CN108596720A - A method of commercial product recommending is carried out according to the behavioral data of user - Google Patents
A method of commercial product recommending is carried out according to the behavioral data of user Download PDFInfo
- Publication number
- CN108596720A CN108596720A CN201810367578.9A CN201810367578A CN108596720A CN 108596720 A CN108596720 A CN 108596720A CN 201810367578 A CN201810367578 A CN 201810367578A CN 108596720 A CN108596720 A CN 108596720A
- Authority
- CN
- China
- Prior art keywords
- user
- commodity
- data
- recommending
- commercial product
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
Abstract
The present invention discloses a kind of method carrying out commercial product recommending according to the behavioral data of user, includes following steps：(1) it is drawn a portrait to user by the behavior record of user；(2) scoring of the user to commodity is generated according to the relationship of user and commodity, rating matrix of the user to commodity is then generated according to collaborative filtering；(3) user's portrait and user is combined to generate recommending data to the rating matrix of commodity.By using the method for the present invention so that commercial product recommending is more fine, can really reflect the commodity that user really actually wants to, more systematicization, it is that user really wants that the commodity of actual recommendation, which are more bonded, is all brought conveniently for businessman and user.
Description
Technical field
The present invention relates to e-commerce field technologies, refer in particular to a kind of according to the behavioral data of user progress commercial product recommending
Method.
Background technology
With the surge of the development and Internet user's number of Internet technology, network information exponentially increases, information mistake
Load problem getting worse.In recent years, as solve problem of information overload major technique --- commending system has obtained widely
Development and application.
Currently, existing store commercial product recommending is not fine enough, it is mainly based upon number of clicks and carries out commercial product recommending, user couple
A certain commodity click checks that often, when user is again turned on webpage, the display order of the commodity is with regard to forward, but such recommendation
The dimension that method analysis is considered is few, cannot really reflect the commodity that user really actually wants to, inadequate system, the quotient of actual recommendation
Product may not be that user is really desired.
Invention content
In view of this, in view of the deficiencies of the prior art, the present invention aims to provide a kind of according to user's
The method that behavioral data carries out commercial product recommending, can effectively solve the problems, such as that existing Method of Commodity Recommendation is not fine enough.
To achieve the above object, the present invention is using following technical solution：
A method of commercial product recommending is carried out according to the behavioral data of user, includes following steps：
(1) it is drawn a portrait to user by the behavior record of user；
(2) scoring of the user to commodity is generated according to the relationship of user and commodity, then generates one according to collaborative filtering
Rating matrix of the user to commodity；
(3) user's portrait and user is combined to generate recommending data to the rating matrix of commodity.
Preferably, the behavior record of the user is to browse, place an order, collect, pay and be added shopping cart.
Preferably, the generation recommending data includes that the online recommending data that generates generates recommending data with offline.
Preferably, the step of online generation recommending data：Data by reading flume or Kafka generate real
When recommending data array, mainly pass through user tag grouping and browsing record generate recommending data, recommending data：
A：User has browsed commodity A, directly recommends correlation type commodity B to user；
B：Recommended according to the type of merchandise of the hobby of user's selection；
B：Much-sought-after item；
C：User is male, just recommends the commodity of man, user is the commodity that Ms just recommends Ms；
Recommend weight：
A>B>C；
A is accounted for：50%；
B is accounted for：40%；
C is accounted for：10%；
Preferably, the step of offline generation recommending data：
(3.1) user data is obtained；
(3.2) scoring of user and commodity are generated；
(3.3) rating matrix of user and commodity are generated according to collaborative filtering；
(3.4) according to scoring come recommending data.
Preferably, user's portrait field source：User's essential attribute, customer consumption feature, user's value characteristic and
Subscriber lifecycle.
