CN109615466A - The mixed method of commending contents and collaborative filtering recommending towards mobile ordering system - Google Patents
The mixed method of commending contents and collaborative filtering recommending towards mobile ordering system Download PDFInfo
- Publication number
- CN109615466A CN109615466A CN201811426408.XA CN201811426408A CN109615466A CN 109615466 A CN109615466 A CN 109615466A CN 201811426408 A CN201811426408 A CN 201811426408A CN 109615466 A CN109615466 A CN 109615466A
- Authority
- CN
- China
- Prior art keywords
- product
- client
- similar
- follows
- value
- 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.)
- Withdrawn
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)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A kind of mixed method of commending contents and collaborative filtering recommending towards mobile ordering system, comprising: according to client purchased product data and product collection before, obtain client characteristics vector sum client to the preference of product feature;Depth is liked according to client characteristics vector sum client characteristics, and the similarity between client is calculated using cosine similarity, client is ranked up according to the size of similar value;According to the product category that similar client buys, is scored based on product, sort from high to low according to score value size, product is arranged into Candidate Recommendation set;Various products set is calculated, consequently recommended set is obtained.The characteristics of present invention combination catering product, realizes that businessman to the personalized recommendation of user, reduces user's scoring and the few limitation of data volume, strengthens the mobile managerial ability made a reservation, recommend to meet its products & services actually required to user to ensure that.
Description
Technical field
The present invention relates to mobile terminal apparatus fields, and the food and drink that mobile phone is made a reservation under in particular to a kind of mobile internet environment
Recommended method.
Technical background
With the development of social economy and science and technology, be hungry, Meituan such as takes out at the mobile phones APP that makes a reservation becomes increasingly to flow
Row, people, which do not have to trip, can select oneself desired food and drink, this greatly facilitates people's lives.It is deposited in catering industry
In such a phenomenon, many food management systems more fall behind, and are not able to satisfy the actual needs of shop administrative staff and client,
Hinder the development of catering industry.In order to effectively solve this problem, recommender system can be introduced in catering trade management.Pass through
Collection and anatomy to consumer's data before, mouse out user to the preference of product and service, give the more satisfied clothes of user
Business, facilitates purchase of the user to food and drink, it helps administrative staff preferably manage catering system.
In traditional recommended method, the most common recommendation for being namely based on content and Collaborative Filtering Recommendation Algorithm.It is based on
The recommendation of content is to recommend the production similar with original purchased product content to user according to the purchase history data before user
Product.Collaborative Filtering Recommendation Algorithm is the product preference record according to user before, finds other use similar with this user interest
Family recommends these other interested product contents of user to this user.Both proposed algorithms all respectively have certain
Limitation, content-based recommendation algorithm cannot specifically obtain the preference profile of other users, institute in Collaborative Filtering Recommendation Algorithm
Data are few, may recommend inaccuracy.Since scoring dependence of the conventional recommendation algorithm to user is very strong, and actually score
User is seldom, results in and has no idea accurately to be recommended.By mixing both proposed algorithms, the standard of recommendation can be improved
True property recommends user's products & services actually required.
Summary of the invention
The present invention in order to overcome the shortcomings of conventional recommendation method, provide a kind of commending contents towards mobile ordering system and
The mixed method of collaborative filtering recommending, comprising:
Step 1. obtains client characteristics vector sum client couple according to client purchased product data and product collection before
The preference of product feature, method include:
A. by product feature value fiIt indicates, by client to fiScoring eiIt indicates, by f in client's purchased productiOccur
Number tiIt indicates, n is product sum, by the analysis to client purchased product data and product collection before, eiIt is
By fiPurchased product number and the ratio between product sum, be formulated are as follows:
Although product is put into the favorites list and is suffered if client does not buy certain product, client is reflected from side
Still there is certain interest to the product, this can also give the characteristic value bonus point of product;
B. client's purchased product have price feature and product classification feature, two characteristic values be all discrete random variable and
It is denoted as Y1And Y2, it is assumed that the standard deviation of the two features is respectively σ1And σ2, mean value is respectively E (Y1) and E (Y2), then two features
The coefficient of variation is respectively as follows:
C. client is respectively as follows: the favorable rating of the two features
Here Q1 adds Q2 to be equal to 1;
D. client characteristics vector is multidimensional, is represented by W=(e with vector1,e2,e3,e4,e5,……,ei), this to
Amount illustrates the preference of client;
Step 2. is calculated between client according to client characteristics vector sum client characteristics fancy grade using cosine similarity
Then similar value is ranked up client according to the size of similar value, calculate the similitude between client using cosine similarity
Method include: given two clients feature vector, be represented sequentially as X1And X2, then the similar value of two clients formula table
It is shown as:
Here | | X | | and | | Y | | the length for respectively indicating two client characteristics vectors, the similarity ranges provided are from 0
To 1, wherein 0 two clients of expression are independent relationships, 1 indicates indifference between two clients;
The product category that step 3. is bought according to similar client is scored based on product, is arranged from high to low according to score value size
Sequence arranges product into Candidate Recommendation set, the method based on product scoring are as follows: after similar users are calculated, give phase
Like client's purchased product score value, the formula for use of giving a mark are as follows:
