CN106960354A - Method and device is recommended in a kind of precision based on customer life cycle - Google Patents
Method and device is recommended in a kind of precision based on customer life cycle Download PDFInfo
- Publication number
- CN106960354A CN106960354A CN201610013249.5A CN201610013249A CN106960354A CN 106960354 A CN106960354 A CN 106960354A CN 201610013249 A CN201610013249 A CN 201610013249A CN 106960354 A CN106960354 A CN 106960354A
- Authority
- CN
- China
- Prior art keywords
- life cycle
- customer
- client
- customer life
- characteristic value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- 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
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- 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
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0203—Market surveys; Market polls
-
- 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
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0241—Advertisements
- G06Q30/0251—Targeted advertisements
- G06Q30/0255—Targeted advertisements based on user history
Landscapes
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Game Theory and Decision Science (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Recommend method and device the invention discloses a kind of precision based on customer life cycle, methods described includes：Set up customer life cycle unified view；The customer life cycle unified view is parsed, customer life cycle characteristic value is obtained, and the customer life cycle characteristic value is updated by sample set training；According to the customer life cycle characteristic value, customers' division is carried out using the collaborative filtering method based on client；Search and match content recommendation with the customer life cycle characteristic value of stating, the content recommendation is recommended into corresponding customers.
Description
Technical field
Recommend the present invention relates to business support field, more particularly to a kind of precision based on customer life cycle
Method and device.
Background technology
With the development of Internet era, being growing more intense of market competition, customer demand also increasingly variation and
It is personalized.Between different industries, such as：Telecom operators, Internet firm, large supermarket, bank etc.,
Each enterprise is owned by the essential information of magnanimity client, service order behavior record etc..How these data to be allowed
Used by enterprise, by technologies such as data minings, its potential value is played, aids in formulating for different clients
Differentiation, personalized business recommended scheme, and be the suitable business of its accurate recommendation in the suitable time,
The key factor possessed as corporate client.
By taking telecommunications industry as an example, existing precision recommended technology is mostly current or recent according to client
The characteristics such as call behavior, consumer behavior, order relations are analyzed and excavated, and then development voice,
Flow or terminal recommendation activity.And precision recommendation method is all based on the data in available data warehouse, including
Client's essential information, client's order record, customer voice ticket and client's internet records etc., pass through these bases
Plinth data, set up customer historical behavior model, excavate classifying rules, and association analysis obtains similar customers,
And then targetedly carry out the recommendation of voice set meal, flow package or terminal device.
Existing technical scheme at least has the following disadvantages：
1st, data storage is disperseed, and the data storage in units of client is not supported, is not easy to statistics client whole
Behavior in communication cycle；
Available data warehouse is not that data storage is carried out in units of individual consumers, but is made with customer action
Based on unit stored, the service order that the data in such as data warehouse are the scattered storage of a rule is closed
System, voice ticket, flow ticket etc. are recorded, this isolated data storage method, it is impossible to which client is entered certainly
Behavior from net in whole communication cycle carries out whole tracking and association analysis.
2nd, data storage has fixed resting period, is unfavorable for analyzing the behavior in the whole communication cycle of client and becomes
Change trend；
The resting period of customer data is fixed in data warehouse, and such as client's ticket is only stored nearest six months
Interior data etc., can only reflect the recent behavioural characteristic of client, it is impossible to learn client from the behavior networked so far
Variation tendency, including voice use the change of duration, flow to use how many change etc..
3rd, it can not reflect that client is presently in the stage of life cycle, recommendation results have one-sidedness.
Existing recommendation method can not set up complete Client view image, few potential phases from client, into
Client's ladder growth management system is set up again to the angle of this Life cycle of decline phase to the maturity period for a long time,
And it is business recommended to carry out precision.It is precisely business recommended not to be truly realized effectively identification potential customers, essence
Recommendation message, guiding client is really delivered to improve brand dependence, reduce customer churn probability etc..Recommendation results are very
There is one-sidedness in big degree, it is difficult to meet client's actual demand, cause recommendation effect difference and spend cost compared with
Height, while influenceing client perception, reduces CSAT.
The content of the invention
In order to solve the above technical problems, the embodiments of the invention provide a kind of based on the accurate of customer life cycle
Change and recommend method and device.
Method is recommended in precision provided in an embodiment of the present invention based on customer life cycle, including：
Set up customer life cycle unified view；
The customer life cycle unified view is parsed, customer life cycle characteristic value is obtained, and lead to
Cross sample set training and update the customer life cycle characteristic value；
According to the customer life cycle characteristic value, customers are carried out using the collaborative filtering method based on client
Divide；
Search and match content recommendation with the customer life cycle characteristic value of stating, the content recommendation is recommended
To corresponding customers.
