CN102592223A - Commodity recommending method and commodity recommending system - Google Patents
Commodity recommending method and commodity recommending system Download PDFInfo
- Publication number
- CN102592223A CN102592223A CN2011100206496A CN201110020649A CN102592223A CN 102592223 A CN102592223 A CN 102592223A CN 2011100206496 A CN2011100206496 A CN 2011100206496A CN 201110020649 A CN201110020649 A CN 201110020649A CN 102592223 A CN102592223 A CN 102592223A
- Authority
- CN
- China
- Prior art keywords
- user
- commercial product
- door
- commodity
- product recommending
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The invention relates to a commodity recommending method and commodity recommending system. The method comprises obtaining sample training data based on browsing history of users and/or user attributes, and establishing an associated analysis model for the sample training data; generating associated data based on the associated analysis model; accepting commodity recommending requests, and obtaining user information based on the commodity recommending requests, recommending commodities associated with the user information to relevant portals based on the user information and the associated data, so that relevant portals can show commodities to users. The commodity recommending system comprises a modeling unit and a commodity recommending server. The technical scheme not only guarantees recommending qualities, but also provides high-speed real-time recommending service, and has the advantages of being high in expandability, and suitable for different application contexts.
Description
Technical field
The present invention relates to the commodity application, more particularly, relate to a kind of commercial product recommending method and commercial product recommending system.
Background technology
The commercial product recommending dissemination system can be recommended appropriate commodity to the user in appropriate scene.Common like Technologies of Recommendation System in E-Commerce, recommend extensive stock to the user on the internet, as recommend the new commodity of putting on the shelf, the commodity of discounting, fast-selling commodity.Technologies of Recommendation System in E-Commerce on the present internet; Generally be based on merchandise sales seniority among brothers and sisters, user commercial product recommending is carried out in the evaluation scoring of commodity; Owing to lack the intellectual analysis factor to the required use of user-customized recommended, the commodity of therefore recommending the user under many circumstances are not that the user likes.Present Technologies of Recommendation System in E-Commerce mainly contains following defective:
1) equilibrium problem between real-time and the recommendation quality.The recommendation precision of commending system and real-time are a pair of contradiction, and most of commending system is to be prerequisite with the recommendation quality of sacrificing commending system when guaranteeing that real-time requires.When real-time recommendation service is provided, how effectively to improve the recommendation quality of commending system, be the difficult problem that commercial product recommending system faces.
2) current Technologies of Recommendation System in E-Commerce architecture imperfection, most Technologies of Recommendation System in E-Commerce all is a single instrument, and a kind of recommended models can only be provided.But because the complicacy of e-commerce system itself, different occasions need dissimilar recommendations.
3) present Technologies of Recommendation System in E-Commerce can only be through simple sales list, provide other users that the modes such as evaluation score information of commodity are achieved the above object to the user; Need further study more effective method produces the reason of recommending to user interpretation; Thereby increase the trusting degree of user to commending system, the persuasion user accepts the recommendation of commending system.
4) object of recommendation service to be provided nearly all be the user that commodity are bought in registration, and ignored the visitor that those access sites are not bought thing; And lack consideration to website expert and analyst's directive function, only be simple sales list.
Summary of the invention
The technical matters that the present invention will solve is, the single defective of object of, real-time difference of low quality to the above-mentioned commercial product recommending of prior art and recommendation service provides a kind of commercial product recommending method and commercial product recommending system.
The technical solution adopted for the present invention to solve the technical problems is: construct a kind of commercial product recommending method, comprising:
According to the user browse record and/or user property obtains the sample training data, and said sample training data are set up the association analysis model;
Generate associated data according to the association analysis model;
Receive the commercial product recommending request,, will give relevant door with the commercial product recommending that said user profile is associated based on said user profile and said associated data, so that be shown to the user by relevant door according to said commercial product recommending acquisition request user profile.
In commercial product recommending method of the present invention, generate associated data according to the association analysis model and specifically comprise:
Generate associated data according to the association analysis model and based on support, degree of confidence, lifting degree.
In commercial product recommending method of the present invention, said user profile comprises user terminal type, the commodity of browsing; Browse or during subscription status, will give relevant door with the commercial product recommending that said user profile is associated when the user has logined and has been in, specifically comprise so that be shown to the user by relevant door based on said user profile and said associated data:
To give relevant door with the commercial product recommending of the said commodity corresponding associated of browsing based on said user terminal type, the commodity of browsing and said associated data, so that be shown to the user by relevant door; Said door comprises WWW door, WAP door, user terminal door.
In commercial product recommending method of the present invention,, then from preset commodity storehouse, choose the commodity that are suitable for said user terminal type if it is not enough to recommend user's commodity amount.
In commercial product recommending method of the present invention, said user profile comprises the commodity of browsing; Do not browse or during subscription status, will give relevant door with the commercial product recommending that said user profile is associated when the user logins and is in, specifically comprise so that be shown to the user by relevant door based on said user profile and said associated data:
To give relevant door with the commercial product recommending of the said commodity corresponding associated of browsing based on said commodity of browsing and said associated data, so that be shown to the user by relevant door; Said door comprises the WWW door.
In commercial product recommending method of the present invention, browse and when not having subscription status, to generate associated data according to the association analysis model and specifically comprise when the user has logined and has been in not have:
Divide a plurality of customer groups based on said user property, and each customer group is provided with corresponding identification, generate associated data according to the association analysis model.
