CN103345698A - Personalized recommendation method based on cloud processing mode and applied in e-business environment - Google Patents
Personalized recommendation method based on cloud processing mode and applied in e-business environment Download PDFInfo
- Publication number
- CN103345698A CN103345698A CN2013102875554A CN201310287555A CN103345698A CN 103345698 A CN103345698 A CN 103345698A CN 2013102875554 A CN2013102875554 A CN 2013102875554A CN 201310287555 A CN201310287555 A CN 201310287555A CN 103345698 A CN103345698 A CN 103345698A
- Authority
- CN
- China
- Prior art keywords
- product
- recommendation
- user
- information
- page
- 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
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a personalized recommendation method based on the cloud processing mode and applied in the e-business environment. The personalized recommendation method mainly solves the problem that an existing personalized recommendation method is low in recommendation efficiency and poor in recommendation precision when processing mass data. The personalized recommendation method is divided into an off-line portion and an on-line portion. According to the off-line portion, the Hadoop frame of the cloud computing technology is used for parallel processing of historical data information, an HDFS is used for storing mass data information, and four kinds of parallelized recommendation methods which are suitable for different business stages of e-business are achieved according to the MapReduce programming model. According to the on-line portion, a lightweight data base is arranged and used for storing a user behavior log, a dynamic data collection mechanism is designed and used for reading data which are processed and obtained by the off-line portion in real time, web display and information statistics service are provided, and real-time recommendation information is provided for a user. The personalized recommendation method based on the cloud processing mode and applied in the e-business environment has the remarkable advantages of processing the mass data generated by e-business application.
Description
Technical field
The present invention relates to computer network and data mining field, specifically realize a kind of personalized recommendation method based on the cloud computing tupe towards e-commerce environment, mass data characteristics and concrete service logic at present E-business applications, utilize Hadoop framework stored record user behavior and product information, and online use Web daily record dynamically recording user current behavior cooperation off-line recommendation result of calculation, thereby provide the service of personalized recommendation rapidly and efficiently.
Background technology
It is in the commerce and trade activity widely of all parts of the world that ecommerce typically refers to, under the open network environment in the Internet, based on the browser/server application mode, both parties do not carry out various commercial activities with meeting, realize a kind of novel commercial operation pattern of consumer's shopping online, the online transaction between the trade company and online E-Payment and various commercial activity, transaction, finance activities and relevant integrated service activity.
Since ecommerce towards the big and inquiry real-time of huge, the product data amount of customer group have relatively high expectations, how can understand user's request rapidly becomes the e-commerce development problem demanding prompt solution, so the personalized recommendation technology is arisen at the historic moment.Personalized recommendation is a kind of subjective interest according to the user and objective usage behavior, initiatively recommends the information filtering technology of the interested product of its possibility to the user.The personalized recommendation technology can effectively solve the product information overload problem that exists in the ecommerce as a kind of important information filtering means.
At present, the personalized recommendation technology has been widely used in all kinds of E-business applications, although obtained certain achievement in research, it still faces a lot of challenges, mainly comprises big data processing, the sparse property of data and cold start-up problem.Big data processing problem refers to that present number of users, product information and purchase information are how much levels and increase, and has reached TB even PB level.Huge data set like this is carried out the personalized recommendation analysis, need take a large amount of calculating and storage resources, if still adopt centralized analyzing and processing pattern, then can cause recommending overlong time, can't requirement of real time, the shopping that has greatly influenced the user is experienced; The sparse property of data problem refers to that the while is very rare by the product that a plurality of users mark under extensive e-commerce environment, thereby influences the excavation precision of similar users; The cold start-up problem refers to new user owing to not having the very difficult neighbours of calculating of purchaser record and new product because the less very difficult acquisition recommendation of scoring.At the problems referred to above, existing collaborative filtering recommending method has been difficult to be applicable to present E-business applications, addresses the above problem so need a kind of brand-new personalized recommendation method badly, realizes recommending fast and efficiently target.