It is described to set up customer life cycle unified view in the embodiment of the present invention, including：
Set up mapping table；
According to the mapping table, the data of the Behavior-based control stored in database are mapped to based on client
Data, form the customer life cycle unified view.
In the embodiment of the present invention, the lookup matches content recommendation with the customer life cycle characteristic value of stating,
Including：
For the different life stage of customer life cycle, search corresponding with the life cycle phase
Content recommendation.
In the embodiment of the present invention, methods described also includes, including：
Using K arest neighbors sorting techniques, customer life cycle characteristic value is classified；
When the sample that target sample subscribes number with exceeding in K most like samples in sample set belongs to same
During one classification, the target sample belongs to the K most like corresponding classifications of sample, and the target
Sample has the sample characteristics of the classification.
In the embodiment of the present invention, methods described also includes, including：
Sample set is set up, wherein, the sample set stores customer life cycle characteristic value；
The sample set is trained, the life cycle characteristic value of whole clients is extracted.
It is described according to the customer life cycle characteristic value in the embodiment of the present invention, using the association based on client
Customers' division is carried out with filter method；Search and match with the customer life cycle characteristic value of stating in recommendation
Hold, including：
According to the customer life cycle characteristic value, search and target customer's similarity highest client set；
The target customer is divided in the client set；
Search and the client set matching degree highest content recommendation.
Precision recommendation apparatus provided in an embodiment of the present invention based on customer life cycle, including：
Unit is set up, for setting up customer life cycle unified view；
Characteristics extraction unit, for being parsed to the customer life cycle unified view, obtains client
Life cycle characteristic value, and the customer life cycle characteristic value is updated by sample set training；
Customers' division unit, for according to the customer life cycle characteristic value, using the association based on client
Customers' division is carried out with filter method；
Recommendation unit, matches content recommendation for searching with the customer life cycle characteristic value of stating, by institute
State content recommendation and recommend corresponding customers.
It is described to set up unit in the embodiment of the present invention, it is further used for：
Set up mapping table；
According to the mapping table, the data of the Behavior-based control stored in database are mapped to based on client
Data, form the customer life cycle unified view.
In the embodiment of the present invention, the recommendation unit is further used for：
For the different life stage of customer life cycle, search corresponding with the life cycle phase
Content recommendation.
In the embodiment of the present invention, described device also includes：Taxon, for using K arest neighbors classification side
Method, classifies to customer life cycle characteristic value；When K in target sample and sample set are most like
When the sample for exceeding reservation number in sample belongs to same category, the target sample belongs to the K most phases
As the corresponding classification of sample, and the target sample has the sample characteristics of the classification.
In the embodiment of the present invention, the characteristics extraction unit is further used for：
Sample set is set up, wherein, the sample set stores customer life cycle characteristic value；
The sample set is trained, the life cycle characteristic value of whole clients is extracted.
In the embodiment of the present invention, customers' division unit is further used for：According to client's life
Periodic proper value, is searched and target customer's similarity highest client set；The target customer is divided to
In the client set；
The recommendation unit, is further used for：Search and the client set matching degree highest content recommendation.
In the technical scheme of the embodiment of the present invention, customer life cycle unified view is set up；The client is given birth to
Life cycle unified view is parsed, and obtains customer life cycle characteristic value, and train renewal by sample set
The customer life cycle characteristic value；According to the customer life cycle characteristic value, using the association based on client
Customers' division is carried out with filter method；Search and match with the customer life cycle characteristic value of stating in recommendation
Hold, the content recommendation is recommended into corresponding customers.It can be seen that, the embodiment of the present invention changes existing number
According to the storage mode in warehouse, storage cell is used as using individual consumers, it is ensured that can accurately depict client and exist
The view of whole communication life cycle.Customer life cycle characterization factor is introduced into proposed algorithm simultaneously, greatly
The big accuracy for improving recommendation.
Brief description of the drawings
Fig. 1 recommends the flow signal of method for the precision based on customer life cycle of the embodiment of the present invention
Figure；
Fig. 2 is the two-stage mapping relations schematic diagram of the embodiment of the present invention；
Fig. 3 is current database storage organization schematic diagram；
Fig. 4 is the structure chart of the customer life cycle unified view of the embodiment of the present invention；
Fig. 5 is the customer life cycle schematic diagram of the embodiment of the present invention；
Fig. 6 shows for the structure composition of the precision recommendation apparatus based on customer life cycle of the embodiment of the present invention
It is intended to.