In commercial product recommending method of the present invention, said user profile comprises user terminal type, user property; Browse and when not having subscription status, will give relevant door with the commercial product recommending that said user profile is associated when the user has logined and has been in not have, specifically to comprise so that be shown to the user by relevant door based on said user profile and said associated data:
User property, user terminal type and said associated data based on logged-in user are carried out the customer group classification logotype to the user; Give relevant door with other user preferences among the same subscriber crowd and with the commercial product recommending that said user terminal type is complementary, so that be shown to the user by relevant door; Said door comprises WWW door, WAP door, user terminal door.
The present invention also constructs a kind of commercial product recommending system, comprising:
The modelling unit, be used for according to the user browse record and/or user property obtains the sample training data, and said sample training data are set up the association analysis model;
The commercial product recommending server is connected with said modelling unit communication, comprising:
Generation unit is used for generating associated data according to the association analysis model;
Receiving element is used to receive the commercial product recommending request;
Recommendation unit is used for according to said commercial product recommending acquisition request user profile, will give relevant door with the commercial product recommending that said user profile is associated based on said user profile and said associated data, so that be shown to the user by relevant door.
In commercial product recommending system of the present invention, said user profile comprises user terminal type, the commodity of browsing; When having logined and be in, the user browses or during subscription status; Said recommendation unit specifically is used for: according to said commercial product recommending acquisition request user profile; To give relevant door with the commercial product recommending of the said commodity corresponding associated of browsing based on said user terminal type, the commodity of browsing and said associated data, so that be shown to the user by relevant door.
In commercial product recommending system of the present invention, said user profile comprises the commodity of browsing; When logining and be in, the user do not browse or during subscription status; Said recommendation unit specifically is used for: according to said commercial product recommending acquisition request user profile; To give relevant door with the commercial product recommending of the said commodity corresponding associated of browsing based on said commodity of browsing and said associated data, so that be shown to the user by relevant door.
Adopt technical scheme of the present invention; Have following beneficial effect: 1) door only need provide real-time service information to commercial product recommending system; Commercial product recommending system just can be given operation system with Recommendations information through setting up the association analysis model and generating associated data.
2) with the WWW door, the WAP door, the mobile phone terminal door combines simultaneously, uses same commercial product recommending system, and ecommerce intelligent recommendation system is become a reality.
3) be different from common search engine; Also be different from the Technologies of Recommendation System in E-Commerce that the modes such as evaluation score information that other users' commodity can only be provided through simple sales list, to the user reach the persuasion customer objective, but the several data digging technology is applied to a plurality of recommendation scenes.
4) with the user mobile phone model also as a Consideration of commercial product recommending.
5) the commercial product recommending method and system had both guaranteed the recommendation quality, and the recommendation service of real time high-speed is provided again.
6) extensibility of commercial product recommending system is strong, can also constantly increase to use other data digging methods or statistical analysis technique, at more scene Recommendations.
Description of drawings
To combine accompanying drawing and embodiment that the present invention is described further below, in the accompanying drawing:
Fig. 1 is the commercial product recommending method flow synoptic diagram according to one embodiment of the invention;
Fig. 2 is the commercial product recommending system structural representation according to one embodiment of the invention;
Fig. 3 is the commercial product recommending server architecture synoptic diagram among Fig. 2.
Embodiment
The present invention to the user at the WWW door; The WAP door, the data mining technology off-line modeling is used in having landed order (browsing), not having to have landed to browse, land not have to browse and do not have the scene of order of mobile phone terminal door; Calculate in conjunction with online recommendation; Use the Memcached caching technology that recommendation results is carried out caching process, to the different scenes of different doors distinct interface is provided, provide the Recommendations data to return to each big door and represent at same recommendation dissemination system; Introduce modern data mining data analysis mechanism such as association, cluster, need not user's input, just can understand user preference; According to business scenario and business support needs, can also use other data mining technology and statistical method etc. to arrange new communication interface for each big online door or off-line system.
Fig. 1 is commercial product recommending method 100 schematic flow sheets according to one embodiment of the invention.Commercial product recommending method 100 comprises the steps:
In step 101, according to the user browse record and/or user property obtains the sample training data, and said sample training data are set up the association analysis model;
In step 102, generate associated data according to the association analysis model, also can be described as correlation rule; The association analysis model here comprises the model that adopts association analysis modeling method or cluster analysis modeling method to set up, and hereinafter will be done detailed elaboration.
In step 103; Receive the commercial product recommending request; According to said commercial product recommending acquisition request user profile, will give relevant door with the commercial product recommending that said user profile is associated based on said user profile and said associated data, so that be shown to the user by relevant door.
In an embodiment of the present invention, step 102 specifically comprises: generate associated data according to the association analysis model and based on support, degree of confidence, lifting degree.Example about support, degree of confidence, lifting degree is as shown in table 1:
Table 1
| Correlation rule | The lifting degree | Support | Degree of confidence | The rule left side | Rule the right |
| Sound worm mobile phone qin=＞the mobile phone torch | 2.48 | 13.9 | 54.47 | Sound worm mobile phone qin | The mobile phone torch |
| Push box=＞number is solely | 4.97 | 4.47 | 59.61 | Push box | Number solely |
Table 1 is that example is set forth with the mobile phone application software, and " sound worm mobile phone qin ", " mobile phone torch ", " pushing box ", " number solely " all are the Games Software titles.The present invention is not limited to software application, can also be the commodity of other reality.