Cloud computing is a kind of new distribution type computation schema that information industry circle proposes, and it has significant advantage aspect processing mass data.Along with appearance and the development of cloud computing in recent years, utilize cloud computing environment to realize becoming the effective way that overcomes the above problems towards the efficient personalized recommendation of ecommerce.The core concept of cloud computing is that the resource with network connection is in a large number carried out unified management, makes up the shared resource pond by Intel Virtualization Technology, and provides corresponding resource in the mode of payment, resilient expansion as required, reduces O﹠M cost when improving service quality.The distributed storage technology of cloud computing and parallel processing framework technology can effectively remedy the various deficiencies that exist in the existing commending system, thereby greatly improve the efficient of commending system.
Summary of the invention
Technical matters: the present invention has realized the personalized recommendation efficiently towards ecommerce.On the one hand, invention is divided, places, is stored and inquire about the magnanimity recommending data collection that user behavior and product information constitute, quickly and efficiently deal with data based on the distributed storage technology in the cloud computing.On the other hand, invention utilizes Web daily record dynamically recording user current behavior, thereby recommends to provide reliable foundation for demand that can real-time ground reaction user.
Technical scheme: the personalized recommendation method based on the cloud computing tupe under a kind of e-commerce environment of the present invention may further comprise the steps:
Step 1): raw data set is carried out pre-service, connect the relevant information that user and product are integrated in inquiry by multilist, and deposit the pre-service result in HDFS (Hadoop distributed file system, Hadoop is a distributed system architecture);
Step 2): in the off-line recommendation process, different phase at the e-commerce transaction flow process, at the different pages, realization is applicable to the different parallelization recommend method of each Service Period of ecommerce according to MapReduce (a kind of parallel processing framework that Google proposes) programming model, satisfy the user in the different demands of each transactional stage to product, recommend method calculates the gained result and deposits HDFS in;
Step 3): the data acquisition algorithm is set regularly reads among the HDFS corresponding data and deposit the lightweight database in, produce recommendation results in real time in conjunction with the behavior daily record of login user and feed back to the user, provide recommendation information statistics and graphical the displaying to serve simultaneously, allow the user be easier to understand the production process of recommendation.
In the described step 1), institute carries out pretreated data to raw data set and comprises production code member, product seller, product key word, product buyer, products catalogue, product business procedure, production area, product hot topic degree, these six together information integrated, is used for the content recommendation of the counting yield page and searched page; Customs Assigned Number, production code member, the visit moment, the inquiry moment, these four together information integrated, is used for calculating the content recommendation of User Page and the product page; Customs Assigned Number, the set of user's inquiry product, products catalogue combines, and the user calculates the content recommendation of the inquiry page; HDFS is the distributed file system under the Hadoop framework, is used for the storage mass data.
Described step 2) in the off-line recommendation process, the different pages at e-commerce website have adopted different recommend methods; The recommendation mechanisms that is based on content in the product page and searched page employing; In the employing of user's login page is the recommendation mechanisms of collaborative filtering; The recommendation that is based on correlation rule that the inquiry page uses, specific as follows:
3.1 the every attribute information of product of the recommend method utilization statistics that the product page is content-based, give weights to these information, and the attribute information of different product is compared, obtain the similarity between the product, in recommendation results, to recommending with the high product of target product similarity;
3.2 products catalogue and the product hot topic degree information of the recommend method utilization statistics that searched page is content-based, occurrence number to its same directory of product accounting in the Search Results, the high catalogue of record occurrence number, and the product that popular degree is high under these catalogues is put into recommendation results recommend;
3.3 User Page based on the user of the recommend method utilization of collaborative filtering statistics to the visit of product, the inquiry information calculations user interest-degree to product, and obtain the similarity height by interest-degree, namely to same or same series products users interest, and recommendation and the high user's interest product of targeted customer's similarity;
3.4 the inquiry page is based on the products catalogue of the recommend method utilization statistics of correlation rule and the inherent correlativity that the inquiry information excavating goes out product, and formation directory rules collection, during recommendation the product that popular degree is high in other catalogues under the rule set is recommended;
The MapReduce distributed algorithm is realized under all recommend methods use Hadoop frameworks, to reach purpose efficient, the fast processing mass data.
