CN107369052A - User's registration behavior prediction method, apparatus and electronic equipment - Google Patents
User's registration behavior prediction method, apparatus and electronic equipment Download PDFInfo
- Publication number
- CN107369052A CN107369052A CN201710754967.2A CN201710754967A CN107369052A CN 107369052 A CN107369052 A CN 107369052A CN 201710754967 A CN201710754967 A CN 201710754967A CN 107369052 A CN107369052 A CN 107369052A
- Authority
- CN
- China
- Prior art keywords
- user
- excitation
- registration
- training data
- behavior prediction
- 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
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- 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
- G06Q30/0207—Discounts or incentives, e.g. coupons or rebates
-
- 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
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)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the present disclosure, which discloses a kind of user's registration behavior prediction method, apparatus and electronic equipment, methods described, to be included:Obtain user's registration Behavioral training data set;The user's registration Behavioral training data set is trained, obtains user's registration behavior prediction model;Registration behavior prediction is carried out to test user according to the user's registration behavior prediction model.The technical scheme provided by the embodiment of the present disclosure, registration behavior prediction is carried out for users, so as to obtain the targeted user population most possibly registered, and then can whole or selected section user in targeted user population perform and send reward voucher, coupons equal excitation behavior.The technical scheme has more specific aim in terms of registered user is got, and success rate is high, while also reduces and get the cost that user is spent.
Description
Technical field
This disclosure relates to technical field of data processing, and in particular to a kind of user's registration behavior prediction method, apparatus and electricity
Sub- equipment.
Background technology
With the development of Internet technology, increasing businessman or service provider are promoted by internet channels
Products & services, and make every effort to strive for more user's registration accounts on the basis of products & services are promoted, further to buy
Its corresponding product or service.At present many businessmans or service provider's generally use to it is popular it is random send favor information,
The modes of reward voucher or coupons is sent to attract user to come login account, but this mode lack of targeted, success rate is low, needs
The cost to be spent is higher.
The content of the invention
The embodiment of the present disclosure provides a kind of user's registration behavior prediction method, apparatus and electronic equipment.
In a first aspect, a kind of user's registration behavior prediction method is provided in the embodiment of the present disclosure.
Specifically, the user's registration behavior prediction method, including:
Obtain user's registration Behavioral training data set;
The user's registration Behavioral training data set is trained, obtains user's registration behavior prediction model;
Registration behavior prediction is carried out to test user according to the user's registration behavior prediction model.
With reference in a first aspect, the disclosure in the first implementation of first aspect, it is described acquisition user's registration behavior
Training dataset, including:
Excitation user data is obtained, the excitation user data includes excitation registered user's data and excitation non-registered users
Data;
Obtain the characteristic of the excitation user;
Characteristic of the excitation user data with encouraging user is associated, obtains encouraging registered user's training data and swashs
Non-registered users training data is encouraged, forms the user's registration Behavioral training data set.
With reference in a first aspect, the disclosure in the first implementation of first aspect, the association excitation user
Data and the characteristic of excitation user, including:
Obtain excitation user's disclosed history feature data before time of origin is encouraged;
Associate history feature data of the excitation user data with encouraging user.
With reference in a first aspect, the disclosure in the first implementation of first aspect, it is described to user's registration behavior instruct
Practice data set to be trained, obtain user's registration behavior prediction model, including:
Using it is described excitation registered user's training data be used as positive sample, using it is described encourage non-registered users training data as
Negative sample is trained, and obtains the user's registration behavior prediction model.
With reference in a first aspect, the disclosure in the first implementation of first aspect, it is described to user's registration behavior instruct
Practice data set to be trained, obtain user's registration behavior prediction model, including:
Obtain excitation registered user's training data and excitation non-registered users training data;
Excitation registered user's training data and excitation non-registered users training data are quantized;
Determine classification function;
Excitation registered user training data after quantizing is as positive sample, the excitation non-registered users after quantizing
Training data determines the parameter of the classification function, obtains the user's registration behavior prediction model as negative sample, training.
With reference in a first aspect, the disclosure in the first implementation of first aspect, it is described according to the user's registration
Behavior prediction model carries out registration behavior prediction to test user, including:
Obtain test user characteristic data;
The test user characteristic data is inputted to the user's registration behavior prediction model, obtained for testing user
Registration behavior prediction result.
With reference to the first of first aspect or first aspect implementation, the disclosure is in second of realization side of first aspect
In formula, methods described also includes:
It is ranked up for the registration behavior prediction result for testing user;
The test user of predetermined quantity before sequence is taken to perform default excitation behavior.
With reference to the first implementation of first aspect, first aspect or second of implementation of first aspect, at this
It is disclosed in the third implementation of first aspect, methods described also includes:
Random selection test user performs default excitation behavior in the test user of predetermined quantity after sequence.
Second of implementation or first of the first implementation, first aspect with reference to first aspect, first aspect
The third implementation of aspect, in four kind implementation of the disclosure in first aspect, methods described also includes:
Obtain the registration behavior feedback information for the user for performing default excitation behavior;
Obtain the characteristic of the user;
Associate the characteristic of registration behavior feedback information and the user of the test user, as training data plus
Enter the user's registration Behavioral training data set.
Second aspect, the embodiment of the present disclosure provide a kind of user's registration behavior prediction device, and described device includes:
First acquisition module, it is configured as obtaining user's registration Behavioral training data set;
Training module, it is configured as being trained the user's registration Behavioral training data set, obtains user's registration row
For forecast model;
Prediction module, is configured as that to carry out registration behavior pre- to test user according to the user's registration behavior prediction model
Survey.
In the first implementation of second aspect, the acquisition module includes the disclosure:
First acquisition submodule, it is configured as obtaining excitation user data, the excitation user data includes excitation and registered
User data and excitation non-registered users data;
Second acquisition submodule, it is configured as obtaining the characteristic of the excitation user;
Submodule is associated, is configured as associating characteristic of the excitation user data with encouraging user, is encouraged
Registered user's training data and excitation non-registered users training data, form the user's registration Behavioral training data set.
With reference to second aspect, in the first implementation of second aspect, the association submodule includes the disclosure:
Acquiring unit, it is configured as obtaining excitation user's disclosed history feature data before time of origin is encouraged;
Associative cell, it is configured as associating history feature data of the excitation user data with encouraging user.
With reference to second aspect, in the first implementation of second aspect, the training module includes the disclosure:
First training submodule, it is configured as, using excitation registered user's training data as positive sample, by described swashing
Encourage non-registered users training data to be trained as negative sample, obtain the user's registration behavior prediction model.
With reference to second aspect, in the first implementation of second aspect, the training module includes the disclosure:
3rd acquisition submodule, it is configured as obtaining excitation registered user's training data and excitation non-registered users training number
According to;
Quantize submodule, is configured as to excitation registered user's training data and excitation non-registered users training number
According to being quantized;
Determination sub-module, it is configured to determine that classification function;
Second training submodule, the excitation registered user training data after quantizing is configured as positive sample, will
Excitation non-registered users training data after quantizing determines the parameter of the classification function, obtains institute as negative sample, training
State user's registration behavior prediction model.
With reference to second aspect, in the first implementation of second aspect, the prediction module includes the disclosure:
4th acquisition submodule, it is configured as obtaining test user characteristic data;
Submodule is predicted, is configured as inputting the test user characteristic data to the user's registration behavior prediction mould
Type, obtain the registration behavior prediction result for testing user.
With reference to the first of second aspect or second aspect implementation, the disclosure is in second of realization side of second aspect
In formula, described device also includes:
Order module, it is configured as being ranked up for testing the registration behavior prediction result of user;
First execution module, it is configured as taking the test user of predetermined quantity before sequence to perform default excitation behavior.
With reference to the first implementation of second aspect, second aspect or second of implementation of second aspect, this public affairs
It is opened in the third implementation of second aspect, described device also includes:
Second execution module, it is configured as random selection test user in the test user of the predetermined quantity after sequence and performs
Default excitation behavior.
