CN110400167A - User's screening technique and device - Google Patents
User's screening technique and device Download PDFInfo
- Publication number
- CN110400167A CN110400167A CN201910579620.8A CN201910579620A CN110400167A CN 110400167 A CN110400167 A CN 110400167A CN 201910579620 A CN201910579620 A CN 201910579620A CN 110400167 A CN110400167 A CN 110400167A
- Authority
- CN
- China
- Prior art keywords
- user
- score
- target user
- time
- network model
- 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
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000012216 screening Methods 0.000 title claims abstract description 18
- 230000003542 behavioural effect Effects 0.000 claims abstract description 48
- 210000005036 nerve Anatomy 0.000 claims abstract description 36
- 230000006399 behavior Effects 0.000 claims description 36
- 238000012549 training Methods 0.000 claims description 11
- 238000003062 neural network model Methods 0.000 claims description 6
- 239000013589 supplement Substances 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 4
- 238000010586 diagram Methods 0.000 description 8
- 230000004913 activation Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002618 waking effect Effects 0.000 description 1
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
- G06Q30/0241—Advertisements
- G06Q30/0251—Targeted advertisements
- G06Q30/0254—Targeted advertisements based on statistics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0241—Advertisements
- G06Q30/0251—Targeted advertisements
- G06Q30/0255—Targeted advertisements based on user history
Landscapes
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Game Theory and Decision Science (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Probability & Statistics with Applications (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
This specification embodiment provides a kind of user's screening technique and device, using target user's behavioural characteristic relevant to e-payment as the input of first nerves network model trained in advance, the first score of target user is obtained by the output of first nerves network model, and the first score is used to indicate target user and probability to the usage behavior for the virtual resource provided in advance occurs in the first preset period of time；Using aforementioned behavioural characteristic as the input of nervus opticus network model trained in advance, the second score of target user is obtained by the output of nervus opticus network model, second score is used to indicate after pushing resource using reminding for target user, and the probability to the usage behavior for the virtual resource provided in advance occurs in the second preset period of time for target user；Variation according to the second score relative to the first score, it is determined whether push resource using reminding to target user, the user that can be activated by push resource using prompting can be filtered out.
Description
Technical field
This specification one or more embodiment is related to computer field more particularly to user's screening technique and device.
Background technique
Corresponding some electronic fare payment systems, registration user is very big, and ten million rank, any active ues of every month are at million grades
Not, there is money inside many user balances, and the remaining sum of many users is all that marketing activity obtains, there is no consume
It goes out, this batch of fund is exactly a loss, invests, but do not take back any active ues for company.It is provided by push
Source may remind the user that using resource that, for example, the mode of short massage notice reminds user, the cost of short massage notice is very using prompting
It is low, but daily magnitude is limited, how the important ring for retracting this crowd of user, being in marketing activity of least cost.
Using operation rule, a collection of user is filtered out, carries out short message dispensing, the method for screening user in this way cannot be sufficiently sharp
With the feature of user, some users, if send short messages, can all enliven, some users do not give short message, he does not know that he has this pen
Money would not enliven.And some users, short message has been given, has exactly been interfered, application (app) perhaps can be unloaded.
Accordingly, it would be desirable to there is improved plan, the user that can be activated by push resource using prompting can be filtered out.
Summary of the invention
This specification one or more embodiment describes a kind of user's screening technique and device, can filter out by pushing away
The user for sending resource that can activate using prompting.
In a first aspect, providing a kind of user's screening technique, method includes:
Using target user's behavioural characteristic relevant to e-payment as the defeated of first nerves network model trained in advance
Enter, the first score of the target user is obtained by the output of the first nerves network model, first score is used for
Indicate that the probability to the usage behavior for the virtual resource provided in advance occurs in the first preset period of time for the target user；
Using target user behavioural characteristic relevant to e-payment as nervus opticus network model trained in advance
Input, the second score of the target user, second score are obtained by the output of the nervus opticus network model
It is used to indicate after pushing resource using reminding for the target user, the target user is in the second preset period of time
The probability to the usage behavior for the virtual resource provided in advance occurs；
Variation according to second score relative to first score, it is determined whether Xiang Suoshu target user pushes money
Source uses prompting.