The user behavior daily record is used for record active user's behavior in the described step 3), and these behaviors are carried out statistical study, in recommendation results, merge the key element of these behaviors, thus a series products of recommending the user to be most interested at present, to reach the purpose of real-time recommendation; Adopt lightweight database storage recommendation results, adopt lightweight database storage recommendation results; Based on the application framework Spring Framework of enterprise-level task scheduling framework Quartz Framework, global function stack, a kind of Java persistence framework MyBatis development data acquisition module and Web display module, use end frame before the BootStrap (a kind of front end tool bag of the quick exploitation web application that is provided by Twitter company) structural interface close friend, the good user interface of interactivity; Utilize the 5th edition HTML5 of HTML (Hypertext Markup Language) and a kind of draw library Chart.js to finish statistics histogram, flowcharting.
Beneficial effect: the present invention compares with existing personalized recommendation method, has the following advantages:
1. at the different phase of e-commerce transaction flow process, the personalized recommendation method that design meets each stage characteristic satisfies the user in the different demands of each transactional stage to product, and recommend method has more specific aim, and the user recommends to experience and is significantly improved;
2. adopt off-line to recommend the mode that combines with online recommendation, when guaranteeing to recommend precision, the corresponding time of recommendation of greatly having reduced client;
3. off-line recommends part to use the distributed storage technology of cloud computing and parallel processing framework technology to handle the magnanimity score data, has effectively improved recommendation efficient;
4. online recommendation part is used lightweight database and stable Web display technique, and the online real-time recommendation information preferably that provides reduces the recommendation response time.
Description of drawings
Fig. 1 is Figure of abstract, the personalized recommendation system process flow diagram,
Fig. 2 is recommendation code overall design Organization Chart,
Fig. 3 is product introduction page recommended flowsheet figure,
Fig. 4 is login page recommended flowsheet figure,
Fig. 5 is inquiry page recommended flowsheet figure.
Embodiment
Based on the personalized recommendation method of cloud computing tupe, comprise four modules under the e-commerce environment, the different pages of the corresponding e-commerce website of each module form different content recommendations.
First module is the Products Show module, may further comprise the steps:
A. extract the needed product information of Products Show module in the raw data, comprise production code member, product seller, product key word, product buyer, products catalogue, product business procedure, production area, product hot topic degree, and be incorporated in the table;
B. be the key value with the production code member, all the other 6 kinds of information are added up as Map as the value value respectively, then record product numbering in twos of identical entry is arranged, be key with production code member in twos, the identical entry number is that value deposits in the different table respectively as Reduce output result, and different tables is integrated into a table again;
C. with 6 the value values weighted sum in this table, carry out record as the similarity between the product;
D. the similar kilsyth basalt of the product of gained is inserted in the MySQL database;
E. get with the highest preceding n item product of target product similarity as the off-line recommendation results, cooperate online user journal, choose the synthetic content recommendation of the present interested product of user and feed back to the user.
Second module is the search recommending module, may further comprise the steps:
A. extract raw data product hot topic degree and products catalogue relevant information, be incorporated in the table by production code member, directly deposit the MySQL database in;
B. add up product associative directory information in the Search Results, and record the catalogue that repeats;
C. repeat other products under the catalogue in the look-up table, the product that popular degree is high feeds back to the user as recommendation results, and catalogue multiplicity and popular degree are in conjunction with considering to determine recommendation order here.
The 3rd module is user's recommending module, may further comprise the steps:
A. extract the needed information of user's recommending module in the raw data, comprise Customs Assigned Number, production code member, inquiry constantly, visit constantly;
B. inquiry is calculated to be specific inquiry and visit numerical value constantly with visit constantly, utilizes formula:
val(t)＝val(0)/λt(1)
Wherein val (0) represents inquiry and the visit numerical value of current time, and λ is attenuation coefficient, and t is current apart from inquiry or visit difference constantly;
C. with in Customs Assigned Number, production code member, inquiry numerical value and visit numerical statistic to a table;
D. calculate the user for the first time to the interest-degree of product according to table information;
E. the similarity between the product of asking before the arrangement is calculated the user to the interest-degree of product for the second time by primary interest-degree and product similarity, remedies the too sparse shortcoming of data;
F. according to the user interest-degree of product is asked dynamic similarity degree between the user, formula is as follows:
G. utilize the user's static information in the raw data to ask for the static similarity of user;
H. integrate user's dynamic similarity degree and static similarity, obtain final user's similarity, formula is as follows:
sim＝σsim
static+(1-σ)sim
dynamic(3)
I. after obtaining similar users, to the interest-degree target of prediction user of the product score value to product, select the highest preceding N item product of score value as the off-line recommendation results according to similar users;
J. utilize online user's daily record, record active user's current operation, understand present user's interest, feed back to the user in conjunction with the off-line recommendation results.