Second of implementation or second of the first implementation, second aspect with reference to second aspect, second aspect
The third implementation of aspect, in the 4th kind of implementation of second aspect, described device also includes the disclosure:
Second acquisition module, it is configured as obtaining the registration behavior feedback information for the user for performing default excitation behavior;
3rd acquisition module, it is configured as obtaining the characteristic of the user;
Relating module, it is configured as associating the registration behavior feedback information and the characteristic of the user of the test user
According to as the training data addition user's registration Behavioral training data set.
The third aspect, the embodiment of the present disclosure provide a kind of electronic equipment, including memory and processor, the memory
User's registration behavior prediction device is supported to perform user's registration behavior prediction in above-mentioned first aspect for storing one or more
The computer instruction of method, the processor are configurable for performing the computer instruction stored in the memory.It is described
User's registration behavior prediction device can also include communication interface, for user's registration behavior prediction device and other equipment or logical
Communication network communicates.
Fourth aspect, the embodiment of the present disclosure provide a kind of computer-readable recording medium, for storing user's registered row
For the computer instruction used in prediction meanss, it is included is for performing user's registration behavior prediction method in above-mentioned first aspect
Computer instruction involved by user's registration behavior prediction device.
The technical scheme that the embodiment of the present disclosure provides can include the following benefits:
Above-mentioned technical proposal, by carrying out registration behavior prediction for users, so as to obtain what is most possibly registered
Targeted user population, and then can whole or selected section user in targeted user population perform and send reward voucher, generation
Gold note equal excitation behavior.The technical scheme has more specific aim in terms of registered user is obtained, and success rate is high, while also reduces and obtain
Take the cost that family is spent.
It should be appreciated that the general description and following detailed description of the above are only exemplary and explanatory, not
The disclosure can be limited.
Brief description of the drawings
With reference to accompanying drawing, by the detailed description of following non-limiting embodiment, the further feature of the disclosure, purpose and excellent
Point will be apparent.In the accompanying drawings:
Fig. 1 shows the flow chart of the user's registration behavior prediction method according to the embodiment of the disclosure one;
Fig. 2 shows the flow chart of the step S101 according to Fig. 1 illustrated embodiments;
Fig. 3 shows the flow chart of the step S203 according to Fig. 2 illustrated embodiments;
Fig. 4 shows the flow chart of the step S102 according to Fig. 1 illustrated embodiments;
Fig. 5 shows the flow chart of the step S103 according to Fig. 1 illustrated embodiments;
Fig. 6 shows the flow chart of the user's registration behavior prediction method according to another embodiment of the disclosure;
Fig. 7 shows the flow chart of the user's registration behavior prediction method according to another embodiment of the disclosure;
Fig. 8 shows the flow chart of the user's registration behavior prediction method according to another embodiment of the disclosure;
Fig. 9 shows the structured flowchart of the user's registration behavior prediction device according to the embodiment of the disclosure one;
Figure 10 shows the structured flowchart of the acquisition module 901 according to Fig. 9 illustrated embodiments;
Figure 11 shows the structured flowchart of the association submodule 1003 according to Figure 10 illustrated embodiments;
Figure 12 shows the structured flowchart of the training module 902 according to Fig. 9 illustrated embodiments;
Figure 13 shows the structured flowchart of the prediction module 903 according to Fig. 9 illustrated embodiments;
Figure 14 shows the structured flowchart of the user's registration behavior prediction device according to another embodiment of the disclosure;
Figure 15 shows the structured flowchart of the user's registration behavior prediction device according to another embodiment of the disclosure;
Figure 16 shows the structured flowchart of the user's registration behavior prediction device according to another embodiment of the disclosure;
Figure 17 shows the structured flowchart of the electronic equipment according to the embodiment of the disclosure one;
Figure 18 is adapted for the computer for realizing the user's registration behavior prediction method according to the embodiment of the disclosure one
The structural representation of system.
Embodiment
Hereinafter, the illustrative embodiments of the disclosure will be described in detail with reference to the attached drawings, so that those skilled in the art can
Easily realize them.In addition, for the sake of clarity, the portion unrelated with description illustrative embodiments is eliminated in the accompanying drawings
Point.
In the disclosure, it should be appreciated that the term of " comprising " or " having " etc. is intended to refer to disclosed in this specification
Feature, numeral, step, behavior, part, part or presence of its combination, and be not intended to exclude other one or more features,
Numeral, step, behavior, part, part or its combination there is a possibility that or be added.
It also should be noted that in the case where not conflicting, the feature in embodiment and embodiment in the disclosure
It can be mutually combined.Describe the disclosure in detail below with reference to the accompanying drawings and in conjunction with the embodiments.
The technical scheme that the embodiment of the present disclosure provides, by carrying out registration behavior prediction for users, so as to obtain
The targeted user population most possibly registered, and then can whole or selected section user in targeted user population perform
Send reward voucher, coupons equal excitation behavior.The technical scheme has more specific aim in terms of registered user is got, and success rate is high,
Also reduce simultaneously and get the cost that user is spent.
Fig. 1 shows the flow chart of the user's registration behavior prediction method according to the embodiment of the disclosure one.As shown in figure 1,
The user's registration behavior prediction method comprises the following steps S101-S103:
In step S101, user's registration Behavioral training data set is obtained;
In step s 102, the user's registration Behavioral training data set is trained, it is pre- obtains user's registration behavior
Survey model;
In step s 103, registration behavior prediction is carried out to test user according to the user's registration behavior prediction model.
Step S101, S102 and S103 will be hereinafter further described respectively.
Step S101
In order to targetedly encourage user, reduce high cost caused by indifference excitation, the present embodiment use for
The method that user's registration behavior is predicted, used to obtain most possible target interested, being most likely to occur registration behavior
Family colony, then whole or selected section user execution transmission reward voucher, coupons etc. swash in the targeted user population again
Behavior is encouraged, can so improve the success rate for striving for registered user, reduces incentive cost.
In this embodiment, it is necessary first to obtain user's registration Behavioral training data set, behavior is registered as training user
The data basis of forecast model.When obtaining user's registration Behavioral training data, excitation behavior user is carried out before can obtaining
Information finally whether is registered, as user's registration Behavioral training data, certainly in actual applications, can also be obtained for note
Volume other helpful data of behavior prediction are as user's registration Behavioral training data.It should be noted that the disclosure for
Family registration Behavioral training data particular content be not construed as limiting, it is all for registration the helpful data of behavior prediction can conduct
Training data, also each fall within the protection domain of the disclosure.
Step S102
Then the user's registration Behavioral training data set of acquisition is trained, obtains user's registration behavior prediction model.
Wherein, the prediction result of the user's registration behavior prediction model can characterize a certain is not carried out action line at present
For user after being activated, the final possibility for performing registration behavior.User's registration possibility prediction result, business are obtained
Family or service provider's can targetedly, for the very possible user for performing registration behavior implement excitation behavior,
The success rate for striving for registered user had so both been improved, has reduced incentive cost again.
Step S103
After training obtains user's registration behavior prediction model, it is possible to carry out the pre- of registration behavior for test user
Survey, wherein, the test user can be unexcited user.
Above-described embodiment for users by carrying out registration behavior prediction, so as to obtain the target most possibly registered
User group, and then can whole or selected section user in targeted user population perform and send reward voucher, coupons
Equal excitation behavior.The technical scheme has more specific aim in terms of registered user is obtained, and success rate is high, while also reduces acquisition and use
The cost that family is spent.
In an optional implementation of the present embodiment, as shown in Fig. 2 the step S101, that is, obtain user's registration
The step of Behavioral training data set, including step S201-S203:
In step s 201, obtain excitation user data, it is described excitation user data include excitation registered user's data and
Encourage non-registered users data;
In step S202, the characteristic of the excitation user is obtained;
In step S203, characteristic of the excitation user data with encouraging user is associated, excitation registration is obtained and uses
Family training data and excitation non-registered users training data, form the user's registration Behavioral training data set.
Wherein, the characteristic includes:Name, sex, phone number, the age, industry, division of life span, Long-term Interest,
One or more in zone of action.