In a kind of possible embodiment, the first nerves network model is trained in the following manner:
Each user in first user set is special as sample in the behavioural characteristic relevant to e-payment of first time
Sign, by each user to the virtual resource provided in advance in first preset period of time after the first time
Usage behavior as sample label, the first nerves network model is trained.
In a kind of possible embodiment, the nervus opticus network model is trained in the following manner:
User each in second user set is special as sample in the behavioural characteristic relevant to e-payment of the second time
Sign, second time are to push resource using the time reminded, by each user described second for each user
It is right to the usage behavior for the virtual resource provided in advance as sample label in second preset period of time after time
The nervus opticus network model is trained.
In a kind of possible embodiment, the behavioural characteristic relevant to e-payment includes at least one of following:
User's transaction count within a certain period of time, user within a certain period of time supplement with money number, user essential information,
User's ties up card information.
In a kind of possible embodiment, the behavioural characteristic relevant to e-payment includes multinomial feature；
The first nerves network model be used for extract respectively the multinomial feature low order feature combination and it is described multinomial
The high-order feature of feature combines, and combines to obtain first score according to low order feature combination and the high-order feature.
In a kind of possible embodiment, the behavioural characteristic relevant to e-payment includes multinomial feature；
The nervus opticus network model be used for extract respectively the multinomial feature low order feature combination and it is described multinomial
The high-order feature of feature combines, and combines to obtain second score according to low order feature combination and the high-order feature.
In a kind of possible embodiment, the push resource includes at least one of following using reminding:
Short message prompting, phone alerts, notification message are reminded.
In a kind of possible embodiment, the variation according to second score relative to first score,
Determine whether that the target user pushes resource and uses prompting, comprising:
Variation by second score relative to first score, is determined as the value added of the target user；
When sequence of the value added in third user set in the value added of each user of the target user be higher than it is default
When ranking, determine that pushing resource to the target user uses prompting.
Second aspect, provides a kind of user's screening plant, and device includes:
First scoring unit, for first using target user's behavioural characteristic relevant to e-payment as training in advance
The input of neural network model obtains the first score of the target user by the output of the first nerves network model,
First score is used to indicate the target user and occurs in the first preset period of time to the virtual resource provided in advance
Usage behavior probability；
Second scoring unit, for using target user behavioural characteristic relevant to e-payment as training in advance
The input of nervus opticus network model, obtain the target user by the output of the nervus opticus network model second obtain
Point, second score is used to indicate after pushing resource using reminding for the target user, and the target user is the
The probability to the usage behavior for the virtual resource provided in advance occurs in two preset period of time；
Determination unit, the second score for being obtained according to the second scoring unit is relative to the first scoring unit
The variation of the first obtained score, it is determined whether Xiang Suoshu target user pushes resource and uses prompting.
The third aspect provides a kind of computer readable storage medium, is stored thereon with computer program, when the calculating
When machine program executes in a computer, enable computer execute first aspect method.
Fourth aspect provides a kind of calculating equipment, including memory and processor, and being stored in the memory can hold
Line code, when the processor executes the executable code, the method for realizing first aspect.
The method and apparatus provided by this specification embodiment, first by target user's behavior relevant to e-payment
Input of the feature as first nerves network model trained in advance, obtains institute by the output of the first nerves network model
The first score of target user is stated, first score is used to indicate the target user and occurs in the first preset period of time
To the probability of the usage behavior for the virtual resource provided in advance；Then target user behavior relevant to e-payment is special
The input as nervus opticus network model trained in advance is levied, is obtained by the output of the nervus opticus network model described
The second score of target user, second score are used to indicate after pushing resource using reminding for the target user,
Probability to the usage behavior for the virtual resource provided in advance occurs in the second preset period of time for the target user；Root again
Variation according to second score relative to first score, it is determined whether Xiang Suoshu target user pushes resource use and mentions
It wakes up, so as to filter out the user that can be activated by push resource using prompting.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, required use in being described below to embodiment
Attached drawing be briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention, for this
For the those of ordinary skill of field, without creative efforts, it can also be obtained according to these attached drawings others
Attached drawing.