The 4th module is the inquiry recommending module, and concrete steps are as follows:
A. raw data is carried out pre-service, extract Customs Assigned Number, production code member, information such as the inquiry moment, and be incorporated in the table, it is handled, according to product classification information, form the affairs collection;
B. according to before popular degree information in the popular kilsyth basalt of the product asked, add up the hot product in every kind of classification;
C. utilize the Apriori algorithm to carry out Mining Association Rules, the formation rule collection, and deposit in the MySQL database;
D. read the product in user's shopping basket, and extract product classification information, product classification information is preposition as rule, rule searching collective data table in the database, it is rearmounted to obtain rule, i.e. Xiang Guan product classification.Extract the hot product in the relevant classification at last, generate recommendation list, in the page, show, finish recommendation process.
The present invention utilizes cloud computing environment to finish processing to mass data in the recommendation process according to the service logic of ecommerce, can generate recommendation results more fast and efficiently, problems such as, cold start-up sparse to data have simultaneously all been carried out effective processing, for original commending system that moves under the centralized processing pattern provides new thinking, simultaneously in conjunction with the Web daily record to user behavior carry out in real time, record dynamically, guaranteed the real-time of recommendation results.Therefore, the present invention will be the strength that further develops contribution oneself of personalized recommendation.
The present invention is further described in more detail below in conjunction with the drawings and the specific embodiments.
Code Design framework of the present invention as shown in Figure 2.What the framework here mainly represented is the design of off-line partial code, and the information that online part only needs to add the Web log statistic on off-line part basis gets final product.Code definition two classes.
First is Product, the various attributes that wherein comprise product, defined the whole bag of tricks in the class then, be used for asking the product similarity of every attribute between any two, similarity result deposits in the defined product similarity matrix of ProductSimStore class after integrating, and obtains net result according to similarity matrix at last.The product similarity matrix here with the document form storage, then can be used for the input and output of MapReduce in actual conditions.
Second is the Fond class, the needed various attributes of user's recommending module have wherein been comprised, and according to the definition method cal_f () obtain the user to the interest-degree of product, obtain dynamic similarity degree between the user according to interest-degree, obtain the static similarity of user by static similarity class UserStaticInfo again, dynamically be combined with static similarity and obtain final user's similarity, deposit user's similarity matrix in.Obtain the recommendation results of User Page according to similarity.
Searched page has defined the SearchRec class, it directly obtains product similarity information by the product similarity matrix, more catalogue number wherein appears in directory information and the statistics of using two method statIndex () and defIndex () to obtain Search Results then, and last method recSearchResult () recommends the hot product under these catalogues.
The class that comprises two keys in the inquiry page, one is AssociationRulesMining (), it has used the Apriori algorithm that rule is excavated, and the rule set that generates is deposited in the database.Another is Classification, and it has added up the hot product under the different classification, after the rule searching collection is determined catalog classification, forms recommendation results according to these hot products.
Searched page recommends whole thinking comparatively simple, and the recommended flowsheet step is as follows:
1. the numbering of every product number in the acquisition Search Results;
2. obtain product place catalogue according to production code member, and these catalogues are added up, the more catalogue of record occurrence number;
3. other hot products under the catalogue that occurrence number is more feed back as recommendation results.
The Products Show flow process as shown in Figure 3, it has defined the recommended flowsheet of product under the off-line case.Step is as follows:
1. definition product class Product, and be each attribute assignment wherein according to raw data;
2. each attribute of product is asked its similarity in twos, utilize the attribute similarity of MapReduce to integrate then, obtain the product similarity at last;
3. the product similarity is deposited in the two-dimensional matrix and (under the MapReduce algorithm, replace with document form);
4. to each product, choose preceding k the most similar to it in similarity matrix product and export as recommendation results.