Embodiment
The characteristics of in order to more fully hereinafter understand the embodiment of the present invention and technology contents, below in conjunction with the accompanying drawings
Realization to the embodiment of the present invention is described in detail, appended accompanying drawing purposes of discussion only for reference, is not used for
Limit the embodiment of the present invention.
The technical scheme of the embodiment of the present invention, 1) storage mode of the foundation in units of individual consumers；Change existing
There is the data storage mode of data warehouse, using individual consumers as storage cell, store the whole Life Cycle of client
The information and behavioral data of phase, it is ensured that recommend the integrality of required customer information.2) by client's Life Cycle
The phase factor, is added in business recommended algorithm, improves the levels of precision recommended.Such as, by telecom client from
Network to off-network and regard a life cycle as, inventionbroadly, client can be ordered to different classes of product and regarded as
The life cycle of client under the product, the whole life cycle of client can be divided into the different stages, in life
The client of life cycle different phase possesses its distinctive behavioural characteristic.The customer life cycle factor is incorporated into industry
It is engaged in recommendation process, and then carries out precision marketing, is the high product of lead referral acceptance, improves telecommunications
Success rate is recommended by enterprise, improves customers' approval degree.
The embodiment of the present invention recommends method using the precision based on the customer life cycle factor, to carry out precisely
It is business recommended.Firstly, it is necessary to set up customer life cycle unified view, by each client is all and life
The feature that cycle is associated is easily storage, data easy to identify by two-stage Mapping and Converting.Then, it is based on
Above-mentioned basic data carries out customer life cycle characteristic value excavation, and trains continuous updating feature by sample set
Value.And then, customers' subdivision is carried out, using the collaborative filtering based on client, there will be similar interests
The user of hobby is divided, and the business of its preference has more similitude.Finally, recommendation list is formed, will
The client that matching degree highest Products Show needs to most probable.The overall flow figure reference of the embodiment of the present invention
Fig. 1.
Fig. 1 recommends the flow signal of method for the precision based on customer life cycle of the embodiment of the present invention
Figure, as shown in figure 1, the precision based on customer life cycle recommends method to comprise the following steps：
Step 101：Set up customer life cycle unified view.
In the embodiment of the present invention, mapping table is set up；According to the mapping table, it will be deposited in database
The data of the Behavior-based control of storage are mapped to the data based on client, form the customer life cycle unified view.
Specifically, customer life cycle unified view is set up using virtual view, by the storage side of data warehouse
Formula is adjusted, and will be changed to the mode based on individual consumers based on the storage mode of customer action.For
The influence to entity table in data warehouse is reduced as far as possible, and the overall situation for establishing customer life cycle is uniformly regarded
Individual consumers are carried out mapping storage, by two-stage mapping relations, by client's life by figure using virtual view
Cycle related table and related data are associated, by the change triggers signaling of customer status, to data
Storehouse table carries out corresponding additions and deletions and changes operation.
Reference picture 2, first extract customer life cycle feature (including client's essential information, customer status,
Client's order relations, customer action data), the mapping for setting up a customer life cycle and backstage entity table is closed
It is table, then mapping table is set up contacts with backstage entity table, is associated using Customer ID.From visitor
Family networks and just sets up such one group of mapping relations.When the life cycle characteristic of client produces change or database
Table produce change, just triggering signaling scanning mapping table, addition or delete correlative connection.Work as client
After off-network, triggering signaling is taken pictures the life cycle unified view of the client, while by the number of the client
According to storehouse relative recording and related mapping table record storage into disk file, database data is deleted.
Reference picture 3, Fig. 3 is current database storage organization schematic diagram, and the data stored in Fig. 3 are based on
The behavior of client carries out data storage.With reference to table 1, table 1 is mapping table, described in mapping table
The corresponding attribute of each user, attribute can be the basic document of client, order record etc..
| Sequence number | Client | Attribute |
| 1 | C1 | P1 |
| 2 | C2 | P2 |
| …… | …… | …… |
| N | Cn | Pn |
Table 1
With virtual view technology, by setting up two-stage mapping relations, form customer life cycle and uniformly regard
Scheme, now data store organisation reference picture 4, it is unit that the data of storage, which are based on individual consumers,.
Step 102：The customer life cycle unified view is parsed, customer life cycle is obtained special
Value indicative, and the customer life cycle characteristic value is updated by sample set training.
In the embodiment of the present invention, using K arest neighbors sorting techniques, customer life cycle characteristic value is divided
Class；When the sample that target sample subscribes number with exceeding in K most like samples in sample set belongs to same
During one classification, the target sample belongs to the K most like corresponding classifications of sample, and the target
Sample has the sample characteristics of the classification.
In the embodiment of the present invention, sample set is set up, wherein, the sample set stores customer life cycle feature
Value；The sample set is trained, the life cycle characteristic value of whole clients is extracted.
