to the integer representation of the scoring of retail shop between 0 ~ 5, and 0 represents that user does not mark, the height of the fancy grade of 1 ~ 5 expression user.
Step 2: user-shop rating matrix that step 1 obtains tentatively is filled up, namely the element in rating matrix being 0 is replaced, singular value decomposition is carried out to the matrix after filling up, utilize matrix after singular value decomposition, dimensionality reduction, reduction to carry out filling up again to the value in former rating matrix being 0, flow process as shown in Figure 4.
Step 21: in compute matrix R, each user is to the average score value of the retail shop of having marked, and this mean value is filled in the missing values of each user and obtains new matrix R '.
Step 22: utilize svd algorithm, is decomposed into U, S, V by the matrix R ' obtained tthree matrixes.
R′＝USV T
Step 23: diagonal matrix S is carried out abbreviation, is all reset to 0 by the singular value being less than 1 in all diagonal matrix, because diagonal matrix singular value is descending arrangement, is therefore that the row and column of 0 is deleted by numerical value all in s-matrix, generates new S kdiagonal matrix (be equivalent to retain front K capable and front K row);
Step 24: according to the S simplified k, corresponding is U by U, V matrix reduction k, matrix, U kfor the two-dimensional matrix of m × k, for the two-dimensional matrix of k × n;
Step 25: by three matrix U after simplification k, S k, and be multiplied, and be filled into the item of not marking of original matrix, obtain last filled matrix R ".
Step 3: to obtaining matrix in step 2, get targeted customer wherein and other user carries out Similarity Measure, produce neighbour user, then carry out score in predicting according to the retail shop of not marking of neighbour user to targeted customer, to recommend user according to predicting the outcome, flow process as shown in Figure 5.
Step 31: calculate the similitude between targeted customer and other users respectively, this method adopts Pearson relevance formula to carry out Similarity measures, and formula is as follows:
s i m ( u , v ) = Σ i ∈ S ( r u , i - r u ‾ ) ( r v , i - r v ‾ ) Σ i ∈ S ( r u , i - r u ‾ ) 2 Σ i ∈ S ( r v , i - r v ‾ ) 2
Wherein, sim (u, v) represents the similarity between user u and user v, r u, irepresent that user u is to the scoring of retail shop i, with represent the set of the average that user u and user v marks to all items, all retail shops corresponding in S representing matrix.
Step 32: the similarity between other users in targeted customer and matrix is done descending, adopt K arest neighbors (kNN, k-NearestNeighbor) algorithm searches user's arest neighbors, find K the user the highest with targeted customer's similarity as its arest neighbors, obtain neighbour user's collection of targeted customer.
Step 33: after the arest neighbors user collection obtaining targeted customer, adopt weighted average method to mark to the retail shop of not marking of targeted customer in original matrix, formula is as follows:
P u , i = r u ‾ + Σ x ∈ N B ( r x , i - r x ‾ ) · s i m ( u , x ) Σ x ∈ N B | s i m ( u , x ) |
Wherein, P u, ibe that user u marks to the prediction of retail shop i, u is targeted customer, and NB is neighbour user's collection of targeted customer u.
Step 4: obtain targeted customer do not mark retail shop prediction scoring after just can recommend user, this method TOP-N method, exactly by all predictions scoring sort, then by prediction mark the highest before N item recommend user.

Claims (6)