The login page recommended flowsheet as shown in Figure 4, it has defined the recommended flowsheet of user's recommending module under the off-line case.Step is as follows:
In raw data according to the company information table, product information table, the visit information table, the inquiry information table obtains the user, product related information;
Try to achieve the user to the interest-degree of product according to these information, utilize the like product attribute to alleviate the influence that sparse property problem produces recommendation results here
By the user interest-degree of product is tried to achieve user's dynamic similarity degree, and in conjunction with the static similarity information of user, utilize formula to obtain the total similarity of user, and deposit in the similarity matrix;
Find neighbours user according to user's similarity matrix, according to the interest-degree information of user to product, obtain last recommendation results again.
Inquiry page recommended flowsheet as shown in Figure 5, concrete steps are as follows:
1. historical data is carried out pre-service, form the affairs collection;
2. carry out association rule mining according to affairs transporting something containerized with the Apriori algorithm, find the potential incidence relation between each product;
3. after the user logins, according to the content in the inquiry basket, search rule, the product classification that obtains recommending;
4. according to preceding n the most popular product under this classification of classification searching, be presented in the page and recommend.
Should be pointed out that for those skilled in the art under the prerequisite that does not break away from the principle of the invention, can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.In the present embodiment not clear and definite each several part all available prior art realized.
Claims (4)
- Under the e-commerce environment based on the personalized recommendation method of cloud computing tupe, it is characterized in that: this method may further comprise the steps:Step 1): raw data set is carried out pre-service, connect the relevant information that user and product are integrated in inquiry by multilist, and deposit the pre-service result in HDFS;Step 2): in the off-line recommendation process, different phase at the e-commerce transaction flow process, at the different pages, realization is applicable to the different parallelization recommend method of each Service Period of ecommerce according to the MapReduce programming model, satisfy the user in the different demands of each transactional stage to product, recommend method calculates the gained result and deposits HDFS in;Step 3): the data acquisition algorithm is set regularly reads among the HDFS corresponding data and deposit the lightweight database in, produce recommendation results in real time in conjunction with the behavior daily record of login user and feed back to the user, provide recommendation information statistics and graphical the displaying to serve simultaneously, allow the user be easier to understand the production process of recommendation.
- Under the e-commerce environment according to claim 1 based on the personalized recommendation method of cloud computing tupe, it is characterized in that: in the described step 1), institute carries out pretreated data to raw data set and comprises production code member, product seller, product key word, product buyer, products catalogue, product business procedure, production area, product hot topic degree, these six together information integrated, is used for the content recommendation of the counting yield page and searched page; Customs Assigned Number, production code member, the visit moment, the inquiry moment, these four together information integrated, is used for calculating the content recommendation of User Page and the product page; Customs Assigned Number, the set of user's inquiry product, products catalogue combines, and the user calculates the content recommendation of the inquiry page; HDFS is the distributed file system under the Hadoop framework, is used for the storage mass data.
- Under the e-commerce environment according to claim 1 based on the personalized recommendation method of cloud computing tupe, it is characterized in that: in the off-line recommendation process described step 2), the different pages at e-commerce website have adopted different recommend methods; The recommendation mechanisms that is based on content in the product page and searched page employing; In the employing of user's login page is the recommendation mechanisms of collaborative filtering; The recommendation that is based on correlation rule that the inquiry page uses, specific as follows:3.1 the every attribute information of product of the recommend method utilization statistics that the product page is content-based, give weights to these information, and the attribute information of different product is compared, obtain the similarity between the product, in recommendation results, to recommending with the high product of target product similarity;3.2 products catalogue and the product hot topic degree information of the recommend method utilization statistics that searched page is content-based, occurrence number to its same directory of product accounting in the Search Results, the high catalogue of record occurrence number, and the product that popular degree is high under these catalogues is put into recommendation results recommend;3.3 User Page based on the user of the recommend method utilization of collaborative filtering statistics to the visit of product, the inquiry information calculations user interest-degree to product, and obtain the similarity height by interest-degree, namely to same or same series products users interest, and recommendation and the high user's interest product of targeted customer's similarity;3.4 the inquiry page is based on the products catalogue of the recommend method utilization statistics of correlation rule and the inherent correlativity that the inquiry information excavating goes out product, and formation directory rules collection, during recommendation the product that popular degree is high in other catalogues under the rule set is recommended;The MapReduce distributed algorithm is realized under all recommend methods use Hadoop frameworks, to reach purpose efficient, the fast processing mass data.