S=c (X1,X2)×(Q1×ea+Q2×eb) (7)
Here S indicates that client gives the score value of similar client's purchased product, c (X1,X2) indicate X1With X2Between it is similar
Degree, eaAnd ebIt is scoring of the current user to two feature values of product;
Step 4. calculates various products set, obtains consequently recommended set, the method for calculating are as follows:
P (U)=P (L)-P (L ∩ N) (8)
P (W)=P (N) (9)
P (V)=P (E)-P (L ∪ N) (10)
Here P refers to certain product set, and P (E) refers to that all product complete or collected works, P (L) refer to that recommended products is waited
Selected works close, and P (N) is the set for the product that client bought;P (U), P (V), P (W) be three priority levels setting up from height to
The low set being arranged successively, wherein P (U) is the product set that do not bought in recommended products candidate collection, and P (W) is to have purchased
The product set bought, P (V) are indicated neither recommended candidate product is also not the set for the product bought;
A kind of commending contents and collaborative filtering recommending mixing towards mobile ordering system provided by the invention in summary
Method, and the characteristics of combining catering product, realizes that businessman to the personalized recommendation of user, reduces user's scoring and data volume
Few limitation strengthens the mobile managerial ability made a reservation, and recommends to meet its product actually required to user to ensure that
And service, for solve the problems, such as under mobile internet environment mobile phone make a reservation management system more it is backward have preferable improve effect
Fruit.
Detailed description of the invention
Fig. 1 show the mixed method flow chart of commending contents and collaborative filtering recommending towards mobile ordering system.
The user that Fig. 2 show the mixed method of commending contents and collaborative filtering recommending towards mobile ordering system comments
Sub-matrix figure.
Specific embodiment
Explanation and specific embodiment are described in further details the present invention with reference to the accompanying drawing;
The method of commending contents and collaborative filtering recommending mixing of the present invention towards mobile ordering system, with Fig. 1
In framework be designed, include the following steps:
Step 1. obtains client characteristics vector sum client couple according to client purchased product data and product collection before
The preference of product feature, method include:
A. by product feature value fiIt indicates, by client to fiScoring eiIt indicates, by f in client's purchased productiOccur
Number tiIt indicates, n is product sum, by the analysis to client purchased product data and product collection before, eiIt is
By fiPurchased product number and the ratio between product sum, be formulated are as follows:
Although product is put into the favorites list and is suffered if client does not buy certain product, client is reflected from side
Still there is certain interest to the product, this can also give the characteristic value bonus point of product;
B. client's purchased product have price feature and product classification feature, two characteristic values be all discrete random variable and
It is denoted as Y1And Y2, it is assumed that the standard deviation of the two features is respectively σ1And σ2, mean value is respectively E (Y1) and E (Y2), then two features
The coefficient of variation is respectively as follows:
C. client is respectively as follows: the favorable rating of the two features
Here Q1 adds Q2 to be equal to 1;
D. client characteristics vector is multidimensional, is represented by W=(e with vector1,e2,e3,e4,e5,……,ei), this to
Amount illustrates the preference of client;
Step 2. is calculated between client according to client characteristics vector sum client characteristics fancy grade using cosine similarity
Then similar value is ranked up client according to the size of similar value, calculate the similitude between client using cosine similarity
Method include: given two clients feature vector, be represented sequentially as X1And X2, then the similar value of two clients formula table
It is shown as:
Here | | X | | and | | Y | | the length for respectively indicating two client characteristics vectors, the similarity ranges provided are from 0
To 1, wherein 0 two clients of expression are independent relationships, 1 indicates indifference between two clients;
The product category that step 3. is bought according to similar client is scored based on product, is arranged from high to low according to score value size
Sequence arranges product into Candidate Recommendation set, the method based on product scoring are as follows: after similar users are calculated, give phase
Like client's purchased product score value, the formula for use of giving a mark are as follows:
S=c (X1,X2)×(Q1×ea+Q2×eb) (7)
Here S indicates that client gives the score value of similar client's purchased product, c (X1,X2) indicate X1With X2Between it is similar
Degree, eaAnd ebIt is scoring of the current user to two feature values of product;
Step 4. calculates various products set, obtains consequently recommended set, the method for calculating are as follows:
P (U)=P (L)-P (L ∩ N) (8)
P (W)=P (N) (9)
P (V)=P (E)-P (L ∪ N) (10)
Here P refers to certain product set, and P (E) refers to that all product complete or collected works, P (L) refer to that recommended products is waited
Selected works close, and P (N) is the set for the product that client bought;P (U), P (V), P (W) be three priority levels setting up from height to
The low set being arranged successively, wherein P (U) is the product set that do not bought in recommended products candidate collection, and P (W) is to have purchased
The product set bought, P (V) are indicated neither recommended candidate product is also not the set for the product bought;
A kind of commending contents and collaborative filtering recommending mixing towards mobile ordering system provided by the invention in summary
Method, and the characteristics of combining catering product, realizes that businessman to the personalized recommendation of user, reduces user's scoring and data volume
Few limitation strengthens the mobile managerial ability made a reservation, and recommends to meet its product actually required to user to ensure that
And service, for solve the problems, such as under mobile internet environment mobile phone make a reservation management system more it is backward have preferable improve effect
Fruit;
Although the present invention be to move the corresponding mixed recommendation technique study being unfolded based on ordering system, it is of the invention
Be not restricted to this, the present invention can also according to applying among other field, solve the problems, such as it is more, therefore, this hair
Bright to have practicability, protection scope is subject to view claims range claimed.