The present invention has clear advantage and advantageous effect compared with prior art, specifically, by above-mentioned technical proposal
Known to：
By using the method for the present invention so that commercial product recommending is more fine, can really reflect that user really actually thinks
The commodity wanted, more systematicization, it is that user really wants that the commodity of actual recommendation, which are more bonded, is businessman and user's all bands
Convenience is carried out.
More clearly to illustrate the structure feature and effect of the present invention, come below in conjunction with the accompanying drawings to this hair with specific embodiment
It is bright to be described in detail：
Description of the drawings
Fig. 1 is the overall procedure schematic diagram of the preferred embodiments of the invention；
Fig. 2 is commodity scoring product process figure in the preferred embodiments of the invention.
Specific implementation mode
Present invention is disclosed a kind of methods carrying out commercial product recommending according to the behavioral data of user, it is characterised in that：Including
There are following steps：
(1) it is drawn a portrait to user by the behavior record of user；The behavior record of the user be browse, place an order,
Collection, payment and addition shopping cart etc..
(2) scoring of the user to commodity is generated according to the relationship of user and commodity, then generates one according to collaborative filtering
Rating matrix of the user to commodity；
(3) user's portrait and user is combined to generate recommending data to the rating matrix of commodity.
The generation recommending data includes that the online recommending data that generates generates recommending data with offline.
The step of online generation recommending data：Data by reading flume or Kafka generate real-time recommendation number
According to array, mainly pass through user tag grouping and browsing record generate recommending data, recommending data：
A：User has browsed commodity A, directly recommends correlation type commodity B to user；
B：Recommended according to the type of merchandise of the hobby of user's selection；
B：Much-sought-after item；
C：User is male, just recommends the commodity of man, user is the commodity that Ms just recommends Ms；
Recommend weight：
A>B>C；
A is accounted for：50%；
B is accounted for：40%；
C is accounted for：10%；
The step of offline generation recommending data：
(3.1) user data is obtained；
(3.2) scoring of user and commodity are generated；
(3.3) rating matrix of user and commodity are generated according to collaborative filtering；
(3.4) according to scoring come recommending data.
User's portrait field source：User's essential attribute, customer consumption feature, user's value characteristic and user's life
Period.Specifically：
1, user's essential attribute：
Demographics：User identifier, name, mobile phone, mailbox, gender are obtained by the essential information of user；
Age of user：The discrete processes age (15 years old or less, -22 years old 15 years old, -30 years old 22 years old, -40 years old 30 years old, 50 years old with
On), it is obtained by the birthday of user information；
Regional Property：Country, province, city, district are obtained by the ship-to of user.
Social property：Whether kinsfolk's association has old man/child, children's age, professional attribute (IT user, company
White collar etc.), whether the specific products bought by user has old man or child to obtain, such as it is exactly 0-3 to have paper nappy
The child in year, it is exactly old man etc. to have health products.
2, customer consumption feature：
Classification preference：User's classification preference (containing commodity classification 1,2,3 grades of classifications), the commodity bought according to user obtain
The classification of commodity；
Brang Preference：Commodity/Brang Preference label, the commodity bought according to user obtain the brand of commodity；
Place an order custom：Lower list time, lower single geographical location, lower single device, summarize according to unirecord under user and obtain；
Period is purchased：Daily necessities, fast-moving consumer goods Buying Cycle label, the record bought according to user obtain multiplicating
The commodity of purchase；
Customer consumption feature：By category distinguish user characteristics (such as mobile phone digital intelligent, makeups intelligent), user touch
Up to preference, promotion sensitivity of user etc. -- temporary nothing, the type of merchandize bought according to user, which summarizes, to be obtained.
3, user's value characteristic：
Consumption level：Divided that (general consumption user, high-end disappears at value user according to amount of money section, purchase type
Take class user), the lower unirecord of user is obtained, then the consumer phase of matching system setting.