Step 103：According to the customer life cycle characteristic value, using the collaborative filtering method based on client
Carry out customers' division.
In the embodiment of the present invention, according to the customer life cycle characteristic value, search and target customer's similarity
Highest client set；The target customer is divided in the client set.
As it was previously stated, having had built up the data storage mechanism of customer life cycle, therefore client can be obtained
From all data till now of networking, and then customer life cycle this crucial sex factor is extracted, as pushing away
Recommend one of condition of algorithm.A proposed algorithm model is set up, all data and recommendation mesh for client are inputted
Mark, defines computation rule, obtains the life cycle index factor of client, then regard these information as triggering bar
Part, is recommended accordingly.
First, customer life cycle feature is classified：Using K arest neighbors sorting techniques, to non-person in charge of reception at ceremonies
The life cycle at family carries out tagsort.If in the K in feature set most like samples of a sample
It is most of belong to same classification, then the sample falls within this classification, and with the sample in this classification
This feature.First, a sample set is set up, wherein, sample set storage is client and its place life
The characteristic value in cycle；By taking telecommunications industry as an example：The flow product growth stage feature of client may be included：Continuously
Flow package is monthly ordered, and increasing trend etc. monthly is presented using flow.Secondly, sample set is instructed
Practice study, so that the life cycle characteristic of full dose client is extracted, and output result, i.e., client is not
With the life cycle ratio characteristics value on product, the input parameter of subsequent recommendation algorithm is used as.
Classification to implement step as follows：
Initialization distance is maximum；
Calculate unknown sample and each training sample apart from d：
Current K are obtained closest to the ultimate range maxd in sample；
If d is less than maxd, the training sample is regard as K- nearest samples；
Repeat step 2,3,4, until the distance of unknown sample and all training samples is all complete；
Count the number of times that each class label cycle occurs in K- nearest samples；
The maximum class label cycle of the frequency of occurrences is selected as the class label cycle of unknown sample.
Then, customers are divided：Customers' subdivision is carried out using the collaborative filtering based on client,
Customer life cycle characteristic value this result that binding modules 1 are drawn, carries out precision and recommends.Its core is thought
Think be：The thing that similar client likes is also likely to be similar.I.e.：If a client wants to read a book,
But not having fixed direction, then this client can seek help from friend or the reading of upper bean cotyledon at one's side
What look at Deng website either with or without good recommendation, common customer often selects to sample close friend with oneself
Recommendation.
Implement step as follows：
Find the client set similar to target customer's interest；
Using revised cosine similarity method, the similarity between target customer and other clients is calculated,
If N (u) is the product set that client u likes, N (v) is the product set that client v likes, and Cycle (u) is visitor
Life cycle characteristic value residing for the u of family, Cycle (v) is the life cycle characteristic value residing for client v, then u
Similarity with v is：
Wherein, μ represents Cycle (u) and Cycle (v) arithmetic mean of instantaneous value.
By the above-mentioned customer life cycle characteristic value cycle calculated, as a factor, entirety is added to
Suggested design in, be favorably improved the accuracy in computation of client's similarity, life cycle characteristic value is more like
Client, the similar probability of interest can be some higher.
Step 104：Search and match content recommendation with the customer life cycle characteristic value of stating, pushed away described
Commending contents are recommended to corresponding customers.
In the embodiment of the present invention, for the different life stage of customer life cycle, search and the life
Order the phase of the cycles corresponding content recommendation.
Specifically, customer life cycle refer to a client be for enterprise have similar life the same birth,
Growth, maturation, aging, dead process.It is also different specific to different definition of industry, such as in telecommunications row
Industry, so-called customer life cycle, refer to client since the client as telecommunications company and generation business consume
Start, consumption is grown up, consumption is stable, consumption declines, finally off-grid process.
Client's whole life cycle can be divided into the different stages by we, positioned at life cycle different phase
Client, its behavior takes on a different character.Potential stage phase, client may just network soon or just
Some classs of business are just ordered；Stage growth stage, voice, GPRS (GPRS, General
Packet Radio Service) etc. business usage amount increase rapidly, constantly order new service lists；And it is ripe
More stable trend is presented in all kinds of business usage amounts that phase then shows as client, and order relations will not occur greatly
Change；Decline phase, all kinds of business usage amounts of client start rapid decline, continuously cancel multinomial business and order
Purchase relation etc..