Fig. 1 is the implement scene schematic diagram of one embodiment that this specification discloses；
Fig. 2 shows user's screening technique flow charts according to one embodiment；
Fig. 3 shows the acquisition process schematic diagram of aforementioned first score；
Fig. 4 shows the acquisition process schematic diagram of aforementioned second score；
Fig. 5 shows the schematic block diagram of user's screening plant according to one embodiment.
Specific embodiment
With reference to the accompanying drawing, the scheme provided this specification is described.
Fig. 1 is the implement scene schematic diagram of one embodiment that this specification discloses.The implement scene is related to the sieve of user
Choosing.Corresponding some electronic fare payment systems, registration user is very big, and ten million rank, any active ues of every month are in million ranks, very
There is money inside multi-user's remaining sum, and the remaining sum of many users is all that marketing activity obtains, there is no consumption to go out,
This batch of fund is exactly a loss, invests, but do not take back any active ues for company.Made by pushing resource
It may remind the user that with prompting using resource.
This specification embodiment can filter out the user that can be activated by push resource using prompting.
Fig. 2 shows user's screening technique flow chart according to one embodiment, this method from a large number of users for screening
Certain customers out push resource to the certain customers and use prompting.As shown in Fig. 2, user's screening technique the following steps are included:
Step 21, using target user's behavioural characteristic relevant to e-payment as the defeated of first nerves network model trained in advance
Enter, the first score of the target user is obtained by the output of the first nerves network model, first score is used for
Indicate that the probability to the usage behavior for the virtual resource provided in advance occurs in the first preset period of time for the target user；
Step 22, using target user behavioural characteristic relevant to e-payment as nervus opticus network model trained in advance
Input obtains the second score of the target user by the output of the nervus opticus network model, and second score is used
In instruction after pushing resource using reminding for the target user, the target user sends out in the second preset period of time
The probability of the raw usage behavior to the virtual resource provided in advance；Step 23, according to second score relative to described first
The variation of score, it is determined whether Xiang Suoshu target user pushes resource and uses prompting.The specific of above each step is described below
Executive mode.
First in step 21, using target user's behavioural characteristic relevant to e-payment as first nerves trained in advance
The input of network model obtains the first score of the target user by the output of the first nerves network model, described
First score is used to indicate the target user and occurs to make the virtual resource provided in advance in the first preset period of time
With the probability of behavior.It is understood that it is not push resource use to the target user to mention that first score is corresponding
Score under waking up.
In one example, the behavioural characteristic relevant to e-payment includes at least one of following:
User's transaction count within a certain period of time, user within a certain period of time supplement with money number, user essential information,
User's ties up card information.
In one example, the first nerves network model is trained in the following manner:
Each user in first user set is special as sample in the behavioural characteristic relevant to e-payment of first time
Sign, by each user to the use row for the virtual resource provided in advance in the first preset period of time after the first time
To be trained to the first nerves network model as sample label.For example, the sample label can serve to indicate that user
The usage behavior to the virtual resource provided in advance has occurred in the first preset period of time after the first time；Or
Person, the sample label can serve to indicate that user does not occur in the first preset period of time after the first time to pre-
The usage behavior for the virtual resource first provided.
In one example, the behavioural characteristic relevant to e-payment includes multinomial feature；The first nerves net
Network model is used to extract the low order feature combination of the multinomial feature respectively and the high-order feature of the multinomial feature combines, according to
The low order feature combination and the high-order feature combine to obtain first score.
Then in step 22, using target user behavioural characteristic relevant to e-payment as the second of training in advance
The input of neural network model obtains the second score of the target user by the output of the nervus opticus network model,
Second score is used to indicate after pushing resource using reminding for the target user, and the target user is pre- second
If the probability to the usage behavior for the virtual resource provided in advance occurs in the time cycle.It is understood that described first
Dividing corresponding is to push resource to the target user to use the score under reminding.