The lifting degree, the multiple that adopts this rule to improve---the validity (deeply) of rule than ratio at population sample;
Support, regular right sides appears at the ratio in all samples simultaneously---the coverage rate (extensively) of rule;
Degree of confidence, rule the right in the conditional probability on the regular left side---regular metric.
Article one, the explanation of correlation rule: ordered and used sound worm mobile phone qin and ordered the user who uses the mobile phone torch simultaneously and account for 13.9% of total customer volume; In having ordered the total user who uses sound worm mobile phone qin, there is 54.47% user to order application mobile phone torch; In having ordered the user who uses sound worm mobile phone qin, also order the probability of using the mobile phone torch, be to order 2.48 times of the probability of using the mobile phone torch among total user.
The explanation of second correlation rule: ordered that using pushes box and ordered simultaneously and used the only user of number and account for 4.47% of total customer volume; Use that to have 59.61% user to order the application number among the total user who pushes box only having ordered; Using among the user who pushes box having ordered, also order and use the only probability of number, is to order 4.97 times that use the only probability of number among total user.
In an embodiment of the present invention, the user profile of Fig. 1 comprises user terminal type, the commodity of browsing.Browse or during subscription status, will give relevant door with the commercial product recommending that said user profile is associated based on said user profile and said associated data in the step 103 when the user has logined and has been in, specifically comprise so that be shown to the user by relevant door:
To give relevant door with the commercial product recommending of the said commodity corresponding associated of browsing based on said user terminal type, the commodity of browsing and said associated data; So that be shown to the user by relevant door, said door comprises WWW door, WAP door, user terminal door.The associated data here is meant the commodity association data.
In an embodiment of the present invention, if it is not enough to recommend user's commodity amount, then from preset commodity storehouse, choose the commodity that are suitable for said user terminal type.The preset commodity storehouse here can be for example success of type coupling and the forward commodity of quantity ordered rank, such as but not limited to 40 commodity before the rank.
Logined and be in about the user and browsed or the application scenarios of subscription status, details are as follows for the process of commercial product recommending method of the present invention:
S1) after the business analyst cleans, filters, changes through the historical data of all users being ordered the behaviors of (browsing) commodity, get 70% sample data sample training data as the association analysis model.Use Data Mining Tools, the sample training data are carried out the association analysis model modeling with the association analysis algorithm.
S2) after model training is accomplished and drawn model result, model result is assessed, chosen effect model preferably, reduced model result (being correlation rule) imports to intelligent distribution platform with correlation rule then.
S3) correlation rule of intelligent distribution platform (being the commercial product recommending server that hereinafter is described, down together) through model is generated confirmed from support, degree of confidence, lifting degree, operational angle, obtains the professional correlation rule that uses.
S4) when the user after the WWW door/the WAP door/the terminal door lands, when ordering (browsing) commodity page, need recommend and related other commodity of these commodity to the user;
S5) WWW door/WAP door/terminal door is to the commodity of intelligent distribution platform request recommendation;
The intelligence distribution platform is prepared Recommendations according to Subscriber Number, user's type, the commodity browsed for the user, and the rule of recommendation as follows;
A, according to the commodity of user's type, order (browsing), through correlation rule, giving the user with the corresponding related commercial product recommending of the commodity of order (browsing); The commodity of ordering in this nearest a period of time are no longer recommended; If need to recommend a plurality of commodity, then need choose many corresponding correlation rules of commodity with order (browsing).
Correlation rule selection principle: correlation rule is chosen (getting 5 such as n) according to degree of confidence TOPn
If b, the commodity amount of under a rule, recommending is not enough, the commodity polishing of the adaptive user terminal of picked at random from the elaboration storehouse then, the commodity that the user ordered in three months reach the commodity of recommending in a step and no longer recommend.
The Rule of judgment of elaboration commodity: adaptive success of type and quantity ordered are at TOP40.
C, can't obtain the commodity of sufficient amount like the b step, then directly return can recommended amount commodity.
If d can't discern user's type, Recommendations not then.
In further embodiment of this invention, the user profile of Fig. 1 comprises the commodity of browsing.Do not browse or during subscription status, will give relevant door with the commercial product recommending that said user profile is associated based on said user profile and said associated data in the step 103 when the user logins and is in, specifically comprise so that be shown to the user by relevant door:
To give relevant door with the commercial product recommending of the said commodity corresponding associated of browsing based on said commodity of browsing and said associated data, so that be shown to the user by relevant door; Said door is the WWW door.Because the user on WAP door/terminal door can get access to user's phone number, this application scenarios is inapplicable.The associated data here is meant the commodity association data.
Do not have login and be in about the user and browse or the application scenarios of subscription status, details are as follows for the process of commercial product recommending method of the present invention:
S1 ') after the business analyst cleans, filters, changes through the historical data of all users being ordered the behaviors of (browsing) commodity, gets 70% sample data sample training data as the association analysis model.Use Data Mining Tools, the sample training data are carried out the association analysis model modeling with the association analysis algorithm.
S2 ') after model training is accomplished and drawn model result, model result is assessed, chosen effect model preferably, reduced model result (being correlation rule) imports to intelligent distribution platform with correlation rule then.
S3 ') correlation rule of intelligent distribution platform through model is generated confirmed from support, degree of confidence, lifting degree, operational angle, obtains the professional correlation rule that uses.
S4 ') as user during, need recommend and related other commodity of these commodity to the user at the WWW door browse commodities page;
S5 ') the WWW door is to the commodity of intelligent distribution platform request recommendation;
S6 ') intelligent distribution platform is prepared Recommendations according to the commodity that the user browses for the user, and the rule of recommendation as follows;
A, the commodity of browsing according to the user are through correlation rule, giving the user with the corresponding related commercial product recommending of the commodity of browsing; If need to recommend a plurality of commodity, then need choose many correlation rules corresponding with the commodity of browsing.