- Under the e-commerce environment according to claim 1 based on the personalized recommendation method of cloud computing tupe, it is characterized in that: the user behavior daily record is used for record active user's behavior in the described step 3), and these behaviors are carried out statistical study, in recommendation results, merge the key element of these behaviors, thereby a series products of recommending the user to be most interested at present is to reach the purpose of real-time recommendation; Adopt lightweight database storage recommendation results, adopt lightweight database storage recommendation results; Based on the application framework Spring Framework of enterprise-level task scheduling framework Quartz Framework, global function stack, a kind of Java persistence framework MyBatis development data acquisition module and Web display module, use BootStrap front end belfry friendly interface, the good user interface of interactivity; Utilize the 5th edition HTML5 of HTML (Hypertext Markup Language) and a kind of draw library Chart.js to finish statistics histogram, flowcharting.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013102875554A CN103345698A (en) | 2013-07-09 | 2013-07-09 | Personalized recommendation method based on cloud processing mode and applied in e-business environment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013102875554A CN103345698A (en) | 2013-07-09 | 2013-07-09 | Personalized recommendation method based on cloud processing mode and applied in e-business environment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103345698A true CN103345698A (en) | 2013-10-09 |
Family
ID=49280491
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2013102875554A Pending CN103345698A (en) | 2013-07-09 | 2013-07-09 | Personalized recommendation method based on cloud processing mode and applied in e-business environment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103345698A (en) |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103729478A (en) * | 2014-01-26 | 2014-04-16 | 重庆邮电大学 | LBS (Location Based Service) interest point discovery method based on MapReduce |
| CN103886487A (en) * | 2014-03-28 | 2014-06-25 | 焦点科技股份有限公司 | Individualized recommendation method and system based on distributed B2B platform |
| CN104111996A (en) * | 2014-07-07 | 2014-10-22 | 山大地纬软件股份有限公司 | Health insurance outpatient clinic big data extraction system and method based on hadoop platform |
| CN104462811A (en) * | 2014-12-05 | 2015-03-25 | 云中万维(北京)科技有限公司 | Network game data processing method |
| CN104503967A (en) * | 2014-10-24 | 2015-04-08 | 浪潮电子信息产业股份有限公司 | Hadoop-based network recommendation method |
| CN105243063A (en) * | 2014-06-18 | 2016-01-13 | 北京新媒传信科技有限公司 | Information recommendation method and device |
| WO2016183898A1 (en) * | 2015-05-18 | 2016-11-24 | 向莉妮 | Personalized commodity matching and recommendation system and method, and electronic device |
| CN106296305A (en) * | 2016-08-23 | 2017-01-04 | 上海海事大学 | Electric business website real-time recommendation System and method under big data environment |
| CN103914528B (en) * | 2014-03-28 | 2017-02-15 | 南京邮电大学 | Parallelizing method of association analytical algorithm |
| CN106844550A (en) * | 2016-12-30 | 2017-06-13 | 郑州云海信息技术有限公司 | Method and device is recommended in a kind of virtual platform operation |
| CN106850750A (en) * | 2016-12-26 | 2017-06-13 | 北京五八信息技术有限公司 | A kind of method and apparatus of real time propelling movement information |
| CN106940801A (en) * | 2016-01-04 | 2017-07-11 | 中国科学院声学研究所 | A kind of deeply for Wide Area Network learns commending system and method |
| CN106971344A (en) * | 2016-01-14 | 2017-07-21 | 平安科技（深圳）有限公司 | Insured amount control method and system |
| CN107256512A (en) * | 2017-06-08 | 2017-10-17 | 贵州优联博睿科技有限公司 | One kind house-purchase personalized recommendation method and system |
| CN107730336A (en) * | 2016-08-12 | 2018-02-23 | 苏宁云商集团股份有限公司 | Commodity method for pushing and device in a kind of online transaction |