In the implementation, the excitation user data in historical user information record is first obtained, that is, is carried out overdriving
The user data of behavior, the excitation user data include:Encourage registered user's data and excitation non-registered users data, i.e. quilt
Implement overdriving behavior and finally perform the user data of registration behavior and be carried out overdriving behavior but be not carried out registering
The user data of behavior.In actual applications, these data may be a phone number or only include very limited amount of
User profile.In order to more fully embody the interest of user, hobby, more accurately it is predicted for registration behavior, at this
In implementation, it is also necessary to the characteristic of the excitation user is obtained, wherein, the characteristic includes:Name, sex,
One or more in phone number, age, industry, division of life span, Long-term Interest, zone of action.The spy of the excitation user
The acquisition of sign data can use various ways, for example can be accumulated from other modules or other applications of same application
Obtained in tired user characteristic data, naturally it is also possible to using other acquisition modes, such as social investigation etc..For example,
Company A exploitations have multiple applications, and company A is to user in order to preferably be managed, its row to user in different application
To be integrated and having been modeled, the user feature database for covering user's various aspects behavior is formd, then for company A's
For subsidiary or cooperative venture, its data interchange between company A is legal and more convenient, thus can be from company A
The characteristic of the excitation user is obtained in user feature database, because company's A serial applications are at home with higher
Coverage, therefore the information that company party A-subscriber characteristic place includes has reference value very much.User data is encouraged obtaining
After the characteristic of corresponding excitation user, both are associated according to the data characteristics that two kinds of corresponding data are jointly comprised
Get up, form new data, the multiple new data obtained here just constitute the user's registration Behavioral training data set.
In an optional implementation of the present embodiment, as shown in figure 3, the step S203, i.e., described in described association
The step of encouraging user data and the characteristic of excitation user, including step S301-S302:
In step S301, excitation user's disclosed history feature data before time of origin is encouraged are obtained;
In step s 302, history feature data of the excitation user data with encouraging user are associated.
In above-mentioned implementation, the excitation user data is gone through with being encouraged before excitation time of origin disclosed in user
History characteristic is associated.In the implementation, it is contemplated that excitation user data and user characteristic data may count each other
According to source, such as, it is mentioned above, described in company A subsidiary or cooperative venture can obtain from company's party A-subscriber's property data base
The characteristic of user is encouraged, and user data caused by company A subsidiary or cooperative venture can also turn into company party A-subscriber
The property data base sample data in company party A-subscriber database in other words, that is to say, that company A subsidiary or cooperation are public
Department and company A data sources each other, form the data loop of complexity, in this case, if by the excitation user data with most
New but also undocumented user characteristic data is associated, then due to the mutual reference of data source, user disclosed in next phase is special
Time of disclosure phase will be quoted to the user data obtained between next time of disclosure phase in sign data, so will result in
Feature leakage in a way.Therefore, in this embodiment, in order to avoid there is features described above leakage, by the excitation
User data before excitation time of origin with encouraging history feature data disclosed in user associated.
In an optional implementation of the present embodiment, the step S102, i.e., to the user's registration Behavioral training
Data set is trained, the step of obtaining user's registration behavior prediction model, including:
Using it is described excitation registered user's training data be used as positive sample, using it is described encourage non-registered users training data as
Negative sample is trained, and obtains the user's registration behavior prediction model.
Further, in an optional implementation of the present embodiment, as shown in figure 4, the step S102, i.e., described
User's registration Behavioral training data set is trained, obtains user's registration behavior prediction model, including step S401-S404:
In step S401, excitation registered user's training data and excitation non-registered users training data are obtained;
In step S402, line number is entered to excitation registered user's training data and excitation non-registered users training data
Value;
In step S403, classification function is determined;
In step s 404, the excitation registered user training data after quantizing is as positive sample, after quantizing
Non-registered users training data is encouraged to determine the parameter of the classification function as negative sample, training, obtain the user's registration
Behavior prediction model.
In this embodiment, when training user registers behavior prediction model, excitation registered user's training data is made
For positive sample, non-registered users training data will be encouraged as negative sample.Wherein, the training side of user's registration behavior prediction model
Method can use a variety of training methods, and the disclosure is not especially limited, and all feasible, rational training methods each fall within the disclosure
In protection domain, such as support vector machine method, logistic regression algorithm etc..In actual applications, can be according to the spy of training data
Point and the suitable model of requirement selection and training method for model result.
Numerical value vector type is only supported in view of many sorting algorithms, it is therefore desirable to is quantized firstly for training data
Processing, for example dummy variable coding method can be used, each training data is launched into the feature that multiple values are 0-100, so
The suitable classification function of reselection, the parameter of training determination classification function afterwards, obtain user's registration behavior prediction model.
In actual applications, the quantity of non-registered users is encouraged to be much larger than the quantity of excitation registered user, so
Positive and negative sample proportion serious unbalance is may result in, so as to reduce the predictablity rate of user's registration behavior prediction model, Wu Fazhi
Hold effective user's excitation activity.In this case, the model of directly output sample probability value may be selected, such as logistic regression mould
Type etc..
In addition, mentioned above, the user characteristic data includes name, sex, phone number, age, industry, life rank
Multiple data labels such as section, Long-term Interest, zone of action, in order to more accurately record each data, in one embodiment, often
Item data label includes one or more enumerated values, wherein, the enumerated value is represented with 0-100 numerical value, represents the data mark
Sign the confidence level of the data referred to.In such case, it is contemplated that the sorting algorithm used during model training only supports numerical value
Vectorial type data, it is therefore desirable to above-mentioned data of enumeration type is encoded, such as using dummy variable coding method by each data
The enumerated value of label is transformed to the characteristic vector that value is 0-100, then carries out model training again.
In an optional implementation of the present embodiment, as shown in figure 5, the step S103, i.e., according to the user
The step of registration behavior prediction model carries out registration behavior prediction to test user, including step S501 and step S502:
In step S501, test user characteristic data is obtained;
In step S502, the test user characteristic data is inputted to the user's registration behavior prediction model, obtained
To the registration behavior prediction result for testing user.
In the implementation, test user characteristic data is obtained according to methods mentioned above, wherein, the test is used
Family characteristic includes:One in name, sex, phone number, age, industry, division of life span, Long-term Interest, zone of action
Kind or it is a variety of, then the test user characteristic data is inputted to the user's registration behavior prediction model, you can obtain pair
In the registration behavior prediction result of test user.When the user's registration behavior prediction model is directly output sample probability value
During model, the prediction result be exactly test user occur registration behavior probability have it is much.Based on the prediction result, just
It may determine that the possibility of registration behavior occurs for a certain test user, then select some of which test user to perform excitation
Behavior, this can accomplishes targetedly to implement excitation behavior, while can also improve the success rate of excitation behavior.
In an optional implementation of the present embodiment, as shown in fig. 6, methods described also includes step S601 and step
S602:
In step s 601, it is ranked up for testing the registration behavior prediction result of user;
In step S602, the default excitation behavior of test user execution of predetermined quantity before sequence is taken.
, can be according to the registration behavior prediction result of test user, depending on the needs of practical application, selection in the implementation
All or part of test user performs default excitation behavior, wherein, the default excitation behavior includes:Transmission favor information,
Send the one or more in reward voucher, transmission coupons.For example it can enter for testing the registration behavior prediction result of user
Row sequence, the forward test user of prediction result sequence are considered as the user for being very likely to perform registration behavior, then just
Take top test user to perform default excitation behavior, further improve the success rate of excitation behavior.
In an optional implementation of the present embodiment, as shown in fig. 7, methods described also includes step S701:
In step s 701, random selection test user performs default excitation in the test user of predetermined quantity after sequence
Behavior.
In the disclosure, caused user data will be supplemented to training data concentration after performing excitation behavior each time,
Training data as model training next time uses, and so as to enrich model training data, realizes the Automatic Optimal of training pattern.
But under many circumstances, because the iteration of training pattern each time fully relies on the output of last training pattern, lack with
Machine, thus after many iterations, training pattern is easily ensnared into the state of local optimum.In order to solve this problem,
It is also random to provide while the test user to predetermined quantity in the top performs default excitation behavior in the implementation
A collection of low point of user, i.e., random selection test user performs default action line in the test user for the predetermined quantity that ranks behind
For that so can obtain globally optimal solution with supplemental training model, improve the accuracy rate of user's registration behavior prediction.