Correlation rule selection principle: correlation rule is chosen (getting 5 such as n) according to degree of confidence TOPn
If b, the commodity amount deficiency of under a rule, recommending, picked at random commodity polishing during then the elaboration from similar commodity is used is no longer recommended at the commodity that a step was recommended.
The Rule of judgment of elaboration commodity: quantity ordered TOP40 in the similar commodity.
S7 ') intelligent distribution platform returns the application that distribution is returned according to intelligence of WWW door to exemplary application, finds the commodity of using correspondence to recommend to the user.
In yet another embodiment of the invention; When having logined and be in not have, the user browses and when not having subscription status; Step 102 among Fig. 1 specifically comprises: divide a plurality of customer groups based on said user property, and each customer group is provided with corresponding identification, generate associated data according to the association analysis model.Association analysis model in the step 102 just is embodied as the cluster analysis model, and the associated data here is the tenant group data, adopts the cluster analysis model to obtain the tenant group data.
In this embodiment, user profile comprises user terminal type, user property; When having logined and be in not have, the user browses and when not having subscription status, step 103 will be given relevant door with the commercial product recommending that said user profile is associated based on said user profile and said associated data, specifically comprise so that be shown to the user by relevant door:
User property, user terminal type and said associated data based on logged-in user are carried out the customer group classification logotype to the user; Give relevant door with other user preferences among the same subscriber crowd and with the commercial product recommending that said user terminal type is complementary, so that be shown to the user by relevant door; Said door comprises WWW door, WAP door, user terminal door.
Having logined and be in about the user does not have the application scenarios browse and do not have subscription status, and details are as follows for the process of commercial product recommending method of the present invention:
S1 ") business analyst through to all UADs (whether comprise the Fetion user, whether the handset customization user, whether group user, whether use weather forecast, belong to which brand, the tariff package expense) clean, filter, change after, get 70% sample data sample training data as the cluster analysis model.Use Data Mining Tools, the sample training data are carried out the Clustering Model modeling with clustering algorithm.
S2 ") after model training accomplishes and to draw model result; model result is assessed; choose effect model preferably, reduced model result (being clustering rule), and each customer group analyzed description; identify for each colony, then the corresponding clustering rule of colony's sign and each colony is imported to intelligent distribution platform.The following form of commercial articles clustering rule:
If the user has following characteristics: be that Fetion, handset customization user, group user, use weather forecast, M-ZONE brand, tariff package are medium, then this user is identified as the preference software bundling.
S3 ") when the user at WWW door/WAP door/when the terminal door lands, the commodity that need recommend the user to like to the user;
S4 ") WWW door/WAP door/terminal door transmits user profile, request Recommendations to intelligent distribution platform;
S5 ") intelligent distribution platform prepares Recommendations according to user property, user's type for the user, and the rule of recommendation is as follows;
A, according to user property, according to clustering rule, the user is carried out classification logotype, choose that other users of same subscriber crowd like and give the user with the commercial product recommending that user's type is complementary;
The commodity selection principle that other users of same cluster like: the commodity to other users of same cluster like are chosen (getting 5 such as n) according to quantity ordered TOPn.
If b, the commodity amount of under a rule, recommending is not enough, the commodity polishing of the adaptive user terminal of picked at random was no longer recommended at the commodity that a step was recommended during then the elaboration from the commodity class that other users of same cluster like was used.
The Rule of judgment of elaboration commodity: adaptive success of type and quantity ordered are at TOP40.
C, can't obtain the commodity of sufficient amount like the b step, then directly return can recommended amount commodity.
If d can't discern user's type, Recommendations not then.
S6 ") intelligent distribution platform returns WWW door/WAP door/terminal door to exemplary application
S7 ") WWW door/WAP door/terminal door commodity that distribution is returned according to intelligence, find the corresponding commodity of user to recommend to the user.
Fig. 2 is commercial product recommending system 200 structural representations according to one embodiment of the invention.Commercial product recommending system 200 comprises modelling unit 201 and commercial product recommending server 202.Commercial product recommending server 202 connects with WWW door/WAP door/terminal door communication link.Commercial product recommending server 202 comprises generation unit 2021, receiving element 2022, recommendation unit 2023, sees Fig. 3 for details.
Commercial product recommending server 202 is connected with 201 communications of modelling unit, comprising:
Generation unit 2021 is used for generating associated data according to the association analysis model;
Receiving element 2022 is used to receive the commercial product recommending request;
Recommendation unit 2023 is used for according to said commercial product recommending acquisition request user profile, will give relevant door with the commercial product recommending that said user profile is associated based on said user profile and said associated data, so that be shown to the user by relevant door.
Commercial product recommending server 202 for example carries out data through for example communication interface between WWW door/WAP door/terminal door with relevant door and transmits.
In an embodiment of the present invention, user profile comprises user terminal type, the commodity of browsing.When having logined and be in, the user browses or during subscription status; Said recommendation unit specifically is used for: according to said commercial product recommending acquisition request user profile, will give the user with the commercial product recommending of the said commodity corresponding associated of browsing based on said user terminal type, the commodity of browsing and said associated data.