In one example, the nervus opticus network model is trained in the following manner:
User each in second user set is special as sample in the behavioural characteristic relevant to e-payment of the second time
Sign, second time are that resource is pushed for each user using the time reminded, by each user after second time
The second preset period of time in the usage behavior of the virtual resource provided in advance as sample label, to the nervus opticus
Network model is trained.For example, the sample label can serve to indicate that user is second default after second time
The usage behavior to the virtual resource provided in advance has occurred in time cycle；Alternatively, the sample label can serve to indicate that use
Usage behavior to the virtual resource provided in advance does not occur in the second preset period of time after second time for family.
In one example, the behavioural characteristic relevant to e-payment includes multinomial feature；The nervus opticus net
Network model is used to extract the low order feature combination of the multinomial feature respectively and the high-order feature of the multinomial feature combines, according to
The low order feature combination and the high-order feature combine to obtain second score.
It is understood that the second preset period of time can be identical as the first preset period of time, can also be different.
Variation finally in step 23, according to second score relative to first score, it is determined whether Xiang Suoshu
Target user pushes resource and uses prompting.It is understood that can be according to the variation compared with preset threshold, to determine whether
Resource, which is pushed, to the target user uses prompting；Alternatively, being compared according to the variation with the variation of other users, with determination
Whether resource is pushed to the target user use prompting.
In one example, the push resource includes at least one of following using reminding: short message prompting, phone alerts,
Notification message is reminded.Wherein, notification message prompting can be, but not limited to as the corresponding interior notification message of application of electronic fare payment system.
In one example, the variation by second score relative to first score is determined as the target and uses
The value added at family；When sequence of the value added in third user set in the value added of each user of the target user be higher than it is pre-
If when ranking, determining that pushing resource to the target user uses prompting.
The method provided by this specification embodiment first makees target user's behavioural characteristic relevant to e-payment
For the input of first nerves network model trained in advance, the target is obtained by the output of the first nerves network model
The first score of user, first score are used to indicate the target user and occur in the first preset period of time to preparatory
The probability of the usage behavior of the virtual resource of granting；Then using target user behavioural characteristic relevant to e-payment as
The input of trained nervus opticus network model in advance obtains the target by the output of the nervus opticus network model and uses
Second score at family, second score are used to indicate after pushing resource using reminding for the target user, the mesh
Probability to the usage behavior for the virtual resource provided in advance occurs in the second preset period of time for mark user；Further according to described
Variation of second score relative to first score, it is determined whether Xiang Suoshu target user pushes resource using prompting, thus
The user that can be activated by push resource using prompting can be filtered out.
Fig. 3 shows the acquisition process schematic diagram of aforementioned first score, wherein it is living that first nerves network model can be described as nature
Jump model, and specifically use DeepFM model: 1., firstly, according to the behavioral data of user in history, extract user behavior characteristics 2.
After user's history uplink is, whether can be enlivened in n days, form training label (label) 3. according to user behavior characteristics and instruction
Practice label, form the input format that user's natural model training dataset 4. forms DeepFM model, training pattern obtains nature
Model 5. is enlivened according to the behavioral data of active user, using model is enlivened naturally, estimates the natural recurrence score of user, it should be certainly
So return score, that is, aforementioned first score.
Fig. 4 shows the acquisition process schematic diagram of aforementioned second score, wherein nervus opticus network model can be described as short message and swash
Model living, specifically use DeepFM model: 1. firstly, random short message dispensing, obtains the day that unbiased data 2. are launched according to short message
Phase, n days after 3. short message of user characteristics that gets Date dispensing, whether user is active, determines sample 4. according to user characteristics
With training label, the training dataset 5. for forming short message activation model forms the input format of DeepFM model, and training pattern obtains
Short message activation model 6. is obtained according to the behavioral data collection of active user, model is activated using short message, estimates user's short message and activate
Point, which activates score, that is, aforementioned second score.