In an optional implementation of the present embodiment, as shown in figure 8, methods described also includes step S801-S803:
In step S801, the registration behavior feedback information for the user for performing default excitation behavior is obtained;
In step S802, the characteristic of the user is obtained;
In step S803, the registration behavior feedback information and the characteristic of the user of the user are associated, as
Training data adds the user's registration Behavioral training data set.
In the implementation, in order to obtain more training datas, Behavioral training data set is enriched, improves prediction result
Accuracy, after default excitation behavior is performed to some test users, also obtain the registration behavior feedback letter of these users
Breath, and the registration behavior feedback information of these users is associated with its characteristic, formed described in new training data addition
User's registration Behavioral training data set.
Following is embodiment of the present disclosure, can be used for performing embodiments of the present disclosure.
Fig. 9 shows the structured flowchart of the user's registration behavior prediction device according to the embodiment of the disclosure one, and the device can
With being implemented in combination with as some or all of of electronic equipment by software, hardware or both.As shown in figure 9, the use
Family registration behavior prediction device includes:First acquisition module 901, training module 902 and prediction module 903:
First acquisition module 901, it is configured as obtaining user's registration Behavioral training data set;
Training module 902, it is configured as being trained the user's registration Behavioral training data set, obtains user's registration
Behavior prediction model;
Prediction module 903, it is configured as carrying out registration row to test user according to the user's registration behavior prediction model
For prediction.
Hereinafter the first acquisition module 901, training module 902 and prediction module 903 will further be retouched respectively
State.
First acquisition module 901
In order to targetedly encourage user, reduce high cost caused by indifference excitation, the present embodiment use for
The method that user's registration behavior is predicted, used to obtain most possible target interested, being most likely to occur registration behavior
Family colony, then whole or selected section user execution transmission reward voucher, coupons etc. swash in the targeted user population again
Behavior is encouraged, can so improve the success rate for striving for registered user, reduces incentive cost.
In this embodiment, it is necessary first to user's registration Behavioral training data set is obtained by the first acquisition module 901, made
The data basis of behavior prediction model is registered for training user.When obtaining user's registration Behavioral training data, before can obtaining
It is carried out excitation behavior user and finally whether registers information, as user's registration Behavioral training data, actually should certainly
In, it can also obtain for registration other helpful data of behavior prediction as user's registration Behavioral training data.Need
Bright, the disclosure is not construed as limiting for the particular content of user's registration Behavioral training data, all for registering behavior prediction
Helpful data can be used as training data, also each fall within the protection domain of the disclosure.
Training module 902
Then training module 902 is trained to the user's registration Behavioral training data set of acquisition, obtains user's registration row
For forecast model.
Wherein, the prediction result of the user's registration behavior prediction model can characterize a certain is not carried out action line at present
For user after being activated, the final possibility for performing registration behavior.User's registration possibility prediction result, business are obtained
Family or service provider's can targetedly, for the very possible user for performing registration behavior implement excitation behavior,
The success rate for striving for registered user had so both been improved, has reduced incentive cost again.
Prediction module 903
After the training of training module 902 obtains user's registration behavior prediction model, the can of prediction module 903 is for surveying
Family on probation carries out the prediction of registration behavior, wherein, the test user can be unexcited user.
Above-described embodiment for users by carrying out registration behavior prediction, so as to obtain the target most possibly registered
User group, and then can whole or selected section user in targeted user population perform and send reward voucher, coupons
Equal excitation behavior.The technical scheme has more specific aim in terms of registered user is obtained, and success rate is high, while also reduces acquisition and use
The cost that family is spent.
In an optional implementation of the present embodiment, as shown in Figure 10, the acquisition module 901 includes first and obtained
Submodule 1001, the second acquisition submodule 1002 and associate submodule 1003:
First acquisition submodule 1001, it is configured as obtaining excitation user data, the excitation user data includes excitation
Registered user's data and excitation non-registered users data;
Second acquisition submodule 1002, it is configured as obtaining the characteristic of the excitation user;
Submodule 1003 is associated, is configured as associating characteristic of the excitation user data with encouraging user, obtains
Registered user's training data and excitation non-registered users training data are encouraged, forms the user's registration Behavioral training data set.
Wherein, the characteristic includes:Name, sex, phone number, the age, industry, division of life span, Long-term Interest,
One or more in zone of action.
In the implementation, the first acquisition submodule 1001 first obtains the excitation number of users in historical user information record
According to, that is, the user data of overdriving behavior is carried out, it is described to encourage user data to include:Encourage registered user's data and excitation not
Registered user's data, that is, be carried out overdriving behavior and the final user data for performing registration behavior is gone with overdriving is carried out
For but be not carried out the user data of registration behavior.In actual applications, these data may be a phone number or
Only include very limited amount of user profile.In order to more fully embody the interest of user, hobby, more accurately gone for registration
To be predicted, in the implementation, it is also necessary to the feature of the excitation user is obtained by the second acquisition submodule 1002
Data, wherein, the characteristic includes:Name, sex, phone number, age, industry, division of life span, Long-term Interest, work
One or more in dynamic region.The acquisition of the characteristic of the excitation user can use various ways, such as can be from same
Obtained in the user characteristic data of other modules or other applications accumulation of one application program, naturally it is also possible to using it
His acquisition modes, such as social investigation etc..For example, company A exploitations have multiple applications, and company A is in order to preferably right
User is managed, and it is integrated and modeled to behavior of the user in different application, forms and covers each side of user
The user feature database of face behavior, then for company A subsidiary or cooperative venture, with the number between company A
It is legal and more convenient according to intercommunication, thus the characteristic of the excitation user can be obtained from company's party A-subscriber's property data base
According to because company's A serial applications have higher coverage, therefore the letter that company party A-subscriber characteristic place includes at home
Breath has reference value very much.After the characteristic of excitation user data and corresponding excitation user is obtained, submodule is associated
1003 data characteristicses jointly comprised according to two kinds of corresponding data associate both, form new data, here
Obtained multiple new data just constitute the user's registration Behavioral training data set.
In an optional implementation of the present embodiment, as shown in figure 11, the association submodule 1003 includes:
Acquiring unit 1101, it is configured as obtaining excitation user's disclosed history feature number before time of origin is encouraged
According to;
Associative cell 1102, it is configured as associating history feature data of the excitation user data with encouraging user.
In above-mentioned implementation, associative cell 1102 swashs the excitation user data with what acquiring unit 1101 obtained
History feature data disclosed in user are encouraged to be associated before encouraging time of origin.In the implementation, it is contemplated that excitation user
Data and user characteristic data may data source each other, such as, be mentioned above, company A subsidiary or cooperative venture's meeting
The characteristic of the excitation user, and company A subsidiary or cooperative venture are obtained from company's party A-subscriber's property data base
Caused user data can also turn into sample data of company's party A-subscriber's property data base in other words in company party A-subscriber database,
That is company A subsidiary or cooperative venture and company A data source each other, the data loop of complexity is formed, in this feelings
Under condition, if associative cell 1102 is associated with newest but also undocumented user characteristic data by the excitation user data,
Then due to the mutual reference of data source, time of disclosure phase will be quoted in user characteristic data disclosed in next phase to next
The user data obtained between time of disclosure phase, it so will result in feature leakage in a way.Therefore, in the embodiment party
In formula, in order to avoid there is features described above leakage, associative cell 1102 by the excitation user data and excitation time of origin it
History feature data disclosed in preceding excitation user are associated.
In an optional implementation of the present embodiment, the training module 902 includes:
First training submodule, it is configured as, using excitation registered user's training data as positive sample, by described swashing
Encourage non-registered users training data to be trained as negative sample, obtain the user's registration behavior prediction model.
Further, in an optional implementation of the present embodiment, as shown in figure 12, the training module 902 wraps
Include the 3rd acquisition submodule 1201, the submodule 1202 that quantizes, determination sub-module 1203 and second and train submodule 1204:
3rd acquisition submodule 1201, it is configured as obtaining excitation registered user's training data and excitation non-registered users instruction
Practice data;
Quantize submodule 1202, is configured as to excitation registered user's training data and excitation non-registered users instruction
Practice data to be quantized;
Determination sub-module 1203, is configured to determine that classification function;
Second training submodule 1204, is configured as the excitation registered user training data after quantizing as positive sample
This, the excitation non-registered users training data after quantizing determines the parameter of the classification function, obtained as negative sample, training
To the user's registration behavior prediction model.