1. based on retail shop's recommend method of location of mobile users, it is characterized in that, described method comprises the steps:
Step 1: user sends the positional information of user to server with mobile phone, server finds out all shops relevant to position according to the positional information that user sends, and finds out all users relevant to shop, finally generates the rating matrix in user-shop;
Step 2: user-shop rating matrix that above-mentioned steps 1 obtains tentatively is filled up, that is: the element in rating matrix being 0 is replaced, singular value decomposition is carried out to the matrix after filling up, utilizes the matrix after singular value decomposition, dimensionality reduction, reduction to carry out filling up again to the value in former rating matrix being 0;
Step 3: to obtaining matrix in above-mentioned steps 2, the targeted customer's vector got wherein carries out Similarity Measure with other user vector, producing neighbour user, then carry out score in predicting according to the retail shop of not marking of neighbour user to targeted customer, according to predicting the outcome user being recommended;
Step 4: obtain targeted customer do not mark retail shop prediction scoring after, user is recommended, by all predictions scoring sort, then by prediction scoring the highest before N item recommend user.
2. a kind of retail shop's recommend method based on location of mobile users according to claim 1, it is characterized in that, the step 1 of described method comprises the steps:
Step 11: mobile subscriber obtains real-time cellphone GPS locating information, and utilize data in mobile phone service to be uploaded onto the server by customer location;
Step 12: server receives the positional information location of user (x, y), with this position for the center of circle, be the retail shop S in half path search peripheral extent with 2km location(s 1, s 2, s 3... .s n), wherein n is the quantity of the retail shop searched;
Step 13: find out all evaluation user U according to the history evaluation of this n retail shop location(u 1, u 2, u 3... .u m), wherein m is the quantity of user;
Step 14: generate user-retail shop rating matrix R (m × n), 0 is filled to the scoring of the retail shop that user marks.
3. a kind of retail shop's recommend method based on location of mobile users according to claim 1, it is characterized in that, the step 2 of described method comprises the steps:
Step 21: in compute matrix R, each user is to the average score value of the retail shop of having marked, and this score value is filled in the missing values of each user and obtains new matrix R ';
Step 22: utilize svd algorithm, is decomposed into U, S, V by the matrix R ' obtained tthree matrixes;
R′＝USV T
Step 23: diagonal matrix S is carried out abbreviation, is all reset to 0 by the singular value being less than 1 in all diagonal matrix, because diagonal matrix singular value is descending arrangement, is therefore that the row and column of 0 is deleted by numerical value all in s-matrix, generates new S kdiagonal matrix, that is: the capable and front K row of K before retaining;
Step 24: according to the S simplified k, corresponding is U by U, V matrix reduction k, matrix, U kfor the two-dimensional matrix of m × k, for the two-dimensional matrix of k × n;
Step 25: by three matrix U after simplification k, S k, and be multiplied, and be filled into the item of not marking of original matrix, obtain last filled matrix R ".
4. a kind of retail shop's recommend method based on location of mobile users according to claim 1, it is characterized in that, the step 3 of described method comprises the steps:
Step 31: calculate the similitude between targeted customer and other users respectively, this method adopts Pearson relevance formula to carry out Similarity measures, and formula is as follows:
s i m ( u , v ) = Σ i ∈ S ( r u , i - r u ‾ ) ( r v , i - r v ‾ ) Σ i ∈ S ( r u , i - r u ‾ ) 2 Σ i ∈ S ( r v , i - r v ‾ ) 2
Wherein, sim (u, v) represents the similarity between user u and user v, r u, irepresent that user u is to the scoring of retail shop i, with represent the set of the average that user u and user v marks to all items, all retail shops corresponding in S representing matrix;
Step 32: the similarity between other users in targeted customer and matrix is done descending, adopt K arest neighbors (kNN, k-NearestNeighbor) algorithm searches user's arest neighbors, find K the user the highest with targeted customer's similarity as its arest neighbors, obtain neighbour user's collection of targeted customer;
Step 33: after the arest neighbors user collection obtaining targeted customer, adopt weighted average method to mark to the retail shop of not marking of targeted customer in original matrix, formula is as follows:
P u , i = r u ‾ + Σ x ∈ N B ( r x , i - r x ‾ ) · s i m ( u , x ) Σ x ∈ N B | s i m ( u , x ) |
Wherein, P u, ibe that user u marks to the prediction of retail shop i, u is targeted customer, and NB is neighbour user's collection of targeted customer u.
5. a kind of retail shop's recommend method based on location of mobile users according to claim 1, it is characterized in that, described method adopts and dynamically generates rating matrix according to customer position information, singular value decomposition is carried out to after the preliminary completion of rating matrix, again through the dimensionality reduction of matrix, reduction obtains final filled matrix, the basis of final filled matrix adopts traditional Collaborative Filtering Recommendation Algorithm find out the retail shop recommended user, and recommendation results is fed back to user.
6. a kind of retail shop's recommend method based on location of mobile users according to claim 1, it is characterized in that, the retail shop that described method is applied to based on location of mobile users is recommended.
CN201510363643.7A 2015-06-26 2015-06-26 Shop recommendation method based on position of mobile user Pending CN105025091A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510363643.7A CN105025091A (en) 2015-06-26 2015-06-26 Shop recommendation method based on position of mobile user