Can iris out short message using the aforementioned score and short message activation score of returning naturally and launch crowd: 1. according to naturally active
The natural recurrence score a 2. that model obtains activates model to obtain short message activation score b 3. and obtains short message bring according to short message
Increment score c=b-a 4. sorts according to score c, and sequence is taken to preset the crowd of ranking at preceding (top), carries out short message dispensing.
This specification embodiment can select the crowd that short message can activate, and for whether launching short message, user can live
The user of jump, is filtered.According to two models, the effect that can maximize short message is selected.Whether filtering launches short message, all
The crowd that meeting nature returns.
According to the embodiment of another aspect, a kind of user's screening plant is also provided, the device is for executing this specification reality
User's screening technique of example offer is provided.Fig. 5 shows the schematic block diagram of user's screening plant according to one embodiment.Such as Fig. 5
Shown, which includes:
First scoring unit 51, for using target user's behavioural characteristic relevant to e-payment as trained in advance the
The input of one neural network model, obtain the target user by the output of the first nerves network model first obtain
Point, first score is used to indicate the target user and occurs in the first preset period of time to the virtual money provided in advance
The probability of the usage behavior in source；
Second scoring unit 52, for using target user behavioural characteristic relevant to e-payment as preparatory training
Nervus opticus network model input, the second of the target user is obtained by the output of the nervus opticus network model
Score, second score are used to indicate after pushing resource using reminding for the target user, and the target user exists
The probability to the usage behavior for the virtual resource provided in advance occurs in second preset period of time；
Determination unit 53, the second score for being obtained according to the second scoring unit 52 is relative to first scoring
The variation for the first score that unit 51 obtains, it is determined whether Xiang Suoshu target user pushes resource and uses prompting.
Optionally, as one embodiment, the first nerves network model is trained in the following manner:
Each user in first user set is special as sample in the behavioural characteristic relevant to e-payment of first time
Sign, by each user to the use row for the virtual resource provided in advance in the first preset period of time after the first time
To be trained to the first nerves network model as sample label.
Optionally, as one embodiment, the nervus opticus network model is trained in the following manner:
User each in second user set is special as sample in the behavioural characteristic relevant to e-payment of the second time
Sign, second time are that resource is pushed for each user using the time reminded, by each user after second time
The second preset period of time in the usage behavior of the virtual resource provided in advance as sample label, to the nervus opticus
Network model is trained.
Optionally, as one embodiment, the behavioural characteristic relevant to e-payment includes at least one of following:
User's transaction count within a certain period of time, user within a certain period of time supplement with money number, user essential information,
User's ties up card information.
Optionally, as one embodiment, the behavioural characteristic relevant to e-payment includes multinomial feature；
The first nerves network model be used for extract respectively the multinomial feature low order feature combination and it is described multinomial
The high-order feature of feature combines, and combines to obtain first score according to low order feature combination and the high-order feature.
Optionally, as one embodiment, the behavioural characteristic relevant to e-payment includes multinomial feature；
The nervus opticus network model be used for extract respectively the multinomial feature low order feature combination and it is described multinomial
The high-order feature of feature combines, and combines to obtain second score according to low order feature combination and the high-order feature.
Optionally, as one embodiment, the push resource includes at least one of following using reminding:
Short message prompting, phone alerts, notification message are reminded.
Optionally, as one embodiment, the determination unit 53 is specifically used for:
Variation by second score relative to first score, is determined as the value added of the target user；
When sequence of the value added in third user set in the value added of each user of the target user be higher than it is default
When ranking, determine that pushing resource to the target user uses prompting.