In this embodiment, when training user registers behavior prediction model, training submodule will encourage registered user
Training data is as positive sample, using excitation non-registered users training data as negative sample.Wherein, user's registration behavior prediction mould
The training method of type can use a variety of training methods, and the disclosure is not especially limited, and all feasible, rational training methods fall
Enter in the protection domain of the disclosure, such as support vector machine method, logistic regression algorithm etc..In actual applications, can be according to instruction
The characteristics of practicing data and the suitable model of requirement selection and training method for model result.
Numerical value vector type is only supported in view of many sorting algorithms, it is therefore desirable to by the submodule 1202 that quantizes for instruction
Practice data and carry out the processing that quantizes, for example dummy variable coding method can be used, each training data is launched into multiple values
For 0-100 feature, it is then determined that the suitable classification function of the reselection of submodule 1203, the second training submodule 1204 are trained really
Determine the parameter of classification function, obtain user's registration behavior prediction model.
In actual applications, the quantity of non-registered users is encouraged to be much larger than the quantity of excitation registered user, so
Positive and negative sample proportion serious unbalance is may result in, so as to reduce the predictablity rate of user's registration behavior prediction model, Wu Fazhi
Hold effective user's excitation activity.In this case, the model of directly output sample probability value may be selected, such as logistic regression mould
Type etc..
In addition, mentioned above, the user characteristic data includes name, sex, phone number, age, industry, life rank
Multiple data labels such as section, Long-term Interest, zone of action, in order to more accurately record each data, in one embodiment, often
Item data label includes one or more enumerated values, wherein, the enumerated value is represented with 0-100 numerical value, represents the data mark
Sign the confidence level of the data referred to.In such case, it is contemplated that the sorting algorithm used during model training only supports numerical value
Vectorial type data, it is therefore desirable to encode to above-mentioned data of enumeration type, for example encoded using dummy variable by each data label
Enumerated value be transformed to value be 0-100 characteristic vector, then carry out model training again.
In an optional implementation of the present embodiment, as shown in figure 13, the prediction module 903 includes the 4th and obtained
Submodule 1301 and prediction submodule 1302:
4th acquisition submodule 1301, it is configured as obtaining test user characteristic data;
Predict submodule 1302, be configured as by it is described test user characteristic data input it is pre- to the user's registration behavior
Model is surveyed, obtains the registration behavior prediction result for testing user.
In the implementation, the 4th acquisition submodule 1301 obtains test user spy according to embodiment mentioned above
Data are levied, wherein, the test user characteristic data includes:Name, sex, phone number, age, industry, division of life span, length
One or more in phase interest, zone of action, then predict submodule 1302 by it is described test user characteristic data input to
The user's registration behavior prediction model, you can obtain the registration behavior prediction result for testing user.When the user notes
When volume behavior prediction model is directly exports the model of sample probability value, the prediction result is exactly that test user registers
The probability of behavior has much.Based on the prediction result, it is possible to judge that the possibility of registration behavior occurs for a certain test user
Property, then select some of which test user to perform excitation behavior, this can accomplishes targetedly to implement action line
For, while the success rate of excitation behavior can also be improved.
In an optional implementation of the present embodiment, as shown in figure 14, described device also includes order module 1401
With the first execution module 1402:
Order module 1401, it is configured as being ranked up for testing the registration behavior prediction result of user;
First execution module 1402, it is configured as taking the test user of predetermined quantity before sequence to perform default excitation behavior.
In the implementation, the first execution module 1402 can be according to the registration behavior prediction result of test user, depending on reality
The needs of border application, all or part of test user is selected to perform default excitation behavior, wherein, the default excitation behavior bag
Include:Send the one or more in favor information, transmission reward voucher, transmission coupons.For example order module 1401 can be passed through
It is ranked up for the registration behavior prediction result for testing user, the forward test user of prediction result sequence is considered as very
It is possible to perform the user of registration behavior, then the first execution module 1402 just takes top test user to perform default action line
For the further success rate for improving excitation behavior.
In an optional implementation of the present embodiment, as shown in figure 15, described device also includes the second execution module
1501:
Second execution module 1501, it is configured as random selection test user in the test user of the predetermined quantity after sequence
Perform default excitation behavior.
In the disclosure, caused user data will be supplemented to training data concentration after performing excitation behavior each time,
Training data as model training next time uses, and so as to enrich model training data, realizes the Automatic Optimal of training pattern.
But under many circumstances, because the iteration of training pattern each time fully relies on the output of last training pattern, lack with
Machine, thus after many iterations, training pattern is easily ensnared into the state of local optimum.In order to solve this problem,
In the implementation, while the test user to predetermined quantity in the top performs default excitation behavior, second performs mould
Block 1501 also provides a collection of low point of user at random, i.e., the random selection test user in the test user for the predetermined quantity that ranks behind
Default excitation behavior is performed, so globally optimal solution can be obtained with supplemental training model, improve the standard of user's registration behavior prediction
True rate.
In an optional implementation of the present embodiment, as shown in figure 16, described device also includes the second acquisition module
1601st, the 3rd acquisition module 1602 and relating module 1603:
Second acquisition module 1601, it is configured as obtaining the registration behavior feedback letter for the user for performing default excitation behavior
Breath;
3rd acquisition module 1602, it is configured as obtaining the characteristic of the user;
Relating module 1603, it is configured as associating registration behavior feedback information and the spy of the user of the test user
Data are levied, the user's registration Behavioral training data set is added as training data.
In the implementation, in order to obtain more training datas, Behavioral training data set is enriched, improves prediction result
Accuracy, to it is some test users perform default excitation behavior after, the second acquisition module 1601 also obtains these users
Registration behavior feedback information, relating module 1603 is by the registration behavior feedback information and the 3rd acquisition module 1602 of these users
The characteristic of acquisition is associated, and forms new training data and adds the user's registration Behavioral training data set.
Above-mentioned embodiment is illustrated exemplified by predicting user's registration behavior for the disclosure, in fact, above-mentioned reality
The mode of applying can be additionally used in the prediction for other users behavior, and principle is similar to foregoing description, and detail repeats no more.
The disclosure also discloses a kind of electronic equipment, and Figure 17 shows the knot of the electronic equipment according to the embodiment of the disclosure one
Structure block diagram, as shown in figure 17, the electronic equipment 1700 include memory 1701 and processor 1702;Wherein,
The memory 1701 is used to store one or more computer instruction, wherein, one or more computer
Instruction is performed by the processor 1702 to realize:
Obtain user's registration Behavioral training data set;
The user's registration Behavioral training data set is trained, obtains user's registration behavior prediction model;
Registration behavior prediction is carried out to test user according to the user's registration behavior prediction model.
One or more computer instruction can be also performed by the processor 1702 to realize:
The acquisition user's registration Behavioral training data set, including:
Excitation user data is obtained, the excitation user data includes excitation registered user's data and excitation non-registered users
Data;
Obtain the characteristic of the excitation user;
Characteristic of the excitation user data with encouraging user is associated, obtains encouraging registered user's training data and swashs
Non-registered users training data is encouraged, forms the user's registration Behavioral training data set.
Characteristic of the association excitation user data with encouraging user, including:
Obtain excitation user's disclosed history feature data before time of origin is encouraged;
Associate history feature data of the excitation user data with encouraging user.
It is described that user's registration Behavioral training data set is trained, user's registration behavior prediction model is obtained, including:
Using it is described excitation registered user's training data be used as positive sample, using it is described encourage non-registered users training data as
Negative sample is trained, and obtains the user's registration behavior prediction model.
It is described that user's registration Behavioral training data set is trained, user's registration behavior prediction model is obtained, including:
Obtain excitation registered user's training data and excitation non-registered users training data;
Excitation registered user's training data and excitation non-registered users training data are quantized;
Determine classification function;
Excitation registered user training data after quantizing is as positive sample, the excitation non-registered users after quantizing
Training data determines the parameter of the classification function, obtains the user's registration behavior prediction model as negative sample, training.