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510363643.7A CN105025091A (en) 2015-06-26 2015-06-26 Shop recommendation method based on position of mobile user

Publications (1)

Publication Number Publication Date
CN105025091A true CN105025091A (en) 2015-11-04

Family

ID=54414785

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510363643.7A Pending CN105025091A (en) 2015-06-26 2015-06-26 Shop recommendation method based on position of mobile user

Country Status (1)

Country Link
CN (1) CN105025091A (en)

Cited By (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105868334A (en) * 2016-03-28 2016-08-17 云南财经大学 Personalized film recommendation method and system based on feature augmentation
CN106971053A (en) * 2016-01-08 2017-07-21 车海莺 A kind of recommendation method based on mixing collaborative filtering
CN107016118A (en) * 2017-04-20 2017-08-04 上海喆之信息科技有限公司 A kind of safe search system for website
CN107093095A (en) * 2016-12-02 2017-08-25 口碑控股有限公司 Method and device is recommended in associated services processing method and processing device, shop
CN107194774A (en) * 2017-05-22 2017-09-22 携程旅游网络技术（上海）有限公司 Personalized hotel's commending system and method in OTA websites
CN107220303A (en) * 2017-05-10 2017-09-29 努比亚技术有限公司 Recommendation method, device and the computer-readable medium of a kind of application
CN107590261A (en) * 2017-09-21 2018-01-16 深圳市晟达机械设计有限公司 A kind of network service commending system based on customer location
CN107784095A (en) * 2017-10-18 2018-03-09 国网内蒙古东部电力有限公司 A kind of education resource auto recommending method based on mobile learning
CN108154425A (en) * 2018-01-19 2018-06-12 广州天源信息科技股份有限公司 Method is recommended by the Xian Xia trade companies of a kind of combination community network and position
CN108701313A (en) * 2015-12-11 2018-10-23 万事达卡国际股份有限公司 The system and method recommended are generated using data corpus
CN108804683A (en) * 2018-06-13 2018-11-13 重庆理工大学 Associate(d) matrix decomposes and the film of collaborative filtering recommends method
CN108924754A (en) * 2018-06-22 2018-11-30 张小勇 Kindergarten's screening technique and system
CN109388756A (en) * 2018-09-10 2019-02-26 浙江口碑网络技术有限公司 Information recommendation method and device
CN109522475A (en) * 2018-10-26 2019-03-26 浙江工业大学之江学院 A kind of merchant recommendation method based on user's history consumption data
CN109727056A (en) * 2018-07-06 2019-05-07 平安科技（深圳）有限公司 Financial institution's recommended method, equipment, storage medium and device
CN110321490A (en) * 2019-07-12 2019-10-11 科大讯飞(苏州)科技有限公司 Recommended method, device, equipment and computer readable storage medium
CN111402003A (en) * 2020-03-13 2020-07-10 第四范式（北京）技术有限公司 System and method for realizing user-related recommendation
CN111782941A (en) * 2016-05-11 2020-10-16 阿里巴巴集团控股有限公司 Information recommendation method and device and server
CN112071401A (en) * 2020-09-05 2020-12-11 苏州贝基电子科技有限公司 Healthy diet management system based on big data
CN113297496A (en) * 2021-06-18 2021-08-24 中山市力泰电子工业有限公司 Collaborative filtering recommendation algorithm based on improved user similarity
CN113449210A (en) * 2021-07-01 2021-09-28 深圳市数字尾巴科技有限公司 Personalized recommendation method and device based on space-time characteristics, electronic equipment and storage medium
US11869015B1 (en) 2022-12-09 2024-01-09 Northern Trust Corporation Computing technologies for benchmarking