4, subscriber lifecycle：
Registration turns first purchase：Registration turns a first purchase user tag, the transformation of realized value client, judge user whether only at
Work(is paid for the first time；
Buy platform transition：PC shifts dynamic, H5 → APP → small routine, tracks user's purchasing habits, realizes multi-platform pushes away
It recommends, judges whether the register platforms and browsing of user, purchase platform have any different.
Flume is introduced：
Apache Flume are a distributing result collection systems, be by Cloudera companies develop it is a efficiently
Energy, high reliability and high restorative system.It can carry out efficiently collecting from a large amount of journal informations of separate sources, polymerize, moves
It is dynamic, it is finally stored into data center's stocking system.Framework is after reconstruct, from original Flume OG till now
Flume NG.Flume NG simply use and are easily adapted to different modes collector journal more like a light-weighted small external member, and
Support Failover and Load Balancing.
Architecture roles explanation：
Mainly there is following core in Flume frameworks:
·Event：One information unit can be attached to an optional message source.Ex:Daily record record, avro.
·Client：Operative position origin Event and be transferred to Flume Agent, mainly generate data, transport
Row is in an independent routine.
·Agent：One independent Flume formula, including Source, Channel, Sink.
·Source：For consuming the Event from the ends Client data collection to this, it is then delivered to Channel.
·Channel：A temporary storage space for converting Event, possesses the Event sent from Source.
·Sink:Event is read and removed from Channel, and Event is transmitted to the next of Flow Pipeline
A Agent (if present).
Kafka explanations：
Kafka is a message system, and exploitation originally is used as the active flow (Activity of LinkedIn from LinkedIn
Stream) and operation data processing pipeline (Pipeline) basis.It is by different types of company of more families as more now
The data pipe and message system of type use.
Movable flow data be in the data that nearly all website will be used when doing report to its website service condition most
Conventional part.Activity data includes page access amount (Page View), checked content in terms of information and search situation
Etc. contents.This common processing mode of data is first various activities to be written in the form of daily record certain file, then the period
Property it is for statistical analysis to these files.Operation data refers to the performance data of server (when CPU, IO utilization rate, request
Between, serve log etc. data).The statistical method type of operation data is various.
In recent years, activity and operation data processing have become one vital group in web site software product characteristic
At part, this just needs a set of slightly more complicated infrastructure to provide support to it.Main design goal is as follows：
Message duration ability is provided in such a way that time complexity is O (1), can be protected to TB grades of data above
Demonstrate,prove the access performance of constant time complexity.
High-throughput.It can accomplish that single machine supports 100K items per second or more to disappear on the business machine being dirt cheap
The transmission of breath.
It supports the message partition between Kafka Server and distributed consumer, while ensureing in each Partition
Message sequence transmits.
Off-line data processing and real time data processing are supported simultaneously.
·Scale out：Support online horizontal extension.
The present invention design focal point be：By using the method for the present invention so that commercial product recommending is more fine, can be real
The commodity that reflection user really actually wants to, more systematicization, it is that user really wants that the commodity of actual recommendation, which are more bonded,
, it is all brought conveniently for businessman and user.
The technical principle of the present invention is described above in association with specific embodiment.These descriptions are intended merely to explain the present invention's
Principle, and it cannot be construed to limiting the scope of the invention in any way.Based on the explanation herein, the technology of this field
Personnel would not require any inventive effort the other specific implementation modes that can associate the present invention, these modes are fallen within
Within protection scope of the present invention.
Claims (6)
1. a kind of method carrying out commercial product recommending according to the behavioral data of user, it is characterised in that：Include following steps：
(1) it is drawn a portrait to user by the behavior record of user；
(2) scoring of the user to commodity is generated according to the relationship of user and commodity, a user is then generated according to collaborative filtering
To the rating matrix of commodity；
(3) user's portrait and user is combined to generate recommending data to the rating matrix of commodity.
2. a kind of method carrying out commercial product recommending according to the behavioral data of user as described in claim 1, it is characterised in that：Institute
The behavior record for stating user is to browse, place an order, collect, pay and be added shopping cart.