It is described that registration behavior prediction is carried out to test user according to the user's registration behavior prediction model, including:
Obtain test user characteristic data;
The test user characteristic data is inputted to the user's registration behavior prediction model, obtained for testing user
Registration behavior prediction result.
Also include:
It is ranked up for the registration behavior prediction result for testing user;
The test user of predetermined quantity before sequence is taken to perform default excitation behavior.
Also include:
Random selection test user performs default excitation behavior in the test user of predetermined quantity after sequence.
Also include:
Obtain the registration behavior feedback information for the user for performing default excitation behavior;
Obtain the characteristic of the user;
Associate the characteristic of registration behavior feedback information and the user of the test user, as training data plus
Enter the user's registration Behavioral training data set.
Figure 18 is suitable to be used for the computer system for realizing the user's registration behavior prediction method according to disclosure embodiment
Structural representation.
As shown in figure 18, computer system 1800 includes CPU (CPU) 1801, its can according to be stored in only
Read the program in memory (ROM) 1802 or be loaded into from storage part 1808 in random access storage device (RAM) 1803
Program and perform the various processing in the embodiment shown in above-mentioned Fig. 1-8.In RAM1803, also it is stored with system 1800 and grasps
Various programs and data needed for making.CPU1801, ROM1802 and RAM1803 are connected with each other by bus 1804.Input/defeated
Go out (I/O) interface 1805 and be also connected to bus 1804.
I/O interfaces 1805 are connected to lower component:Importation 1806 including keyboard, mouse etc.;Including such as negative electrode
The output par, c 1807 of ray tube (CRT), liquid crystal display (LCD) etc. and loudspeaker etc.;Storage part including hard disk etc.
1808;And the communications portion 1809 of the NIC including LAN card, modem etc..Communications portion 1809 passes through
Communication process is performed by the network of such as internet.Driver 1810 is also according to needing to be connected to I/O interfaces 1805.It is detachable to be situated between
Matter 1811, such as disk, CD, magneto-optic disk, semiconductor memory etc., it is arranged on as needed on driver 1810, so as to
Storage part 1808 is mounted into as needed in the computer program read from it.
Especially, according to embodiment of the present disclosure, it is soft to may be implemented as computer above with reference to Fig. 1 methods described
Part program.For example, embodiment of the present disclosure includes a kind of computer program product, it includes being tangibly embodied in and its readable
Computer program on medium, the computer program include the journey for the user's registration behavior prediction method for being used to perform Fig. 1-8
Sequence code.In such embodiment, the computer program can be downloaded and pacified from network by communications portion 1809
Dress, and/or be mounted from detachable media 1811.
Flow chart and block diagram in accompanying drawing, it is illustrated that according to the system, method and computer of the various embodiments of the disclosure
Architectural framework in the cards, function and the operation of program product.At this point, each square frame in course diagram or block diagram can be with
A part for a module, program segment or code is represented, a part for the module, program segment or code includes one or more
For realizing the executable instruction of defined logic function.It should also be noted that some as replace realization in, institute in square frame
The function of mark can also be with different from the order marked in accompanying drawing generation.For example, two square frames succeedingly represented are actual
On can perform substantially in parallel, they can also be performed in the opposite order sometimes, and this is depending on involved function.Also
It is noted that the combination of each square frame and block diagram in block diagram and/or flow chart and/or the square frame in flow chart, Ke Yiyong
Function as defined in execution or the special hardware based system of operation are realized, or can be referred to specialized hardware and computer
The combination of order is realized.
Being described in unit or module involved in disclosure embodiment can be realized by way of software, also may be used
Realized in a manner of by hardware.Described unit or module can also be set within a processor, these units or module
Title do not form restriction to the unit or module in itself under certain conditions.
As on the other hand, the disclosure additionally provides a kind of computer-readable recording medium, the computer-readable storage medium
Matter can be the computer-readable recording medium included in device described in above-mentioned embodiment;Can also be individualism,
Without the computer-readable recording medium in supplying equipment.Computer-readable recording medium storage has one or more than one journey
Sequence, described program is used for performing by one or more than one processor is described in disclosed method.
Above description is only the preferred embodiment of the disclosure and the explanation to institute's application technology principle.People in the art
Member should be appreciated that invention scope involved in the disclosure, however it is not limited to the technology that the particular combination of above-mentioned technical characteristic forms
Scheme, while should also cover in the case where not departing from the inventive concept, carried out by above-mentioned technical characteristic or its equivalent feature
The other technical schemes for being combined and being formed.Such as features described above has similar work(with the (but not limited to) disclosed in the disclosure
The technical scheme that the technical characteristic of energy is replaced mutually and formed.
The present disclosure discloses A1, a kind of user's registration behavior prediction method, methods described includes:Obtain user's registration behavior
Training dataset;The user's registration Behavioral training data set is trained, obtains user's registration behavior prediction model;According to
The user's registration behavior prediction model carries out registration behavior prediction to test user.A2, the method according to A1, it is described to obtain
User's registration Behavioral training data set is taken, including:Excitation user data is obtained, the excitation user data includes excitation registration and used
User data and excitation non-registered users data;Obtain the characteristic of the excitation user;Associate it is described excitation user data with
The characteristic of user is encouraged, obtains encouraging registered user's training data and encourages non-registered users training data, described in formation
User's registration Behavioral training data set.A3, the method according to A2, the association excitation user data and excitation user
Characteristic, including:Obtain excitation user's disclosed history feature data before time of origin is encouraged;Associate the excitation
User data and the history feature data of excitation user.A4, the method according to A2, it is described to user's registration Behavioral training number
It is trained according to collection, obtains user's registration behavior prediction model, including:Using excitation registered user's training data as positive sample
This, the excitation non-registered users training data is trained as negative sample, obtains the user's registration behavior prediction mould
Type.A5, the method according to A2, it is described that user's registration Behavioral training data set is trained, obtain user's registration behavior
Forecast model, including:Obtain excitation registered user's training data and excitation non-registered users training data;To the excitation registration
User's training data and excitation non-registered users training data are quantized;Determine classification function;By the excitation after quantizing
Registered user's training data is as positive sample, and the excitation non-registered users training data after quantizing is as negative sample, training
The parameter of the classification function is determined, obtains the user's registration behavior prediction model.A6, the method according to A1, it is described
Registration behavior prediction is carried out to test user according to the user's registration behavior prediction model, including:Obtain test user characteristics
Data;The test user characteristic data is inputted to the user's registration behavior prediction model, obtained for test user's
Register behavior prediction result.A7, the method according to A1, it is characterised in that also include:Registration behavior for testing user
Prediction result is ranked up;The test user of predetermined quantity before sequence is taken to perform default excitation behavior.A8, the side according to A7
Method, in addition to:Random selection test user performs default excitation behavior in the test user of predetermined quantity after sequence.A9, root
According to the method described in A7 or A8, in addition to:Obtain the registration behavior feedback information for the user for performing default excitation behavior;Obtain
The characteristic of the user;The registration behavior feedback information and the characteristic of the user of the test user is associated, is made
The user's registration Behavioral training data set is added for training data.