Claims (1)
1. the mixed method of commending contents and collaborative filtering recommending towards mobile ordering system, includes the following steps:
Step 1. obtains client characteristics vector sum client to product according to client purchased product data and product collection before
The preference of feature, method include:
A. by product feature value fiIt indicates, by client to fiScoring eiIt indicates, by f in client's purchased productiTime occurred
Number uses tiIt indicates, n is product sum, by the analysis to client purchased product data and product collection before, eiIt is by fi
Purchased product number and the ratio between product sum, be formulated are as follows:
Although product is put into the favorites list and is suffered if client does not buy certain product, reflect client to this from side
Product still has certain interest, this can also give the characteristic value bonus point of product;
B. client's purchased product has price feature and product classification feature, and two characteristic values are all discrete random variables and are denoted as
Y1And Y2, it is assumed that the standard deviation of the two features is respectively σ1And σ2, mean value is respectively E (Y1) and E (Y2), then the variation of two features
Coefficient is respectively as follows:
C. client is respectively as follows: the favorable rating of the two features
Here Q1 adds Q2 to be equal to 1;
D. client characteristics vector is multidimensional, is represented by W=(e with vector1,e2,e3,e4,e5,……,ei), this vector is said
The preference of bright client;
Step 2. is calculated similar between client according to client characteristics vector sum client characteristics fancy grade using cosine similarity
Value, is then ranked up client according to the size of similar value, and the side of the similitude between client is calculated using cosine similarity
Method includes: the feature vector of given two clients, is represented sequentially as X1And X2, then the similar value of two clients is formulated are as follows:
Here | | X | | and | | Y | | the length for respectively indicating two client characteristics vectors, the similarity ranges provided from 0 to 1,
Wherein 0 two clients of expression are independent relationships, and 1 indicates indifference between two clients;
The product category that step 3. is bought according to similar client is scored based on product, is sorted from high to low according to score value size, will
Product is arranged into Candidate Recommendation set, the method based on product scoring are as follows: after similar users are calculated, give similar client
Purchased product score value, the formula for use of giving a mark are as follows:
S=c (X1,X2)×(Q1×ea+Q2×eb) (7)
Here S indicates that client gives the score value of similar client's purchased product, c (X1,X2) indicate X1With X2Between similarity, eaWith
ebIt is scoring of the current user to two feature values of product;
Step 4. calculates various products set, obtains consequently recommended set, the method for calculating are as follows:
P (U)=P (L)-P (L ∩ N) (8)
P (W)=P (N) (9)
P (V)=P (E)-P (L ∪ N) (10)
Here P refers to certain product set, and P (E) refers to that all product complete or collected works, P (L) refer to recommended products Candidate Set
It closes, P (N) is the set for the product that client bought;P (U), P (V), P (W) be three priority levels setting up from high to low according to
The set of secondary arrangement, wherein P (U) is the product set that do not bought in recommended products candidate collection, and P (W) is to have bought
Product set, P (V) indicates neither recommended candidate product is also not the set for the product bought.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811426408.XA CN109615466A (en) | 2018-11-27 | 2018-11-27 | The mixed method of commending contents and collaborative filtering recommending towards mobile ordering system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811426408.XA CN109615466A (en) | 2018-11-27 | 2018-11-27 | The mixed method of commending contents and collaborative filtering recommending towards mobile ordering system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109615466A true CN109615466A (en) | 2019-04-12 |
Family
ID=66005305
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811426408.XA Withdrawn CN109615466A (en) | 2018-11-27 | 2018-11-27 | The mixed method of commending contents and collaborative filtering recommending towards mobile ordering system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109615466A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110287410A (en) * | 2019-06-05 | 2019-09-27 | 达疆网络科技(上海)有限公司 | The fusion method of a variety of proposed algorithms of user under a kind of O2O electric business scene |