In further embodiment of this invention, user profile comprises the commodity of browsing.When logining and be in, the user do not browse or during subscription status; Said recommendation unit specifically is used for: according to said commercial product recommending acquisition request user profile; To give relevant door with the commercial product recommending of the said commodity corresponding associated of browsing based on said commodity of browsing and said associated data, so that be shown to the user by relevant door.
In yet another embodiment of the invention, user profile comprises user terminal type, user property.When having logined and be in not have, the user browses and when not having subscription status, generation unit 2021 specifically is used for: divide a plurality of customer groups based on said user property, and each customer group is provided with corresponding identification, generate associated data according to the association analysis model.
Recommendation unit 2023 specifically is used for: user property, user terminal type and said associated data based on logged-in user are carried out the customer group classification logotype to the user; Give relevant door with other user preferences among the same subscriber crowd and with the commercial product recommending that said user terminal type is complementary, so that be shown to the user by relevant door.Said door comprises WWW door, WAP door, user terminal door.
Preceding text are equally applicable to commercial product recommending system about the description of commercial product recommending method, do not do here and give unnecessary details.
Technical scheme of the present invention has the following advantages:
1, intelligent distribution platform is firstly to use the data service of doing in the store for the internet cell phone and recommend dissemination system, uses the WWW door in store to cell phone, the WAP door, and the operation of mobile phone terminal door provides data to support.
2, first traditional data digging technology and each door are combined, aspect commercial product recommending, play a role.The WWW door; The WAP door; Each operation system of mobile phone terminal door need be concerned about the numerous and complicated mixed and disorderly realization details of intelligent distribution platform, only need provide real-time service information to intelligent distribution platform according to interface, and intelligent distribution platform returns Recommendations information to operation system with the interface of agreement.
3, first with the WWW door, the WAP door, the mobile phone terminal door combines simultaneously, uses same recommendation distribution platform.The data mining technology of having used RECENT DEVELOPMENTS to get up simultaneously becomes a reality ecommerce intelligent recommendation system.
4, be different from common search engine; Also be different from the Technologies of Recommendation System in E-Commerce that the present modes such as evaluation score information that other users' commodity can only be provided through simple sales list, to the user reach the persuasion customer objective, but the several data digging technology is applied to a plurality of recommendation scenes.
5, first with the user mobile phone model also as a Consideration of commercial product recommending.
6, intelligent distribution platform is placed on the off-line part through data mining model modeling and model evaluation with complicated and time consumption; Recommend calculating section to be placed on online part; Use caching technology that recommendation results is carried out caching process simultaneously; Again they are organically combined, both guaranteed the recommendation quality, the recommendation service of real time high-speed is provided again.
Because data mining model is longer in the time of model training, assessment, and the sample data of using is historic data, so be placed on the off-line part; And be to need instant recommendation service on door, so need online the recommendation to calculate.Be placed on caching process generate good recommendation rules data by data mining model again, just real-time recommendation can be provided.
7. the extensibility of intelligent distribution platform is strong, can also constantly increase to use other data digging methods or statistical analysis technique, at more scene Recommendations.
Technical scheme of the present invention to the user at the WWW door; The WAP door, the data mining technology off-line modeling is used in having landed order (browsing), not having to have landed to browse, land not have to browse and do not have the scene of order of mobile phone terminal door; Calculate in conjunction with online recommendation; For example use that the Memcached caching technology carries out caching process to recommendation results, to the different scenes of different doors distinct interface is provided, provide the Recommendations data to return to each big door and represent at same recommendation dissemination system; Introduce modern data mining data analysis mechanism such as association, cluster, need not user's input, just can understand user preference; According to business scenario and business support needs, can also use other data mining technology and statistical method etc. to arrange new communication interface for each big online door or off-line system.
Claims (10)
1. a commercial product recommending method is characterized in that, comprising:
According to the user browse record and/or user property obtains the sample training data, and said sample training data are set up the association analysis model;
Generate associated data according to the association analysis model;
Receive the commercial product recommending request,, will give relevant door with the commercial product recommending that said user profile is associated based on said user profile and said associated data, so that be shown to the user by relevant door according to said commercial product recommending acquisition request user profile.
2. commercial product recommending method according to claim 1 is characterized in that, generates associated data according to the association analysis model and specifically comprises:
Generate associated data according to the association analysis model and based on support, degree of confidence, lifting degree.
3. commercial product recommending method according to claim 2 is characterized in that, said user profile comprises user terminal type, the commodity of browsing; Browse or during subscription status, will give relevant door with the commercial product recommending that said user profile is associated when the user has logined and has been in, specifically comprise so that be shown to the user by relevant door based on said user profile and said associated data:
To give relevant door with the commercial product recommending of the said commodity corresponding associated of browsing based on said user terminal type, the commodity of browsing and said associated data; So that be shown to the user by relevant door, said door comprises WWW door, WAP door, user terminal door.
4. commercial product recommending method according to claim 3 is characterized in that, if it is not enough to recommend user's commodity amount, then from preset commodity storehouse, chooses the commodity that are suitable for said user terminal type.
5. commercial product recommending method according to claim 2 is characterized in that said user profile comprises the commodity of browsing; Do not browse or during subscription status, will give relevant door with the commercial product recommending that said user profile is associated when the user logins and is in, specifically comprise so that be shown to the user by relevant door based on said user profile and said associated data:
To give relevant door with the commercial product recommending of the said commodity corresponding associated of browsing based on said commodity of browsing and said associated data, so that be shown to the user by relevant door; Said door comprises the WWW door.