The present disclosure discloses B10, a kind of user's registration behavior prediction device, described device includes:First acquisition module, quilt
It is configured to obtain user's registration Behavioral training data set;Training module, it is configured as to the user's registration Behavioral training data
Collection is trained, and obtains user's registration behavior prediction model;Prediction module, it is configured as according to the user's registration behavior prediction
Model carries out registration behavior prediction to test user.B11, the device according to B10, the acquisition module include:First obtains
Submodule is taken, is configured as obtaining excitation user data, the excitation user data includes excitation registered user's data and excitation
Non-registered users data;Second acquisition submodule, it is configured as obtaining the characteristic of the excitation user;Associate submodule,
It is configured as associating characteristic of the excitation user data with encouraging user, obtains encouraging registered user's training data and swash
Non-registered users training data is encouraged, forms the user's registration Behavioral training data set.B12, the device according to B11, institute
Stating association submodule includes:Acquiring unit, it is configured as obtaining excitation user disclosed history spy before time of origin is encouraged
Levy data;Associative cell, it is configured as associating history feature data of the excitation user data with encouraging user.B13, basis
Device described in B11, the training module include:First training submodule, it is configured as the excitation registered user training
The excitation non-registered users training data is trained as positive sample, obtains user's note by data as negative sample
Volume behavior prediction model.B14, the device according to B11, the training module include:3rd acquisition submodule, is configured as
Obtain excitation registered user's training data and excitation non-registered users training data;Quantize submodule, is configured as to described
Excitation registered user's training data and excitation non-registered users training data are quantized;Determination sub-module, it is configured as really
Determine classification function;Second training submodule, is configured as the excitation registered user training data after quantizing as positive sample,
Excitation non-registered users training data after quantizing determines the parameter of the classification function, obtained as negative sample, training
The user's registration behavior prediction model.B15, the device according to B10, the prediction module include:4th obtains submodule
Block, it is configured as obtaining test user characteristic data;Submodule is predicted, is configured as the test user characteristic data input
To the user's registration behavior prediction model, the registration behavior prediction result for testing user is obtained.B16, according to B10
Device, in addition to:Order module, it is configured as being ranked up for testing the registration behavior prediction result of user;First holds
Row module, it is configured as taking the test user of predetermined quantity before sequence to perform default excitation behavior.B17, the dress according to B16
Put, in addition to:Second execution module, it is configured as random selection test user in the test user of the predetermined quantity after sequence and holds
The default excitation behavior of row.B18, the device according to B16 or B17, in addition to:Second acquisition module, it is configured as acquisition and holds
Gone default excitation behavior user registration behavior feedback information;3rd acquisition module, it is configured as obtaining the user's
Characteristic;Relating module, it is configured as associating the registration behavior feedback information and the feature of the user of the test user
Data, the user's registration Behavioral training data set is added as training data.
The present disclosure discloses C19, a kind of electronic equipment, including memory and processor;Wherein, the memory is used to deposit
One or more computer instruction is stored up, wherein, one or more computer instruction is by the computing device to realize:Obtain
Take user's registration Behavioral training data set;The user's registration Behavioral training data set is trained, obtains user's registration row
For forecast model;Registration behavior prediction is carried out to test user according to the user's registration behavior prediction model.C20, according to C19
Described electronic equipment, the acquisition user's registration Behavioral training data set, including:Obtain excitation user data, the excitation
User data includes excitation registered user's data and excitation non-registered users data;Obtain the characteristic of the excitation user;
Characteristic of the excitation user data with encouraging user is associated, obtains encouraging registered user's training data and encourages unregistered
User's training data, form the user's registration Behavioral training data set.C21, the electronic equipment according to C20, the pass
Join characteristic of the excitation user data with encouraging user, including:It is public before time of origin is encouraged to obtain excitation user
The history feature data opened;Associate history feature data of the excitation user data with encouraging user.C22, according to C20
Electronic equipment, it is described that user's registration Behavioral training data set is trained, obtain user's registration behavior prediction model, wrap
Include:Using excitation registered user's training data as positive sample, using the excitation non-registered users training data as negative sample
Originally it is trained, obtains the user's registration behavior prediction model.C23, the electronic equipment according to C20, it is described to user
Registration Behavioral training data set is trained, and obtains user's registration behavior prediction model, including:Obtain excitation registered user's training
Data and excitation non-registered users training data;To excitation registered user's training data and excitation non-registered users training number
According to being quantized;Determine classification function;Excitation registered user training data after quantizing will quantize as positive sample
Excitation non-registered users training data afterwards determines the parameter of the classification function, obtains the user as negative sample, training
Register behavior prediction model.C24, the electronic equipment according to C19, it is described according to the user's registration behavior prediction model pair
Test user carries out registration behavior prediction, including:Obtain test user characteristic data;By the test user characteristic data input
To the user's registration behavior prediction model, the registration behavior prediction result for testing user is obtained.C25, according to C19
Electronic equipment, one or more computer instruction is also by the computing device to realize:Note for testing user
Volume behavior prediction result is ranked up;The test user of predetermined quantity before sequence is taken to perform default excitation behavior.C26, according to C25
Described electronic equipment, one or more computer instruction is also by the computing device to realize:Make a reservation for after sequence
Random selection test user performs default excitation behavior in the test user of quantity.C27, the electronics according to C25 or C26 are set
Standby, one or more computer instruction is also by the computing device to realize:Acquisition performs default excitation behavior
The registration behavior feedback information of user;Obtain the characteristic of the user;Associate the registration behavior feedback of the test user
Information and the characteristic of the user, the user's registration Behavioral training data set is added as training data.
The disclosure also discloses D28, a kind of computer-readable recording medium, is stored thereon with computer instruction, the calculating
The method as described in any one of A1-A9 is realized in machine instruction when being executed by processor.
Claims (10)
- A kind of 1. user's registration behavior prediction method, it is characterised in that methods described includes:Obtain user's registration Behavioral training data set;The user's registration Behavioral training data set is trained, obtains user's registration behavior prediction model;Registration behavior prediction is carried out to test user according to the user's registration behavior prediction model.
- 2. according to the method for claim 1, it is characterised in that the acquisition user's registration Behavioral training data set, including:Excitation user data is obtained, the excitation user data includes excitation registered user's data and excitation non-registered users number According to;Obtain the characteristic of the excitation user;Characteristic of the excitation user data with encouraging user is associated, obtains encouraging registered user's training data and excitation not Registered user's training data, form the user's registration Behavioral training data set.
- 3. according to the method for claim 2, it is characterised in that described that user's registration Behavioral training data set is instructed Practice, obtain user's registration behavior prediction model, including:Using excitation registered user's training data as positive sample, using the excitation non-registered users training data as negative sample Originally it is trained, obtains the user's registration behavior prediction model.
- 4. according to the method for claim 2, it is characterised in that described that user's registration Behavioral training data set is instructed Practice, obtain user's registration behavior prediction model, including:Obtain excitation registered user's training data and excitation non-registered users training data;Excitation registered user's training data and excitation non-registered users training data are quantized;Determine classification function;Excitation registered user training data after quantizing is as positive sample, the excitation non-registered users training after quantizing Data determine the parameter of the classification function, obtain the user's registration behavior prediction model as negative sample, training.
- 5. a kind of user's registration behavior prediction device, it is characterised in that described device includes:First acquisition module, it is configured as obtaining user's registration Behavioral training data set;Training module, it is configured as being trained the user's registration Behavioral training data set, it is pre- obtains user's registration behavior Survey model;Prediction module, it is configured as carrying out registration behavior prediction to test user according to the user's registration behavior prediction model.
- 6. a kind of electronic equipment, it is characterised in that including memory and processor;Wherein,The memory is used to store one or more computer instruction, wherein, one or more computer instruction is by institute Computing device is stated to realize:Obtain user's registration Behavioral training data set;The user's registration Behavioral training data set is trained, obtains user's registration behavior prediction model;Registration behavior prediction is carried out to test user according to the user's registration behavior prediction model.
- 7. electronic equipment according to claim 6, it is characterised in that the acquisition user's registration Behavioral training data set, Including:Excitation user data is obtained, the excitation user data includes excitation registered user's data and excitation non-registered users number According to;Obtain the characteristic of the excitation user;Characteristic of the excitation user data with encouraging user is associated, obtains encouraging registered user's training data and excitation not Registered user's training data, form the user's registration Behavioral training data set.
- 8. electronic equipment according to claim 6, it is characterised in that described to be carried out to user's registration Behavioral training data set Training, obtains user's registration behavior prediction model, including:Using excitation registered user's training data as positive sample, using the excitation non-registered users training data as negative sample Originally it is trained, obtains the user's registration behavior prediction model.
- 9. electronic equipment according to claim 6, it is characterised in that described to be carried out to user's registration Behavioral training data set Training, obtains user's registration behavior prediction model, including:Obtain excitation registered user's training data and excitation non-registered users training data;Excitation registered user's training data and excitation non-registered users training data are quantized;Determine classification function;Excitation registered user training data after quantizing is as positive sample, the excitation non-registered users training after quantizing Data determine the parameter of the classification function, obtain the user's registration behavior prediction model as negative sample, training.