3. a kind of method carrying out commercial product recommending according to the behavioral data of user as described in claim 1, it is characterised in that：Institute
It includes that the online recommending data that generates generates recommending data with offline to state and generate recommending data.
4. a kind of method carrying out commercial product recommending according to the behavioral data of user as claimed in claim 3, it is characterised in that：Institute
State online the step of generating recommending data：Data by reading flume or Kafka generate real-time recommendation array of data,
Record is mainly grouped and browsed by user tag generates recommending data, recommending data：
A：User has browsed commodity A, directly recommends correlation type commodity B to user；
B：Recommended according to the type of merchandise of the hobby of user's selection；
B：Much-sought-after item；
C：User is male, just recommends the commodity of man, user is the commodity that Ms just recommends Ms；
Recommend weight：
A>B>C；
A is accounted for：50%；
B is accounted for：40%；
C is accounted for：10%.
5. a kind of method carrying out commercial product recommending according to the behavioral data of user as claimed in claim 3, it is characterised in that：Institute
State offline the step of generating recommending data：
(3.1) user data is obtained；
(3.2) scoring of user and commodity are generated；
(3.3) rating matrix of user and commodity are generated according to collaborative filtering；
(3.4) according to scoring come recommending data.
6. a kind of method carrying out commercial product recommending according to the behavioral data of user as described in claim 1, it is characterised in that：Institute
State user's portrait field source：User's essential attribute, customer consumption feature, user's value characteristic and subscriber lifecycle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810367578.9A CN108596720A (en) | 2018-04-23 | 2018-04-23 | A method of commercial product recommending is carried out according to the behavioral data of user |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810367578.9A CN108596720A (en) | 2018-04-23 | 2018-04-23 | A method of commercial product recommending is carried out according to the behavioral data of user |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108596720A true CN108596720A (en) | 2018-09-28 |
Family
ID=63614001
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810367578.9A Pending CN108596720A (en) | 2018-04-23 | 2018-04-23 | A method of commercial product recommending is carried out according to the behavioral data of user |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108596720A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112036984A (en) * | 2020-09-04 | 2020-12-04 | 烟台冰兔网络科技有限公司 | E-commerce operation big data management system |
| CN112269932A (en) * | 2020-10-30 | 2021-01-26 | 金天国际医疗科技有限公司 | Big data-based small and medium enterprise resource integration processing system |
| CN113158023A (en) * | 2021-02-05 | 2021-07-23 | 杭州码全信息科技有限公司 | Public digital life accurate classification service method based on mixed recommendation algorithm |
| CN113407827A (en) * | 2021-06-11 | 2021-09-17 | 广州三七极创网络科技有限公司 | Information recommendation method, device, equipment and medium based on user value classification |
| CN113434770A (en) * | 2021-07-08 | 2021-09-24 | 广州康乾信息科技有限公司 | Business portrait analysis method and system combining electronic commerce and big data |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2096558A1 (en) * | 2008-02-29 | 2009-09-02 | Sony Corporation | Method for generating an ordered list of content items |
| 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 |
| US8781915B2 (en) * | 2008-10-17 | 2014-07-15 | Microsoft Corporation | Recommending items to users utilizing a bi-linear collaborative filtering model |
| 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 |
| CN106651533A (en) * | 2016-12-29 | 2017-05-10 | 合肥华凌股份有限公司 | User behavior-based personalized product recommendation method and apparatus |
| CN107633430A (en) * | 2017-09-20 | 2018-01-26 | 哈尔滨工业大学 | A kind of Method of Commodity Recommendation based on community of colony |
-
2018
- 2018-04-23 CN CN201810367578.9A patent/CN108596720A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2096558A1 (en) * | 2008-02-29 | 2009-09-02 | Sony Corporation | Method for generating an ordered list of content items |