6. commercial product recommending method according to claim 1 is characterized in that, browses and when not having subscription status, to generate associated data according to the association analysis model and specifically comprise when the user has logined and has been in not have:
Divide a plurality of customer groups based on said user property, and each customer group is provided with corresponding identification, generate associated data according to the association analysis model.
7. commercial product recommending method according to claim 6 is characterized in that said user profile comprises user terminal type, user property; Browse and when not having subscription status, will give relevant door with the commercial product recommending that said user profile is associated when the user has logined and has been in not have, specifically to comprise so that be shown to the user by relevant door based on said user profile and said associated data:
User property, user terminal type and said associated data based on logged-in user are carried out the customer group classification logotype to the user; Give relevant door with other user preferences among the same subscriber crowd and with the commercial product recommending that said user terminal type is complementary, so that be shown to the user by relevant door; Said door comprises WWW door, WAP door, user terminal door.
8. a commercial product recommending system is characterized in that, comprising:
The modelling unit, be used for according to the user browse record and/or user property obtains the sample training data, and said sample training data are set up the association analysis model;
The commercial product recommending server is connected with said modelling unit communication, comprising:
Generation unit is used for generating associated data according to the association analysis model;
Receiving element is used to receive the commercial product recommending request;
Recommendation unit is used for according to said commercial product recommending acquisition request user profile, will give relevant door with the commercial product recommending that said user profile is associated based on said user profile and said associated data, so that be shown to the user by relevant door.
9. commercial product recommending system according to claim 8 is characterized in that, said user profile comprises user terminal type, the commodity of browsing; When having logined and be in, the user browses or during subscription status; Said recommendation unit specifically is used for: according to said commercial product recommending acquisition request user profile; To give relevant door with the commercial product recommending of the said commodity corresponding associated of browsing based on said user terminal type, the commodity of browsing and said associated data, so that be shown to the user by relevant door.
10. commercial product recommending system according to claim 8 is characterized in that said user profile comprises the commodity of browsing; When logining and be in, the user do not browse or during subscription status; Said recommendation unit specifically is used for: according to said commercial product recommending acquisition request user profile; To give relevant door with the commercial product recommending of the said commodity corresponding associated of browsing based on said commodity of browsing and said associated data, so that be shown to the user by relevant door.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011100206496A CN102592223A (en) | 2011-01-18 | 2011-01-18 | Commodity recommending method and commodity recommending system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011100206496A CN102592223A (en) | 2011-01-18 | 2011-01-18 | Commodity recommending method and commodity recommending system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102592223A true CN102592223A (en) | 2012-07-18 |
Family
ID=46480824
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2011100206496A Pending CN102592223A (en) | 2011-01-18 | 2011-01-18 | Commodity recommending method and commodity recommending system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102592223A (en) |
Cited By (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103345699A (en) * | 2013-07-10 | 2013-10-09 | 湖南大学 | Personalized food recommendation method based on commodity forest system |
| CN103353880A (en) * | 2013-06-20 | 2013-10-16 | 兰州交通大学 | Data mining method adopting dissimilarity degree clustering and association |
| CN103646341A (en) * | 2013-11-29 | 2014-03-19 | 北京奇虎科技有限公司 | A method and an apparatus for recommending website-provided objects |
| 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 |
| CN103716338A (en) * | 2012-09-28 | 2014-04-09 | 腾讯科技（深圳）有限公司 | Information push method and device |
| CN103942217A (en) * | 2013-01-21 | 2014-07-23 | 阿里巴巴集团控股有限公司 | Webpage information recommending method and device |
| CN104216885A (en) * | 2013-05-29 | 2014-12-17 | 酷盛(天津）科技有限公司 | Recommending system and method with static and dynamic recommending reasons automatically combined |
| CN104317945A (en) * | 2014-10-31 | 2015-01-28 | 亚信科技（南京）有限公司 | E-commerce website commodity recommending method on basis of search behaviors |
| CN104636947A (en) * | 2013-11-13 | 2015-05-20 | 同济大学 | Calculation method of confidence degree of recommendation information in Internet of Things |
| CN105183781A (en) * | 2015-08-14 | 2015-12-23 | 百度在线网络技术（北京）有限公司 | Information recommendation method and apparatus |