Reference picture 5, Fig. 5 is the customer life cycle schematic diagram of the embodiment of the present invention, according to client's Life Cycle
Phase development figure can be seen that the different phase of life cycle according to residing for client can pointedly formulate business and push away
Recommend strategy, carry out client's precision marketing and client maintains and keep work, from keeping into thing afterwards,
Transformation is kept in advance, so as to lift the lifelong value and loyalty of client.When client is in growth stage and maturation
Period, secondary way of recommendation extension client can be taken in net use time, enhancing client uses stickiness；Work as visitor
During family is in decline, when having off-network tendency, the personalized ground hand such as business presentation or activity discount can be taken
Section, promotes client to recover service, improves customer satisfaction and loyalty.
Customer life cycle in this programme is a more wide in range concept, is not only referred to for networking
The life cycle of whole industry, and the life cycle comprising different product, i.e. different phonetic set meal, flow
Use life cycle of order life cycle or different terminals series of set meal etc..Client after networking,
For different products, different life-cycle stages can be corresponded to, be can be used to respectively for difference
Product carry out precision recommendation.Life cycle phase residing for client, precision business is carried out for enterprise
Recommendation has important reference value, can greatly improve the matching degree of recommendation business and client's actual demand, increases
Business handling success rate and utilization rate after big precision recommendation.
During content recommendation, find it is that client in client set likes and can match to the full extent client work as
The product list of preceding actual demand：
Firstly the need of the K client most like with target customer u is found out from matrix, with set S (u, K) table
Show, the product that client in S likes all is extracted, and remove the product that u has liked.For every
Individual candidate products i, degree interested in it client u is calculated with equation below：
Wherein, rviRepresent that client v likes degree to i.
Assuming that we will give A recommended products, K=3 similar clients are chosen, similar client is assumed to be：B、
C, D, it can be seen from customer life cycle unified view, the product that they like and A was not subscribed to has：
C, e, then calculating p respectively, (A, c) (A, value e), computational methods are as follows with p：
P (A, c)=wAB+wAD
P (A, e)=wAC+wAD
Comparing p, (A, c) (it is the higher Products Show of value then to choose preference for A, value size e) with p
Client A is given, the successful probability of recommendation is higher, and client's acceptance and satisfaction are more preferable.
The customer priorities and the information of target product produced according to above-mentioned steps, obtain client to its similar client
The higher product list of ordered product fancy grade, as the lead referral product similar to customer priorities.
The technical scheme of the embodiment of the present invention, carries out data storage classification by customer life cycle, enhances number
According to break-up value；Added customer life cycle as label in proposed algorithm, it is perfect to recommend to calculate
The isolatism of method, the critical-path analysis data and feature of client's different cycles greatly improve recommendation accurate
Property, recommendation hit rate and actual purchase rate are improved, client has effectively been carried out and has possessed.According to client's Life Cycle
Phase is theoretical, for the client in different cycles, can be excavated according to its key characteristics potential unknown
Preference, most suitable Products Show is given the client needed most.
Fig. 6 shows for the structure composition of the precision recommendation apparatus based on customer life cycle of the embodiment of the present invention
It is intended to, as shown in fig. 6, the precision recommendation apparatus based on customer life cycle includes：
Unit 61 is set up, for setting up customer life cycle unified view；
Characteristics extraction unit 62, for being parsed to the customer life cycle unified view, obtains visitor
Family life cycle characteristic value, and the customer life cycle characteristic value is updated by sample set training；
Customers' division unit 63, for according to the customer life cycle characteristic value, using based on client's
Collaborative filtering method carries out customers' division；
Recommendation unit 64, matches content recommendation for searching with the customer life cycle characteristic value of stating, will
The content recommendation recommends corresponding customers.
It is described to set up unit 61, it is further used for：
Set up mapping table；
According to the mapping table, the data of the Behavior-based control stored in database are mapped to based on client
Data, form the customer life cycle unified view.
The recommendation unit 64, is further used for：
For the different life stage of customer life cycle, search corresponding with the life cycle phase
Content recommendation.
Described device also includes：Taxon 65, for using K arest neighbors sorting techniques, to client's life
Periodic proper value is classified；Subscribed when exceeding in K most like samples in target sample and sample set
When the sample of number belongs to same category, it is corresponding that the target sample belongs to the K most like samples
Classification, and the target sample has the sample characteristics of the classification.
The characteristics extraction unit 62, is further used for：
Sample set is set up, wherein, the sample set stores customer life cycle characteristic value；
The sample set is trained, the life cycle characteristic value of whole clients is extracted.
Customers' division unit 63, is further used for：According to the customer life cycle characteristic value, look into
Look for and target customer's similarity highest client set；The target customer is divided in the client set；
The recommendation unit 64, is further used for：Search interior with client set matching degree highest recommendation
Hold.