CN110930184A (en) * | 2019-11-14 | 2020-03-27 | 杭州天宽科技有限公司 | Potential customer mining and customer type selection method based on mixed recommendation algorithm |
CN112733067A (en) * | 2020-12-22 | 2021-04-30 | 上海机器人产业技术研究院有限公司 | Data set selection method for robot target detection algorithm |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020178057A1 (en) * | 2001-05-10 | 2002-11-28 | International Business Machines Corporation | System and method for item recommendations |
CN102567900A (en) * | 2011-12-28 | 2012-07-11 | 尚明生 | Method for recommending commodities to customers |
CN102968506A (en) * | 2012-12-14 | 2013-03-13 | 北京理工大学 | Personalized collaborative filtering recommendation method based on extension characteristic vectors |
CN104424247A (en) * | 2013-08-28 | 2015-03-18 | 北京闹米科技有限公司 | Product information filtering recommendation method and device |
-
2018
- 2018-11-27 CN CN201811426408.XA patent/CN109615466A/en not_active Withdrawn
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020178057A1 (en) * | 2001-05-10 | 2002-11-28 | International Business Machines Corporation | System and method for item recommendations |
CN102567900A (en) * | 2011-12-28 | 2012-07-11 | 尚明生 | Method for recommending commodities to customers |
CN102968506A (en) * | 2012-12-14 | 2013-03-13 | 北京理工大学 | Personalized collaborative filtering recommendation method based on extension characteristic vectors |
CN104424247A (en) * | 2013-08-28 | 2015-03-18 | 北京闹米科技有限公司 | Product information filtering recommendation method and device |
Non-Patent Citations (1)
Title |
---|
许翀寰 等: "基于协同过滤的推荐方法", 《浙江工商大学出版社》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110287410A (en) * | 2019-06-05 | 2019-09-27 | 达疆网络科技(上海)有限公司 | The fusion method of a variety of proposed algorithms of user under a kind of O2O electric business scene |
CN110930184A (en) * | 2019-11-14 | 2020-03-27 | 杭州天宽科技有限公司 | Potential customer mining and customer type selection method based on mixed recommendation algorithm |
CN112733067A (en) * | 2020-12-22 | 2021-04-30 | 上海机器人产业技术研究院有限公司 | Data set selection method for robot target detection algorithm |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11216861B2 (en) | Color based social networking recommendations | |
US10402917B2 (en) | Color-related social networking recommendations using affiliated colors | |
CN103246980B (en) | Information output method and server | |
TWI534733B (en) | Based on interpersonal distance ranking, search methods and devices | |
CA2846025C (en) | Recommendations based upon explicit user similarity | |
CN107194774A (en) | Personalized hotel's commending system and method in OTA websites | |
CN109815402A (en) | Collaborative Filtering Recommendation Algorithm based on user characteristics | |
EP1156424A2 (en) | Information recommendation apparatus and information recommendation system | |
US20130297608A1 (en) | Content presentation device, content presentation terminal, content presentation system, content presentation program, and content presentation method | |
CN102063433A (en) | Method and device for recommending related items | |
CN109615466A (en) | The mixed method of commending contents and collaborative filtering recommending towards mobile ordering system | |
CN103858142A (en) | Store information provision system | |
CN108876537A (en) | A kind of mixed recommendation method for on-line mall system | |
CN106408377A (en) | Shopping recommended method and system | |
CN104731866B (en) | Personalized food recommendation method based on position | |
CN107329994A (en) | A kind of improvement collaborative filtering recommending method based on user characteristics | |
CN110619559B (en) | Method for accurately recommending commodities in electronic commerce based on big data information | |
CN106980646A (en) | The method applied based on popularity to the influencing mechanism analysis of user interest and its in proposed algorithm | |
CN107679239A (en) | Recommend method in a kind of personalized community based on user behavior | |
Gaberamos et al. | The effect of information quality, customer experience, price, and service quality on purchase intention by using customer perceived value as mediation variables (Study On Gofood Applications On The Millenial Generation) | |
CN108241619A (en) | A kind of recommendation method based on the more interest of user | |
CN104008193B (en) | A kind of information recommendation method based on group of typical user discovery technique | |
CN106354855A (en) | Recommendation method and system | |
CN109299360B (en) | Dish recommending method | |
Peemanee et al. | Technology acceptance and service convenience on repeat purchase decision of food delivery business in Thailand |
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 | ||
WW01 | Invention patent application withdrawn after publication | ||
WW01 | Invention patent application withdrawn after publication |
Application publication date: 20190412 |