| CN105224623A (en) * | 2015-09-22 | 2016-01-06 | 北京百度网讯科技有限公司 | The training method of data model and device |
| CN105335386A (en) * | 2014-07-01 | 2016-02-17 | 阿里巴巴集团控股有限公司 | Method and apparatus for providing navigation tag |
| CN105894332A (en) * | 2016-04-22 | 2016-08-24 | 深圳市永兴元科技有限公司 | Commodity recommendation method, device and system based on user behavior analysis |
| CN105989074A (en) * | 2015-02-09 | 2016-10-05 | 北京字节跳动科技有限公司 | Method and device for recommending cold start through mobile equipment information |
| CN106022881A (en) * | 2016-05-20 | 2016-10-12 | 四川省英联国泰科技有限公司 | Information processing method and device |
| CN106095867A (en) * | 2016-06-03 | 2016-11-09 | 北京奇虎科技有限公司 | A kind of book recommendation method based on industry analysis and device |
| CN106157043A (en) * | 2015-03-24 | 2016-11-23 | 联想(北京)有限公司 | The processing method of a kind of recommended and electronic equipment |
| CN106354734A (en) * | 2015-07-17 | 2017-01-25 | 阿里巴巴集团控股有限公司 | Method and device for providing business object information |
| CN106446696A (en) * | 2015-08-10 | 2017-02-22 | 联想(北京)有限公司 | Information processing method and electronic device |
| WO2017028687A1 (en) * | 2015-08-14 | 2017-02-23 | 阿里巴巴集团控股有限公司 | Unused commodity object information processing method and device |
| CN106779926A (en) * | 2016-12-02 | 2017-05-31 | 乐视控股（北京）有限公司 | Correlation rule generation method, device and terminal |
| CN106897293A (en) * | 2015-12-17 | 2017-06-27 | 中国移动通信集团公司 | A kind of data processing method and device |
| CN107123014A (en) * | 2017-03-03 | 2017-09-01 | 德基网络科技南京有限公司 | One kind is based on e-commerce platform personalized recommendation algorithm |
| CN107332910A (en) * | 2017-07-03 | 2017-11-07 | 北京京东尚科信息技术有限公司 | Information-pushing method and device |
| CN107481114A (en) * | 2017-08-16 | 2017-12-15 | 北京京东尚科信息技术有限公司 | Method of Commodity Recommendation, device, e-commerce system and storage medium |
| CN108428189A (en) * | 2018-02-27 | 2018-08-21 | 上海掌门科技有限公司 | A kind of social resource processing method, equipment and readable medium |
| CN108596705A (en) * | 2018-03-23 | 2018-09-28 | 郑州大学西亚斯国际学院 | A kind of commodity suitable for e-commerce recommend method and system with information classification |
| CN111461801A (en) * | 2019-01-18 | 2020-07-28 | 阿里巴巴集团控股有限公司 | Page generation method and device |
Citations (4)
| 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 |
| CN101520878A (en) * | 2009-04-03 | 2009-09-02 | 华为技术有限公司 | Method, device and system for pushing advertisements to users |
| CN101783004A (en) * | 2010-03-03 | 2010-07-21 | 陈嵘 | Fast intelligent commodity recommendation system |
| CN101826114A (en) * | 2010-05-26 | 2010-09-08 | 南京大学 | Multi Markov chain-based content recommendation method |
-
2011
- 2011-01-18 CN CN2011100206496A patent/CN102592223A/en active Pending
Patent Citations (4)
| 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 |
| CN101520878A (en) * | 2009-04-03 | 2009-09-02 | 华为技术有限公司 | Method, device and system for pushing advertisements to users |
| CN101783004A (en) * | 2010-03-03 | 2010-07-21 | 陈嵘 | Fast intelligent commodity recommendation system |
| CN101826114A (en) * | 2010-05-26 | 2010-09-08 | 南京大学 | Multi Markov chain-based content recommendation method |
Cited By (41)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103716338A (en) * | 2012-09-28 | 2014-04-09 | 腾讯科技（深圳）有限公司 | Information push method and device |
| CN103942217B (en) * | 2013-01-21 | 2017-09-01 | 阿里巴巴集团控股有限公司 | The recommendation method and device of a kind of info web |
| CN103942217A (en) * | 2013-01-21 | 2014-07-23 | 阿里巴巴集团控股有限公司 | Webpage information recommending method and device |
| CN104216885A (en) * | 2013-05-29 | 2014-12-17 | 酷盛(天津）科技有限公司 | Recommending system and method with static and dynamic recommending reasons automatically combined |
| CN103353880B (en) * | 2013-06-20 | 2017-03-15 | 兰州交通大学 | A kind of utilization distinctiveness ratio cluster and the data digging method for associating |
| CN103353880A (en) * | 2013-06-20 | 2013-10-16 | 兰州交通大学 | Data mining method adopting dissimilarity degree clustering and association |
| CN103345699A (en) * | 2013-07-10 | 2013-10-09 | 湖南大学 | Personalized food recommendation method based on commodity forest system |
| CN104636947A (en) * | 2013-11-13 | 2015-05-20 | 同济大学 | Calculation method of confidence degree of recommendation information in Internet of Things |
| CN103646341A (en) * | 2013-11-29 | 2014-03-19 | 北京奇虎科技有限公司 | A method and an apparatus for recommending website-provided objects |
| 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 |
| CN103700005B (en) * | 2013-12-17 | 2016-08-31 | 南京信息工程大学 | A kind of correlation rule based on the many minimum supports of self adaptation recommends method |
| CN105335386B (en) * | 2014-07-01 | 2018-10-16 | 阿里巴巴集团控股有限公司 | A kind of method and device that navigation tag is provided |
| CN105335386A (en) * | 2014-07-01 | 2016-02-17 | 阿里巴巴集团控股有限公司 | Method and apparatus for providing navigation tag |
| CN104317945A (en) * | 2014-10-31 | 2015-01-28 | 亚信科技（南京）有限公司 | E-commerce website commodity recommending method on basis of search behaviors |
| CN105989074A (en) * | 2015-02-09 | 2016-10-05 | 北京字节跳动科技有限公司 | Method and device for recommending cold start through mobile equipment information |
| CN106157043A (en) * | 2015-03-24 | 2016-11-23 | 联想(北京)有限公司 | The processing method of a kind of recommended and electronic equipment |
| CN106354734A (en) * | 2015-07-17 | 2017-01-25 | 阿里巴巴集团控股有限公司 | Method and device for providing business object information |