It will be appreciated by those skilled in the art that dress is recommended in the precision based on customer life cycle shown in Fig. 6
Each unit in putting realizes that function can refer to the phase that method is recommended in the foregoing precision based on customer life cycle
Close description and understand.
, can be in any combination in the case where not conflicting between technical scheme described in the embodiment of the present invention.
In several embodiments provided by the present invention, it should be understood that disclosed method and smart machine,
It can realize by another way.Apparatus embodiments described above are only schematical, for example,
The division of the unit, only a kind of division of logic function, can there is other division side when actually realizing
Formula, such as：Multiple units or component can be combined, or be desirably integrated into another system, or some features can
To ignore, or do not perform.In addition, the coupling or straight of shown or discussed each part each other
Connect coupling or communication connection can be by some interfaces, the INDIRECT COUPLING or communication connection of equipment or unit,
Can be electrical, machinery or other forms.
The above-mentioned unit illustrated as separating component can be or may not be it is physically separate, as
The part that unit is shown can be or may not be physical location, you can positioned at a place, also may be used
To be distributed on multiple NEs；Part or all of unit therein can be selected according to the actual needs
Realize the purpose of this embodiment scheme.
In addition, each functional unit in various embodiments of the present invention can be fully integrated into a second processing list
, can also two or more unit collection in member or each unit is individually as a unit
Into in a unit；Above-mentioned integrated unit can both be realized in the form of hardware, it would however also be possible to employ hard
Part adds the form of SFU software functional unit to realize.
The foregoing is only a specific embodiment of the invention, but protection scope of the present invention is not limited to
This, any one skilled in the art the invention discloses technical scope in, can readily occur in
Change or replacement, should all be included within the scope of the present invention.
Claims (12)
1. method is recommended in a kind of precision based on customer life cycle, it is characterised in that methods described includes：
Set up customer life cycle unified view；
The customer life cycle unified view is parsed, customer life cycle characteristic value is obtained, and lead to
Cross sample set training and update the customer life cycle characteristic value；
According to the customer life cycle characteristic value, customers are carried out using the collaborative filtering method based on client
Divide；
Search and match content recommendation with the customer life cycle characteristic value of stating, the content recommendation is recommended
To corresponding customers.
2. method is recommended in the precision according to claim 1 based on customer life cycle, its feature exists
In, it is described to set up customer life cycle unified view, including：
Set up mapping table；
According to the mapping table, the data of the Behavior-based control stored in database are mapped to based on client
Data, form the customer life cycle unified view.
3. method is recommended in the precision according to claim 1 based on customer life cycle, its feature exists
In, the lookup matches content recommendation with the customer life cycle characteristic value of stating, including：
For the different life stage of customer life cycle, search corresponding with the life cycle phase
Content recommendation.
4. method is recommended in the precision according to claim 1 based on customer life cycle, its feature exists
In, methods described also includes, including：
Using K arest neighbors sorting techniques, customer life cycle characteristic value is classified；
When the sample that target sample subscribes number with exceeding in K most like samples in sample set belongs to same
During one classification, the target sample belongs to the K most like corresponding classifications of sample, and the target
Sample has the sample characteristics of the classification.
5. method is recommended in the precision according to claim 4 based on customer life cycle, its feature exists
In, methods described also includes, including：
Sample set is set up, wherein, the sample set stores customer life cycle characteristic value；
The sample set is trained, the life cycle characteristic value of whole clients is extracted.
6. method is recommended in the precision based on customer life cycle according to any one of claim 1 to 5,
Characterized in that, it is described according to the customer life cycle characteristic value, using the collaborative filtering side based on client
Method carries out customers' division；Search and match content recommendation with the customer life cycle characteristic value of stating, including：
According to the customer life cycle characteristic value, search and target customer's similarity highest client set；
The target customer is divided in the client set；
Search and the client set matching degree highest content recommendation.
7. a kind of precision recommendation apparatus based on customer life cycle, it is characterised in that described device includes：
Unit is set up, for setting up customer life cycle unified view；
Characteristics extraction unit, for being parsed to the customer life cycle unified view, obtains client
Life cycle characteristic value, and the customer life cycle characteristic value is updated by sample set training；
Customers' division unit, for according to the customer life cycle characteristic value, using the association based on client
Customers' division is carried out with filter method；
Recommendation unit, matches content recommendation for searching with the customer life cycle characteristic value of stating, by institute
State content recommendation and recommend corresponding customers.
8. the precision recommendation apparatus according to claim 7 based on customer life cycle, its feature exists
In, it is described to set up unit, it is further used for：
Set up mapping table；
According to the mapping table, the data of the Behavior-based control stored in database are mapped to based on client
Data, form the customer life cycle unified view.