The device provided by this specification embodiment, the first first scoring unit 51 is by target user and e-payment phase
Input of the behavioural characteristic of pass as first nerves network model trained in advance, passes through the defeated of the first nerves network model
The first score of the target user is obtained out, and first score is used to indicate the target user in the first preset time week
The probability to the usage behavior for the virtual resource provided in advance occurs in phase；Then the second scoring unit 52 is by the target user
Input of the behavioural characteristic relevant to e-payment as nervus opticus network model trained in advance, passes through the nervus opticus
The output of network model obtains the second score of the target user, and second score is used to indicate to be used for the target
After family pushes resource using reminding, the target user occurs in the second preset period of time to the virtual resource provided in advance
Usage behavior probability；Last variation of the determination unit 53 further according to second score relative to first score, really
Determine whether to use to target user push resource and remind, can be swashed by push resource using prompting so as to filter out
User living.
According to the embodiment of another aspect, a kind of computer readable storage medium is also provided, is stored thereon with computer journey
Sequence enables computer execute method described in conjunction with Figure 2 when the computer program executes in a computer.
According to the embodiment of another further aspect, a kind of calculating equipment, including memory and processor, the memory are also provided
In be stored with executable code, when the processor executes the executable code, realize method described in conjunction with Figure 2.
Those skilled in the art are it will be appreciated that in said one or multiple examples, function described in the invention
It can be realized with hardware, software, firmware or their any combination.It when implemented in software, can be by these functions
Storage in computer-readable medium or as on computer-readable medium one or more instructions or code transmitted.
Above-described specific embodiment has carried out further the purpose of the present invention, technical scheme and beneficial effects
It is described in detail, it should be understood that being not intended to limit the present invention the foregoing is merely a specific embodiment of the invention
Protection scope, all any modification, equivalent substitution, improvement and etc. on the basis of technical solution of the present invention, done should all
Including within protection scope of the present invention.
Claims (18)
1. a kind of user's screening technique, which comprises
Using target user's behavioural characteristic relevant to e-payment as the input of first nerves network model trained in advance, lead to
The output for crossing the first nerves network model obtains the first score of the target user, and first score is used to indicate institute
It states target user and probability to the usage behavior for the virtual resource provided in advance occurs in the first preset period of time；
Using target user behavioural characteristic relevant to e-payment as the defeated of nervus opticus network model trained in advance
Enter, the second score of the target user is obtained by the output of the nervus opticus network model, second score is used for
After pushing resource using reminding for the target user, the target user occurs in the second preset period of time for instruction
To the probability of the usage behavior for the virtual resource provided in advance；
Variation according to second score relative to first score, it is determined whether Xiang Suoshu target user, which pushes resource, to be made
With prompting.
2. the method for claim 1, wherein the first nerves network model is trained in the following manner:
Using the first user set in each user first time behavioural characteristic relevant to e-payment be used as sample characteristics, general
Each user makes the virtual resource provided in advance in first preset period of time after the first time
It uses behavior as sample label, the first nerves network model is trained.
3. the method for claim 1, wherein the nervus opticus network model is trained in the following manner:
Using user each in second user set the second time behavioural characteristic relevant to e-payment as sample characteristics, institute
Stating for the second time is that resource is pushed for each user using the time reminded, by each user second time it
To the usage behavior for the virtual resource provided in advance as sample label in rear second preset period of time, to described the
Two neural network models are trained.
4. the method for claim 1, wherein the behavioural characteristic relevant to e-payment includes following at least one
:
User's transaction count within a certain period of time, user within a certain period of time supplement number, the essential information of user, user with money
Tie up card information.
5. the method for claim 1, wherein the behavioural characteristic relevant to e-payment includes multinomial feature；
The first nerves network model is used to extract the combination of low order feature and the multinomial feature of the multinomial feature respectively
High-order feature combination, according to the low order feature combination and the high-order feature combine to obtain first score.
6. the method for claim 1, wherein the behavioural characteristic relevant to e-payment includes multinomial feature；
The nervus opticus network model is used to extract the combination of low order feature and the multinomial feature of the multinomial feature respectively
High-order feature combination, according to the low order feature combination and the high-order feature combine to obtain second score.
7. the method for claim 1, wherein the push resource includes at least one of following using reminding:
Short message prompting, phone alerts, notification message are reminded.