| CN109034869A (en) * | 2018-06-26 | 2018-12-18 | 杭州排列科技有限公司 | Real-time recommendation system and method based on similar audient |
| CN109087123A (en) * | 2018-06-29 | 2018-12-25 | 安客诚全球信息服务（南通）有限公司 | A kind of accurate screening technique of audient based on artificial intelligence |
| CN109255586A (en) * | 2018-08-24 | 2019-01-22 | 安徽讯飞智能科技有限公司 | A kind of online personalized recommendation method that E-Governance Oriented is handled affairs |
| CN110808833A (en) * | 2019-11-12 | 2020-02-18 | 电子科技大学 | Lightweight online and offline certificateless signature method |
| CN111310042A (en) * | 2020-02-13 | 2020-06-19 | 研祥智能科技股份有限公司 | Collaborative filtering recommendation method and system based on cloud computing |
| CN114168862A (en) * | 2021-12-14 | 2022-03-11 | 平安国际智慧城市科技股份有限公司 | Product recommendation method, device, equipment and storage medium based on collaborative filtering |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102169505A (en) * | 2011-05-16 | 2011-08-31 | 苏州两江科技有限公司 | Recommendation system building method based on cloud computing |
| CN102750647A (en) * | 2012-06-29 | 2012-10-24 | 南京大学 | Merchant recommendation method based on transaction network |
-
2013
- 2013-07-09 CN CN2013102875554A patent/CN103345698A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102169505A (en) * | 2011-05-16 | 2011-08-31 | 苏州两江科技有限公司 | Recommendation system building method based on cloud computing |
| CN102750647A (en) * | 2012-06-29 | 2012-10-24 | 南京大学 | Merchant recommendation method based on transaction network |
Non-Patent Citations (2)
| Title |
|---|
| 张永霞: "一种基于Hadoop 的个性化推荐系统架构", 《中国科技论文在线》 * |
| 汪彦红: "基于组合算法的个性化推荐系统的研究和应用", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103729478A (en) * | 2014-01-26 | 2014-04-16 | 重庆邮电大学 | LBS (Location Based Service) interest point discovery method based on MapReduce |
| CN103729478B (en) * | 2014-01-26 | 2016-11-09 | 重庆邮电大学 | LBS interest point discovery method based on MapReduce |
| CN103914528B (en) * | 2014-03-28 | 2017-02-15 | 南京邮电大学 | Parallelizing method of association analytical algorithm |
| CN103886487A (en) * | 2014-03-28 | 2014-06-25 | 焦点科技股份有限公司 | Individualized recommendation method and system based on distributed B2B platform |
| CN103886487B (en) * | 2014-03-28 | 2016-01-27 | 焦点科技股份有限公司 | Based on personalized recommendation method and the system of distributed B2B platform |
| CN105243063A (en) * | 2014-06-18 | 2016-01-13 | 北京新媒传信科技有限公司 | Information recommendation method and device |
| CN105243063B (en) * | 2014-06-18 | 2019-11-15 | 北京新媒传信科技有限公司 | The method and apparatus of information recommendation |
| CN104111996A (en) * | 2014-07-07 | 2014-10-22 | 山大地纬软件股份有限公司 | Health insurance outpatient clinic big data extraction system and method based on hadoop platform |
| CN104503967A (en) * | 2014-10-24 | 2015-04-08 | 浪潮电子信息产业股份有限公司 | Hadoop-based network recommendation method |
| CN104462811A (en) * | 2014-12-05 | 2015-03-25 | 云中万维(北京)科技有限公司 | Network game data processing method |
| WO2016183898A1 (en) * | 2015-05-18 | 2016-11-24 | 向莉妮 | Personalized commodity matching and recommendation system and method, and electronic device |
| CN106940801A (en) * | 2016-01-04 | 2017-07-11 | 中国科学院声学研究所 | A kind of deeply for Wide Area Network learns commending system and method |
| CN106971344A (en) * | 2016-01-14 | 2017-07-21 | 平安科技（深圳）有限公司 | Insured amount control method and system |
| CN107730336A (en) * | 2016-08-12 | 2018-02-23 | 苏宁云商集团股份有限公司 | Commodity method for pushing and device in a kind of online transaction |
| CN106296305A (en) * | 2016-08-23 | 2017-01-04 | 上海海事大学 | Electric business website real-time recommendation System and method under big data environment |
| CN106850750A (en) * | 2016-12-26 | 2017-06-13 | 北京五八信息技术有限公司 | A kind of method and apparatus of real time propelling movement information |
| CN106844550A (en) * | 2016-12-30 | 2017-06-13 | 郑州云海信息技术有限公司 | Method and device is recommended in a kind of virtual platform operation |