9. the precision recommendation apparatus according to claim 7 based on customer life cycle, its feature exists
In the recommendation unit is further used for：
For the different life stage of customer life cycle, search corresponding with the life cycle phase
Content recommendation.
10. the precision recommendation apparatus according to claim 7 based on customer life cycle, its feature
It is that described device also includes：Taxon, for using K arest neighbors sorting techniques, to client's life
Periodic proper value is classified；Subscribed when exceeding in K most like samples in target sample and sample set
When the sample of number belongs to same category, it is corresponding that the target sample belongs to the K most like samples
Classification, and the target sample has the sample characteristics of the classification.
11. the precision recommendation apparatus according to claim 10 based on customer life cycle, its feature
It is that the characteristics extraction unit is further used for：
Sample set is set up, wherein, the sample set stores customer life cycle characteristic value；
The sample set is trained, the life cycle characteristic value of whole clients is extracted.
12. dress is recommended in the precision based on customer life cycle according to any one of claim 7 to 11
Put, it is characterised in that customers' division unit, be further used for：According to the customer life cycle
Characteristic value, is searched and target customer's similarity highest client set；The target customer is divided to described
In client set；
The recommendation unit, is further used for：Search and the client set matching degree highest content recommendation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610013249.5A CN106960354A (en) | 2016-01-11 | 2016-01-11 | Method and device is recommended in a kind of precision based on customer life cycle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610013249.5A CN106960354A (en) | 2016-01-11 | 2016-01-11 | Method and device is recommended in a kind of precision based on customer life cycle |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106960354A true CN106960354A (en) | 2017-07-18 |
Family
ID=59480851
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610013249.5A Pending CN106960354A (en) | 2016-01-11 | 2016-01-11 | Method and device is recommended in a kind of precision based on customer life cycle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106960354A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108090785A (en) * | 2017-12-07 | 2018-05-29 | 北京小度信息科技有限公司 | Determine the method, apparatus and electronic equipment of user behavior decline tendency |
| CN109491641A (en) * | 2018-10-25 | 2019-03-19 | 广西大学 | A kind of construction method of product operational management data model |
| TWI658420B (en) * | 2016-11-15 | 2019-05-01 | 平安科技（深圳）有限公司 | Method, device, server and computer readable storage medium for integrate collaborative filtering with time factor |
| CN112508607A (en) * | 2020-12-02 | 2021-03-16 | 上海中商网络股份有限公司 | Data pushing method and device, computer equipment and readable storage medium |
| CN112819520A (en) * | 2021-01-27 | 2021-05-18 | 北京思特奇信息技术股份有限公司 | Method and system for calculating life cycle of operator customer |
| TWI730273B (en) * | 2018-12-17 | 2021-06-11 | 兆豐國際商業銀行股份有限公司 | Potential client searching method and potential client searching device |
| CN112967102A (en) * | 2021-02-04 | 2021-06-15 | 江苏警官学院 | Method for establishing customer portrait by logistics data |
| CN112989183A (en) * | 2021-02-20 | 2021-06-18 | 湖南视拓信息技术股份有限公司 | Product information recommendation method and device based on life cycle and related equipment |
| CN113297478A (en) * | 2021-04-25 | 2021-08-24 | 上海淇玥信息技术有限公司 | Information pushing method and device based on user life cycle and electronic equipment |
| CN113591018A (en) * | 2021-07-30 | 2021-11-02 | 中国联合网络通信集团有限公司 | Communication client classification management method, system, electronic device and storage medium |
| CN114143772A (en) * | 2021-11-18 | 2022-03-04 | 北京思特奇信息技术股份有限公司 | Method and system for reducing user off-network rate |
| CN114742569A (en) * | 2021-01-08 | 2022-07-12 | 广州视源电子科技股份有限公司 | User life stage prediction method and device, computer equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2003067555A (en) * | 2001-08-24 | 2003-03-07 | Takashi Ishino | System for increasing repeat rate of fishing customer for leisure fishing boat operator |
| CN104104724A (en) * | 2014-07-23 | 2014-10-15 | 深圳市富途网络科技有限公司 | Method for pushing operating mode of client side and communication system |
| CN104298755A (en) * | 2014-10-20 | 2015-01-21 | 北京奇虎科技有限公司 | Content push method, content push system and server |
-
2016
- 2016-01-11 CN CN201610013249.5A patent/CN106960354A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2003067555A (en) * | 2001-08-24 | 2003-03-07 | Takashi Ishino | System for increasing repeat rate of fishing customer for leisure fishing boat operator |