8. the method for claim 1, wherein change according to second score relative to first score
Change, it is determined whether Xiang Suoshu target user pushes resource and uses prompting, comprising:
Variation by second score relative to first score, is determined as the value added of the target user；
When sequence of the value added in third user set in the value added of each user of the target user is higher than default ranking
When, determine that pushing resource to the target user uses prompting.
9. a kind of user's screening plant, described device include:
First scoring unit, for using target user's behavioural characteristic relevant to e-payment as first nerves trained in advance
The input of network model obtains the first score of the target user by the output of the first nerves network model, described
First score is used to indicate the target user and occurs to make the virtual resource provided in advance in the first preset period of time
With the probability of behavior；
Second scoring unit, for second using target user behavioural characteristic relevant to e-payment as training in advance
The input of neural network model obtains the second score of the target user by the output of the nervus opticus network model,
Second score is used to indicate after pushing resource using reminding for the target user, and the target user is pre- second
If the probability to the usage behavior for the virtual resource provided in advance occurs in the time cycle；
Determination unit, the second score for being obtained according to the second scoring unit are obtained relative to the first scoring unit
The first score variation, it is determined whether Xiang Suoshu target user push resource using remind.
10. device as claimed in claim 9, wherein the first nerves network model is trained in the following manner:
Using the first user set in each user first time behavioural characteristic relevant to e-payment be used as sample characteristics, general
Each user makes the virtual resource provided in advance in first preset period of time after the first time
It uses behavior as sample label, the first nerves network model is trained.
11. device as claimed in claim 9, wherein the nervus opticus network model is trained in the following manner:
Using user each in second user set the second time behavioural characteristic relevant to e-payment as sample characteristics, institute
Stating for the second time is that resource is pushed for each user using the time reminded, by each user second time it
To the usage behavior for the virtual resource provided in advance as sample label in rear second preset period of time, to described the
Two neural network models are trained.
12. device as claimed in claim 9, wherein the behavioural characteristic relevant to e-payment includes following at least one
:
User's transaction count within a certain period of time, user within a certain period of time supplement number, the essential information of user, user with money
Tie up card information.
13. device as claimed in claim 9, wherein the behavioural characteristic relevant to e-payment includes multinomial feature；
The first nerves network model is used to extract the combination of low order feature and the multinomial feature of the multinomial feature respectively
High-order feature combination, according to the low order feature combination and the high-order feature combine to obtain first score.
14. device as claimed in claim 9, wherein the behavioural characteristic relevant to e-payment includes multinomial feature；
The nervus opticus network model is used to extract the combination of low order feature and the multinomial feature of the multinomial feature respectively
High-order feature combination, according to the low order feature combination and the high-order feature combine to obtain second score.
15. device as claimed in claim 9, wherein the push resource includes at least one of following using reminding:
Short message prompting, phone alerts, notification message are reminded.
16. device as claimed in claim 9, wherein the determination unit is specifically used for:
Variation by second score relative to first score, is determined as the value added of the target user；
When sequence of the value added in third user set in the value added of each user of the target user is higher than default ranking
When, determine that pushing resource to the target user uses prompting.
17. a kind of computer readable storage medium, is stored thereon with computer program, when the computer program in a computer
When execution, computer perform claim is enabled to require the method for any one of 1-8.