- 10. a kind of computer-readable recording medium, is stored thereon with computer instruction, it is characterised in that the computer instruction quilt The method as described in claim any one of 1-4 is realized during computing device.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710754967.2A CN107369052A (en) | 2017-08-29 | 2017-08-29 | User's registration behavior prediction method, apparatus and electronic equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710754967.2A CN107369052A (en) | 2017-08-29 | 2017-08-29 | User's registration behavior prediction method, apparatus and electronic equipment |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107369052A true CN107369052A (en) | 2017-11-21 |
Family
ID=60311753
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710754967.2A Pending CN107369052A (en) | 2017-08-29 | 2017-08-29 | User's registration behavior prediction method, apparatus and electronic equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107369052A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108133390A (en) * | 2017-12-22 | 2018-06-08 | 北京三快在线科技有限公司 | For predicting the method and apparatus of user behavior and computing device |
CN108647990A (en) * | 2018-04-04 | 2018-10-12 | 北京奇艺世纪科技有限公司 | A kind of method, apparatus and electronic equipment of determining target user |
CN108647986A (en) * | 2018-03-28 | 2018-10-12 | 北京奇艺世纪科技有限公司 | A kind of target user determines method, apparatus and electronic equipment |
CN109146562A (en) * | 2018-08-13 | 2019-01-04 | 宜人恒业科技发展(北京)有限公司 | It is a kind of for retrieving the intelligent recommendation system that will be lost user |
CN109345277A (en) * | 2018-08-13 | 2019-02-15 | 宜人恒业科技发展(北京)有限公司 | A kind of intelligent recalling system for registration failure user |
CN111260383A (en) * | 2018-11-30 | 2020-06-09 | 北京嘀嘀无限科技发展有限公司 | Registration probability estimation method and device and probability estimation model construction method and device |
CN111262716A (en) * | 2018-11-30 | 2020-06-09 | 北京嘀嘀无限科技发展有限公司 | Registration probability estimation method and device and probability estimation model construction method and device |
CN111275470A (en) * | 2018-12-04 | 2020-06-12 | 北京嘀嘀无限科技发展有限公司 | Service initiation probability prediction method and training method and device of model thereof |
CN113129080A (en) * | 2021-05-13 | 2021-07-16 | 北京大米科技有限公司 | Data processing method and device |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102075352A (en) * | 2010-12-17 | 2011-05-25 | 北京邮电大学 | Method and device for predicting network user behavior |
CN102663617A (en) * | 2012-03-20 | 2012-09-12 | 亿赞普(北京)科技有限公司 | Method and system for prediction of advertisement clicking rate |
CN103150696A (en) * | 2011-12-06 | 2013-06-12 | 中兴通讯股份有限公司 | Method and device for selecting potential customer of target value-added service |
CN106127566A (en) * | 2016-06-20 | 2016-11-16 | 中国农业银行股份有限公司 | A kind of data managing method and device |
CN106156809A (en) * | 2015-04-24 | 2016-11-23 | 阿里巴巴集团控股有限公司 | For updating the method and device of disaggregated model |
CN106204091A (en) * | 2015-05-28 | 2016-12-07 | 阿里巴巴集团控股有限公司 | Data processing method and device |
-
2017
- 2017-08-29 CN CN201710754967.2A patent/CN107369052A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102075352A (en) * | 2010-12-17 | 2011-05-25 | 北京邮电大学 | Method and device for predicting network user behavior |
CN103150696A (en) * | 2011-12-06 | 2013-06-12 | 中兴通讯股份有限公司 | Method and device for selecting potential customer of target value-added service |
CN102663617A (en) * | 2012-03-20 | 2012-09-12 | 亿赞普(北京)科技有限公司 | Method and system for prediction of advertisement clicking rate |
CN106156809A (en) * | 2015-04-24 | 2016-11-23 | 阿里巴巴集团控股有限公司 | For updating the method and device of disaggregated model |
CN106204091A (en) * | 2015-05-28 | 2016-12-07 | 阿里巴巴集团控股有限公司 | Data processing method and device |
CN106127566A (en) * | 2016-06-20 | 2016-11-16 | 中国农业银行股份有限公司 | A kind of data managing method and device |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108133390A (en) * | 2017-12-22 | 2018-06-08 | 北京三快在线科技有限公司 | For predicting the method and apparatus of user behavior and computing device |
CN108647986A (en) * | 2018-03-28 | 2018-10-12 | 北京奇艺世纪科技有限公司 | A kind of target user determines method, apparatus and electronic equipment |
CN108647986B (en) * | 2018-03-28 | 2021-10-01 | 北京奇艺世纪科技有限公司 | Target user determination method and device and electronic equipment |
CN108647990A (en) * | 2018-04-04 | 2018-10-12 | 北京奇艺世纪科技有限公司 | A kind of method, apparatus and electronic equipment of determining target user |
CN109146562A (en) * | 2018-08-13 | 2019-01-04 | 宜人恒业科技发展(北京)有限公司 | It is a kind of for retrieving the intelligent recommendation system that will be lost user |
CN109345277A (en) * | 2018-08-13 | 2019-02-15 | 宜人恒业科技发展(北京)有限公司 | A kind of intelligent recalling system for registration failure user |
CN111260383A (en) * | 2018-11-30 | 2020-06-09 | 北京嘀嘀无限科技发展有限公司 | Registration probability estimation method and device and probability estimation model construction method and device |
CN111262716A (en) * | 2018-11-30 | 2020-06-09 | 北京嘀嘀无限科技发展有限公司 | Registration probability estimation method and device and probability estimation model construction method and device |
CN111260383B (en) * | 2018-11-30 | 2023-08-29 | 北京嘀嘀无限科技发展有限公司 | Registration probability estimation method and device and probability estimation model construction method and device |
CN111275470A (en) * | 2018-12-04 | 2020-06-12 | 北京嘀嘀无限科技发展有限公司 | Service initiation probability prediction method and training method and device of model thereof |
CN111275470B (en) * | 2018-12-04 | 2023-12-01 | 北京嘀嘀无限科技发展有限公司 | Service initiation probability prediction method and training method and device of model thereof |
CN113129080A (en) * | 2021-05-13 | 2021-07-16 | 北京大米科技有限公司 | Data processing method and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107369052A (en) | User's registration behavior prediction method, apparatus and electronic equipment | |
CN107578294B (en) | User behavior prediction method and device and electronic equipment | |
Chaney et al. | How algorithmic confounding in recommendation systems increases homogeneity and decreases utility | |
JP4806034B2 (en) | Method for calculating user expert index by keyword and system for executing this method | |
CN107784390A (en) | Recognition methods, device, electronic equipment and the storage medium of subscriber lifecycle | |
Wiuf et al. | A likelihood approach to analysis of network data | |
CN115002200B (en) | Message pushing method, device, equipment and storage medium based on user portrait | |
CN109345263A (en) | Predict the method and system of customer satisfaction | |
CN109903095A (en) | Data processing method, device, electronic equipment and computer readable storage medium | |
CN107103036A (en) | Using acquisition methods, equipment and the programmable device for downloading probability | |
CN110196809A (en) | Interface test method and device | |
WO2018069817A1 (en) | System and method for predicting repeat behavior of customers | |
Kontopantelis et al. | Longitudinal multiple imputation approaches for body mass index or other variables with very low individual-level variability: the mibmi command in Stata | |
CN106909560A (en) | Point of interest sort method | |
CN106452809A (en) | Data processing method and device | |
CN107844584A (en) | Usage mining method, apparatus, electronic equipment and computer-readable recording medium | |
CN111127074B (en) | Data recommendation method | |
JP2015118698A (en) | Method, system, computer program for directing communications to nodes of social network using elastic map | |
CN107632971A (en) | Method and apparatus for generating multidimensional form | |
CN110555749B (en) | Credit behavior prediction method and device based on neural network | |
CN110009128A (en) | Industry public opinion index prediction technique, device, computer equipment and storage medium | |
CN107992570A (en) | Character string method for digging, device, electronic equipment and computer-readable recording medium | |
CN108090785B (en) | Method and device for determining user behavior decline tendency and electronic equipment | |
CN112241417A (en) | Verification method, device, medium and electronic equipment of page data | |
Yan et al. | synth2: Synthetic control method with placebo tests, robustness test, and visualization |
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: 20171121 |