| CN104104724A (en) * | 2014-07-23 | 2014-10-15 | 深圳市富途网络科技有限公司 | Method for pushing operating mode of client side and communication system |
| CN104298755A (en) * | 2014-10-20 | 2015-01-21 | 北京奇虎科技有限公司 | Content push method, content push system and server |
Non-Patent Citations (2)
| Title |
|---|
| 王建: "移动数据业务营销平台的需求分析与设计", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 田颖: "基于用户消费行为的移动业务推荐", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI658420B (en) * | 2016-11-15 | 2019-05-01 | 平安科技（深圳）有限公司 | Method, device, server and computer readable storage medium for integrate collaborative filtering with time factor |
| CN108090785A (en) * | 2017-12-07 | 2018-05-29 | 北京小度信息科技有限公司 | Determine the method, apparatus and electronic equipment of user behavior decline tendency |
| CN108090785B (en) * | 2017-12-07 | 2021-03-02 | 北京小度信息科技有限公司 | Method and device for determining user behavior decline tendency and electronic equipment |
| CN109491641A (en) * | 2018-10-25 | 2019-03-19 | 广西大学 | A kind of construction method of product operational management data model |
| CN109491641B (en) * | 2018-10-25 | 2021-09-17 | 广西大学 | Method for constructing product operation management data model |
| TWI730273B (en) * | 2018-12-17 | 2021-06-11 | 兆豐國際商業銀行股份有限公司 | Potential client searching method and potential client searching device |
| CN112508607A (en) * | 2020-12-02 | 2021-03-16 | 上海中商网络股份有限公司 | Data pushing method and device, computer equipment and readable storage medium |
| CN114742569A (en) * | 2021-01-08 | 2022-07-12 | 广州视源电子科技股份有限公司 | User life stage prediction method and device, computer equipment and storage medium |
| CN112819520A (en) * | 2021-01-27 | 2021-05-18 | 北京思特奇信息技术股份有限公司 | Method and system for calculating life cycle of operator customer |
| CN112967102A (en) * | 2021-02-04 | 2021-06-15 | 江苏警官学院 | Method for establishing customer portrait by logistics data |
| CN112989183A (en) * | 2021-02-20 | 2021-06-18 | 湖南视拓信息技术股份有限公司 | Product information recommendation method and device based on life cycle and related equipment |
| CN113297478A (en) * | 2021-04-25 | 2021-08-24 | 上海淇玥信息技术有限公司 | Information pushing method and device based on user life cycle and electronic equipment |
| CN113591018A (en) * | 2021-07-30 | 2021-11-02 | 中国联合网络通信集团有限公司 | Communication client classification management method, system, electronic device and storage medium |
| CN114143772A (en) * | 2021-11-18 | 2022-03-04 | 北京思特奇信息技术股份有限公司 | Method and system for reducing user off-network rate |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106960354A (en) | Method and device is recommended in a kind of precision based on customer life cycle | |
| CN105574159B (en) | A kind of user's portrait method for building up and user's portrait management system based on big data | |
| CN103714130B (en) | Video recommendation system and method | |
| CN103246980B (en) | Information output method and server | |
| WO2017121251A1 (en) | Information push method and device | |
| CN104050187B (en) | Search result methods of exhibiting and system | |
| CN102629360B (en) | A kind of effective dynamic commodity recommend method and commercial product recommending system | |
| CN103902538B (en) | Information recommending apparatus and method based on decision tree | |
| CN107220365A (en) | Accurate commending system and method based on collaborative filtering and correlation rule parallel processing | |
| CN108205766A (en) | Information-pushing method, apparatus and system | |
| US20040138958A1 (en) | Sales prediction using client value represented by three index axes as criteron | |
| CN103810162B (en) | The method and system of recommendation network information | |
| CN104866474A (en) | Personalized data searching method and device | |
| CN104679771A (en) | Individual data searching method and device | |
| CN106570718A (en) | Information releasing method and releasing system | |
| CN109409928A (en) | A kind of material recommended method, device, storage medium, terminal | |
| KR100970335B1 (en) | Method for updating interest keyword of user and system for executing the method | |
| CN109300039A (en) | The method and system of intellectual product recommendation are carried out based on artificial intelligence and big data | |
| CN112418956A (en) | Financial product recommendation method and device | |
| CN103064880B (en) | A kind of methods, devices and systems providing a user with website selection based on search information | |
| CN104778237A (en) | Individual recommending method and system based on key users | |
| CN103186595A (en) | Method and system for recommending audios/videos | |
| CN113469730A (en) | Customer repurchase prediction method and device based on RF-LightGBM fusion model under non-contract scene | |
| CN114663132A (en) | Intelligent marketing method and device based on real-time user image | |
| Zheng et al. | A scalable purchase intention prediction system using extreme gradient boosting machines with browsing content entropy |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170718 |