18. a kind of calculating equipment, including memory and processor, executable code, the processing are stored in the memory
When device executes the executable code, the method for any one of claim 1-8 is realized.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910579620.8A CN110400167A (en) | 2019-06-28 | 2019-06-28 | User's screening technique and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910579620.8A CN110400167A (en) | 2019-06-28 | 2019-06-28 | User's screening technique and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110400167A true CN110400167A (en) | 2019-11-01 |
Family
ID=68323583
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910579620.8A Pending CN110400167A (en) | 2019-06-28 | 2019-06-28 | User's screening technique and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110400167A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110852755A (en) * | 2019-11-06 | 2020-02-28 | 支付宝(杭州)信息技术有限公司 | User identity identification method and device for transaction scene |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106846041A (en) * | 2016-12-26 | 2017-06-13 | 携程计算机技术（上海）有限公司 | The distribution method and system of reward voucher |
| CN107944593A (en) * | 2017-10-11 | 2018-04-20 | 北京三快在线科技有限公司 | A kind of resource allocation methods and device, electronic equipment |
| CN109509030A (en) * | 2018-11-15 | 2019-03-22 | 北京旷视科技有限公司 | Method for Sales Forecast method and its training method of model, device and electronic system |
| CN109711887A (en) * | 2018-12-28 | 2019-05-03 | 拉扎斯网络科技（上海）有限公司 | Generation method, device, electronic equipment and the computer media of store recommendation list |
-
2019
- 2019-06-28 CN CN201910579620.8A patent/CN110400167A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106846041A (en) * | 2016-12-26 | 2017-06-13 | 携程计算机技术（上海）有限公司 | The distribution method and system of reward voucher |
| CN107944593A (en) * | 2017-10-11 | 2018-04-20 | 北京三快在线科技有限公司 | A kind of resource allocation methods and device, electronic equipment |
| CN109509030A (en) * | 2018-11-15 | 2019-03-22 | 北京旷视科技有限公司 | Method for Sales Forecast method and its training method of model, device and electronic system |
| CN109711887A (en) * | 2018-12-28 | 2019-05-03 | 拉扎斯网络科技（上海）有限公司 | Generation method, device, electronic equipment and the computer media of store recommendation list |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110852755A (en) * | 2019-11-06 | 2020-02-28 | 支付宝(杭州)信息技术有限公司 | User identity identification method and device for transaction scene |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107862782B (en) | Lottery drawing method and device based on Ether house block chain | |
| CN104866484B (en) | A kind of data processing method and device | |
| CN108460629A (en) | User, which markets, recommends method, apparatus, terminal device and storage medium | |
| CN107547214B (en) | Group's reading method, electronic equipment and computer storage medium based on e-book | |
| CN104298755B (en) | Content delivery method, content push system and server | |
| JP5722505B2 (en) | Server, game control method, and program | |
| CN109685559A (en) | Lottery drawing method, device, equipment and storage medium based on big data analysis | |
| CN109685336A (en) | Collection task distribution method, device, computer equipment and storage medium | |
| CN103714133A (en) | Data operation and maintenance management method and device | |
| CN111078742A (en) | User classification model training method, user classification method and device | |
| CN107741986A (en) | User's behavior prediction and corresponding information recommend method and apparatus | |
| CN109493110A (en) | Reward control method and device based on funds transaction | |
| CN110400167A (en) | User's screening technique and device | |
| Lim et al. | App epidemics: Modelling the effects of publicity in a mobile app ecosystem | |
| CN107169786B (en) | Data resource processing method and device | |
| CN113190210A (en) | Random data acquisition method and device based on user task and electronic equipment | |
| CN112308582B (en) | Resource processing method and device, storage medium and electronic device | |
| CN112084044A (en) | Event processing method in system and related device | |
| CN111009299A (en) | Similar medicine recommendation method and system, server and medium | |
| CN109299355B (en) | Recommended book list display method and device and storage medium | |
| CN109544218A (en) | One kind, which is registered, handles method, system and device | |
| CN111309774A (en) | Data processing method and device, electronic equipment and storage medium | |
| CN109858942A (en) | Promotion message methods of exhibiting, device, electronic equipment and readable storage medium storing program for executing | |
| CN114201696A (en) | Message pushing method and device, storage medium and computer equipment | |
| CN110956494B (en) | Virtual article putting method and device, storage medium and electronic device |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200925 Address after: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant after: Innovative advanced technology Co.,Ltd. Address before: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant before: Advanced innovation technology Co.,Ltd. Effective date of registration: 20200925 Address after: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant after: Advanced innovation technology Co.,Ltd. Address before: A four-storey 847 mailbox in Grand Cayman Capital Building, British Cayman Islands Applicant before: Alibaba Group Holding Ltd. |
|
| TA01 | Transfer of patent application right | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191101 |
|
| RJ01 | Rejection of invention patent application after publication |