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102594905A (en) * 2012-03-07 2012-07-18 南京邮电大学 Method for recommending social network position interest points based on scene
CN103023977A (en) * 2012-11-19 2013-04-03 华南理工大学 Recommendation system and method based on positional information
CN103093376A (en) * 2013-01-16 2013-05-08 北京邮电大学 Clustering collaborative filtering recommendation system based on singular value decomposition algorithm
CN103412865A (en) * 2013-06-07 2013-11-27 广州唯品会信息科技有限公司 Method and system for carrying out notification on website projects
CN103995823A (en) * 2014-03-25 2014-08-20 南京邮电大学 Information recommending method based on social network
CN104050258A (en) * 2014-06-15 2014-09-17 中国传媒大学 Group recommendation method based on interest groups
US20140370844A1 (en) * 2012-01-20 2014-12-18 Telefonica, S.A. Method for the automatic detection and labelling of user point of interest
CN104731866A (en) * 2015-02-27 2015-06-24 湖南大学 Individual gourmet recommending method based on position

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140370844A1 (en) * 2012-01-20 2014-12-18 Telefonica, S.A. Method for the automatic detection and labelling of user point of interest
CN102594905A (en) * 2012-03-07 2012-07-18 南京邮电大学 Method for recommending social network position interest points based on scene
CN103023977A (en) * 2012-11-19 2013-04-03 华南理工大学 Recommendation system and method based on positional information
CN103093376A (en) * 2013-01-16 2013-05-08 北京邮电大学 Clustering collaborative filtering recommendation system based on singular value decomposition algorithm
CN103412865A (en) * 2013-06-07 2013-11-27 广州唯品会信息科技有限公司 Method and system for carrying out notification on website projects
CN103995823A (en) * 2014-03-25 2014-08-20 南京邮电大学 Information recommending method based on social network
CN104050258A (en) * 2014-06-15 2014-09-17 中国传媒大学 Group recommendation method based on interest groups
CN104731866A (en) * 2015-02-27 2015-06-24 湖南大学 Individual gourmet recommending method based on position