| CN106354734B (en) * | 2015-07-17 | 2019-06-11 | 阿里巴巴集团控股有限公司 | The method and device of business object information is provided |
| WO2017012474A1 (en) * | 2015-07-17 | 2017-01-26 | 阿里巴巴集团控股有限公司 | Method and apparatus for providing service object information |
| CN106446696A (en) * | 2015-08-10 | 2017-02-22 | 联想(北京)有限公司 | Information processing method and electronic device |
| CN106446696B (en) * | 2015-08-10 | 2020-06-23 | 联想(北京)有限公司 | Information processing method and electronic equipment |
| CN105183781B (en) * | 2015-08-14 | 2018-11-20 | 百度在线网络技术（北京）有限公司 | Information recommendation method and device |
| WO2017028687A1 (en) * | 2015-08-14 | 2017-02-23 | 阿里巴巴集团控股有限公司 | Unused commodity object information processing method and device |
| CN105183781A (en) * | 2015-08-14 | 2015-12-23 | 百度在线网络技术（北京）有限公司 | Information recommendation method and apparatus |
| CN105224623A (en) * | 2015-09-22 | 2016-01-06 | 北京百度网讯科技有限公司 | The training method of data model and device |
| CN105224623B (en) * | 2015-09-22 | 2019-06-18 | 北京百度网讯科技有限公司 | The training method and device of data model |
| CN106897293A (en) * | 2015-12-17 | 2017-06-27 | 中国移动通信集团公司 | A kind of data processing method and device |
| CN106897293B (en) * | 2015-12-17 | 2020-09-11 | 中国移动通信集团公司 | Data processing method and device |
| CN105894332A (en) * | 2016-04-22 | 2016-08-24 | 深圳市永兴元科技有限公司 | Commodity recommendation method, device and system based on user behavior analysis |
| CN106022881A (en) * | 2016-05-20 | 2016-10-12 | 四川省英联国泰科技有限公司 | Information processing method and device |
| CN106022881B (en) * | 2016-05-20 | 2020-02-14 | 四川省英联国泰科技有限公司 | Information processing method and device |
| CN106095867A (en) * | 2016-06-03 | 2016-11-09 | 北京奇虎科技有限公司 | A kind of book recommendation method based on industry analysis and device |
| CN106095867B (en) * | 2016-06-03 | 2019-08-23 | 北京奇虎科技有限公司 | A kind of book recommendation method and device based on industry analysis |
| CN106779926A (en) * | 2016-12-02 | 2017-05-31 | 乐视控股（北京）有限公司 | Correlation rule generation method, device and terminal |
| CN107123014A (en) * | 2017-03-03 | 2017-09-01 | 德基网络科技南京有限公司 | One kind is based on e-commerce platform personalized recommendation algorithm |
| CN107332910A (en) * | 2017-07-03 | 2017-11-07 | 北京京东尚科信息技术有限公司 | Information-pushing method and device |
| CN107332910B (en) * | 2017-07-03 | 2020-07-31 | 北京京东尚科信息技术有限公司 | Information pushing method and device |
| CN107481114A (en) * | 2017-08-16 | 2017-12-15 | 北京京东尚科信息技术有限公司 | Method of Commodity Recommendation, device, e-commerce system and storage medium |
| CN108428189A (en) * | 2018-02-27 | 2018-08-21 | 上海掌门科技有限公司 | A kind of social resource processing method, equipment and readable medium |
| CN108596705A (en) * | 2018-03-23 | 2018-09-28 | 郑州大学西亚斯国际学院 | A kind of commodity suitable for e-commerce recommend method and system with information classification |
| CN111461801A (en) * | 2019-01-18 | 2020-07-28 | 阿里巴巴集团控股有限公司 | Page generation method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102592223A (en) | Commodity recommending method and commodity recommending system | |
| Gunawan et al. | Determining Factors in the Use of Digital Marketing and Its Effect on Marketing Performance in the Creative Industries in Tasikmalaya | |
| Hossain et al. | The impact of social media marketing on university students’ brand loyalty | |
| CN103841122B (en) | Target object information recommends method, server and client | |
| EP3561757A1 (en) | Method and system for engaging real-time-human interaction into media presented online | |
| TWI668641B (en) | Object selection method and device | |
| Dickey et al. | An overview of digital media and advertising | |
| Beheshtinia et al. | A multi-objective and integrated model for supply chain scheduling optimization in a multi-site manufacturing system | |
| US20140324571A1 (en) | System and method for selecting and rendering content | |
| CN104636473A (en) | Data processing method and system based on electronic payment behaviors | |
| CN106022816A (en) | Rebate method | |
| CN104091270A (en) | Method for commodity comparison in code scanning mode | |
| Van Couvering | The political economy of new media revisited | |
| Gligorijevic et al. | Engaging social customers–Influencing new marketing strategies for social media information sources | |
| Aris et al. | Crowdsourcing evolution: Towards a taxonomy of crowdsourcing initiatives | |
| CN104484406A (en) | Recommended data issuing analysis method and recommended data issuing analysis device | |
| Fanta et al. | Social media as effective tool for understanding customer experience: a systematized review | |
| CN109299368B (en) | Method and system for intelligent and personalized recommendation of environmental information resources AI | |
| Tan et al. | Dynamic model and simulation of open innovation in product development | |
| CN103164407A (en) | Information searching method and system | |
| Alam | An investigation on the use of digital marketing towards the customer satisfaction and brand loyalty of resturants in Saudi Arabia | |
| Kumah | The role of social media as a platform for E-commerce | |
| CN113034231B (en) | Multi-supply chain commodity intelligent recommendation system and method based on SaaS cloud service | |
| CN110347923B (en) | Traceable fast fission type user portrait construction method | |
| Wang et al. | Systematic approach for digital marketing strategy through data mining technology |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20120718 |