| CN107256512A (en) * | 2017-06-08 | 2017-10-17 | 贵州优联博睿科技有限公司 | One kind house-purchase personalized recommendation method and system |
| CN109034869A (en) * | 2018-06-26 | 2018-12-18 | 杭州排列科技有限公司 | Real-time recommendation system and method based on similar audient |
| CN109087123A (en) * | 2018-06-29 | 2018-12-25 | 安客诚全球信息服务（南通）有限公司 | A kind of accurate screening technique of audient based on artificial intelligence |
| CN109255586A (en) * | 2018-08-24 | 2019-01-22 | 安徽讯飞智能科技有限公司 | A kind of online personalized recommendation method that E-Governance Oriented is handled affairs |
| CN109255586B (en) * | 2018-08-24 | 2022-03-29 | 安徽讯飞智能科技有限公司 | Online personalized recommendation method for e-government affairs handling |
| CN110808833A (en) * | 2019-11-12 | 2020-02-18 | 电子科技大学 | Lightweight online and offline certificateless signature method |
| CN110808833B (en) * | 2019-11-12 | 2021-08-06 | 电子科技大学 | Lightweight online and offline certificateless signature method |
| CN111310042A (en) * | 2020-02-13 | 2020-06-19 | 研祥智能科技股份有限公司 | Collaborative filtering recommendation method and system based on cloud computing |
| CN114168862A (en) * | 2021-12-14 | 2022-03-11 | 平安国际智慧城市科技股份有限公司 | Product recommendation method, device, equipment and storage medium based on collaborative filtering |
| CN114168862B (en) * | 2021-12-14 | 2024-07-19 | 平安国际智慧城市科技股份有限公司 | Collaborative filtering-based product recommendation method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103345698A (en) | Personalized recommendation method based on cloud processing mode and applied in e-business environment | |
| Ren et al. | Demand forecasting in retail operations for fashionable products: methods, practices, and real case study | |
| Niu et al. | Organizational business intelligence and decision making using big data analytics | |
| Zhu et al. | Online purchase decisions for tourism e-commerce | |
| CN103886487B (en) | Based on personalized recommendation method and the system of distributed B2B platform | |
| Lu et al. | BizSeeker: a hybrid semantic recommendation system for personalized government‐to‐business e‐services | |
| CN105488216A (en) | Recommendation system and method based on implicit feedback collaborative filtering algorithm | |
| CN108416625A (en) | The recommendation method and apparatus of marketing product | |
| Verma et al. | An intelligent approach to Big Data analytics for sustainable retail environment using Apriori-MapReduce framework | |
| CN108122153A (en) | Personalized recommendation method based on cloud computing tupe under e-commerce environment | |
| CN103914468A (en) | Method and device for searching for released information | |
| CN102789462A (en) | Project recommendation method and system | |
| Akpan et al. | Foreign direct investment and industrial sector performance: Assessing the long-run implication on economic growth in Nigeria | |
| Stahl et al. | Marketplaces for digital data: Quo vadis? | |
| Chen et al. | Behavior-aware user response modeling in social media: Learning from diverse heterogeneous data | |
| CN104899229A (en) | Swarm intelligence based behavior clustering system | |
| Zhao | A Study on the Applications of Big Data in Cross-border E-commerce | |
| Xu et al. | Novel model of e-commerce marketing based on big data analysis and processing | |
| Devlin et al. | Big data comes of age | |
| Saha | Analytics and Big Data: Emerging trends and their impact on our lives | |
| Luo et al. | Exploring the dynamic influence of visit behavior on online store sales performance: An empirical investigation | |
| Saxena et al. | Business intelligence | |
| Sheriff | Big data revolution: Is it a business disruption? | |
| CN104715430A (en) | Data collecting engine adopting novel program slicing technology and orienting commercial real estate industry | |
| Sahu et al. | Personalized recommendation engine using HADOOP |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20131009 |