Cited By (31)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108701313A (en) * 2015-12-11 2018-10-23 万事达卡国际股份有限公司 The system and method recommended are generated using data corpus
US11790431B2 (en) 2015-12-11 2023-10-17 Mastercard International Incorporated Systems and methods for generating recommendations using a corpus of data
CN106971053A (en) * 2016-01-08 2017-07-21 车海莺 A kind of recommendation method based on mixing collaborative filtering
CN105868334A (en) * 2016-03-28 2016-08-17 云南财经大学 Personalized film recommendation method and system based on feature augmentation
CN111782941A (en) * 2016-05-11 2020-10-16 阿里巴巴集团控股有限公司 Information recommendation method and device and server
CN111782941B (en) * 2016-05-11 2023-12-12 创新先进技术有限公司 Information recommendation method, device and server
CN107093095A (en) * 2016-12-02 2017-08-25 口碑控股有限公司 Method and device is recommended in associated services processing method and processing device, shop
CN107093095B (en) * 2016-12-02 2018-05-25 口碑(上海)信息技术有限公司 Method and device is recommended in associated services processing method and processing device, shop
CN107016118A (en) * 2017-04-20 2017-08-04 上海喆之信息科技有限公司 A kind of safe search system for website
CN107220303A (en) * 2017-05-10 2017-09-29 努比亚技术有限公司 Recommendation method, device and the computer-readable medium of a kind of application
CN107194774A (en) * 2017-05-22 2017-09-22 携程旅游网络技术（上海）有限公司 Personalized hotel's commending system and method in OTA websites
CN107590261A (en) * 2017-09-21 2018-01-16 深圳市晟达机械设计有限公司 A kind of network service commending system based on customer location
CN107784095A (en) * 2017-10-18 2018-03-09 国网内蒙古东部电力有限公司 A kind of education resource auto recommending method based on mobile learning
CN107784095B (en) * 2017-10-18 2022-04-01 国网内蒙古东部电力有限公司 Learning resource automatic recommendation method based on mobile learning
CN108154425A (en) * 2018-01-19 2018-06-12 广州天源信息科技股份有限公司 Method is recommended by the Xian Xia trade companies of a kind of combination community network and position
CN108804683A (en) * 2018-06-13 2018-11-13 重庆理工大学 Associate(d) matrix decomposes and the film of collaborative filtering recommends method
CN108804683B (en) * 2018-06-13 2021-11-23 重庆理工大学 Movie recommendation method combining matrix decomposition and collaborative filtering algorithm
CN108924754A (en) * 2018-06-22 2018-11-30 张小勇 Kindergarten's screening technique and system
CN109727056A (en) * 2018-07-06 2019-05-07 平安科技（深圳）有限公司 Financial institution's recommended method, equipment, storage medium and device
WO2020006834A1 (en) * 2018-07-06 2020-01-09 平安科技（深圳）有限公司 Financial institution recommending method, equipment, storage medium, and device
CN109727056B (en) * 2018-07-06 2023-04-18 平安科技（深圳）有限公司 Financial institution recommendation method, device, storage medium and device
CN109388756A (en) * 2018-09-10 2019-02-26 浙江口碑网络技术有限公司 Information recommendation method and device
CN109522475A (en) * 2018-10-26 2019-03-26 浙江工业大学之江学院 A kind of merchant recommendation method based on user's history consumption data
CN109522475B (en) * 2018-10-26 2022-04-22 浙江工业大学之江学院 Merchant recommendation method based on user historical consumption data
CN110321490A (en) * 2019-07-12 2019-10-11 科大讯飞(苏州)科技有限公司 Recommended method, device, equipment and computer readable storage medium
CN111402003A (en) * 2020-03-13 2020-07-10 第四范式（北京）技术有限公司 System and method for realizing user-related recommendation
CN112071401A (en) * 2020-09-05 2020-12-11 苏州贝基电子科技有限公司 Healthy diet management system based on big data
CN113297496A (en) * 2021-06-18 2021-08-24 中山市力泰电子工业有限公司 Collaborative filtering recommendation algorithm based on improved user similarity
CN113449210A (en) * 2021-07-01 2021-09-28 深圳市数字尾巴科技有限公司 Personalized recommendation method and device based on space-time characteristics, electronic equipment and storage medium
CN113449210B (en) * 2021-07-01 2023-01-31 深圳市数字尾巴科技有限公司 Personalized recommendation method and device based on space-time characteristics, electronic equipment and storage medium
US11869015B1 (en) 2022-12-09 2024-01-09 Northern Trust Corporation Computing technologies for benchmarking

Similar Documents

Publication Publication Date Title
CN105025091A (en) Shop recommendation method based on position of mobile user
Dao et al. A novel recommendation model of location-based advertising: Context-Aware Collaborative Filtering using GA approach
US20110218859A1 (en) Method, Apparatus and System for Increasing Website Data Transfer Speed
CN105095219B (en) Micro-blog recommendation method and terminal
US20110010245A1 (en) Location-based advertising method and system
CN106202331A (en) The commending system of secret protection and operational method based on this commending system by different level
CN103995823A (en) Information recommending method based on social network
CN102695121A (en) Method and system for pushing friend information for user in social network
CN102999588A (en) Method and system for recommending multimedia applications
CN103077220A (en) User group correlation degree-based personalized recommendation method and system
US20100131335A1 (en) User interest mining method based on user behavior sensed in mobile device
CN106846053A (en) A kind of recommendation method and device of the page advertisement that is polymerized
CN104268155A (en) Immediate mobile terminal personalized recommendation feedback and optimization method and device
US9734519B2 (en) Native advertisement smart injection
CN105808680A (en) Tensor decomposition based context-dependent position recommendation method
CN108595493A (en) Method for pushing and device, storage medium, the electronic device of media content
CN110473040A (en) A kind of Products Show method and device, electronic equipment
CN104794635A (en) Cellphone model recommendation system based on collaborative filtering recommendation algorithm
CN102959539B (en) Item recommendation method during a kind of repeat in work and system
CN104111959B (en) Service recommendation method based on social networks
US20110258131A1 (en) System and method for sharing information
CN106302610A (en) User interaction method based on geographical position and device
CN112215664A (en) Information recommendation method and device
CN101925044A (en) Method for judging wireless user attributes
CN116304332A (en) Knowledge graph mixed recommendation method

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20151104

RJ01 Rejection of invention patent application after publication