| US8781915B2 (en) * | 2008-10-17 | 2014-07-15 | Microsoft Corporation | Recommending items to users utilizing a bi-linear collaborative filtering model |
| 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 |
| 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 |
| CN106651533A (en) * | 2016-12-29 | 2017-05-10 | 合肥华凌股份有限公司 | User behavior-based personalized product recommendation method and apparatus |
| CN107633430A (en) * | 2017-09-20 | 2018-01-26 | 哈尔滨工业大学 | A kind of Method of Commodity Recommendation based on community of colony |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112036984A (en) * | 2020-09-04 | 2020-12-04 | 烟台冰兔网络科技有限公司 | E-commerce operation big data management system |
| CN112269932A (en) * | 2020-10-30 | 2021-01-26 | 金天国际医疗科技有限公司 | Big data-based small and medium enterprise resource integration processing system |
| CN113158023A (en) * | 2021-02-05 | 2021-07-23 | 杭州码全信息科技有限公司 | Public digital life accurate classification service method based on mixed recommendation algorithm |
| CN113407827A (en) * | 2021-06-11 | 2021-09-17 | 广州三七极创网络科技有限公司 | Information recommendation method, device, equipment and medium based on user value classification |
| CN113434770A (en) * | 2021-07-08 | 2021-09-24 | 广州康乾信息科技有限公司 | Business portrait analysis method and system combining electronic commerce and big data |
| CN113434770B (en) * | 2021-07-08 | 2022-09-09 | 上海识致信息科技有限责任公司 | Business portrait analysis method and system combining electronic commerce and big data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108596720A (en) | A method of commercial product recommending is carried out according to the behavioral data of user | |
| Rabinovich et al. | Physical distribution service quality in Internet retailing: service pricing, transaction attributes, and firm attributes | |
| Heinemann et al. | New online retailing | |
| CN104820928A (en) | Rapid shopping realization system based on mobile internet and method thereof | |
| CN107430618A (en) | Realize the system and method interacted with host computer device progress user speech | |
| CN106886910A (en) | The recommendation method and apparatus of consumption information, the method and apparatus for spelling list | |
| CN102521355A (en) | System for making statistics on relativity of network information promoters and network visitor target behavior | |
| CN105590233A (en) | Method and system for releasing information of commodities or service | |
| JP2015149096A (en) | Advertisement providing system and method for providing interface for integrated payment for products in integrated marketplace | |
| JP2015011500A (en) | Information processor and information processing method | |
| CN110335106A (en) | A method of increasing conversion in e-commerce website and sale is urgent | |
| CN107403329A (en) | A kind of business voucher collocation method and device | |
| Heyns et al. | Online shopping behaviour and service quality perceptions of young people in South Africa: A COVID-19 perspective | |
| KR101539625B1 (en) | Server and system for mobile market platform | |
| TWI488135B (en) | Combined with the group preferences of the real-time regional buy merchandise media system | |
| TWI482111B (en) | Affiliate marketing management device, affiliate marketing management method, affiliate marketing management program and computer program that memorizes its program | |
| JP7011552B2 (en) | Ad management system, ad management method, and ad management program | |
| JP2018521435A (en) | System and method for accurate marketing by sharing data, gifts and logistics | |
| JP7339688B2 (en) | Space station-based product resale method, system and recording medium | |
| Reveiu et al. | Mining social media to identify the immediate impact of COVID-19 pandemic on the Romanian retailers: early findings | |
| CN103208074A (en) | Service system and method | |
| Kulshrestha et al. | The Analytical Study of an E-Business Model for Establishing the Concept of Customer Retention (A Case Study of Myntra. com) | |
| Teltzrow et al. | Web metrics for retailers | |
| JP2014170419A (en) | Sales promotion apparatus, sales promotion system, sales promotion method, and sales promotion program | |
| JP7339686B2 (en) | Gift smart matching method, system and recording medium based on space station |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180928 |
|
| RJ01 | Rejection of invention patent application after publication |