CN107622326A - User's classification, available resources Forecasting Methodology, device and equipment - Google Patents
User's classification, available resources Forecasting Methodology, device and equipment Download PDFInfo
- Publication number
- CN107622326A CN107622326A CN201710823670.7A CN201710823670A CN107622326A CN 107622326 A CN107622326 A CN 107622326A CN 201710823670 A CN201710823670 A CN 201710823670A CN 107622326 A CN107622326 A CN 107622326A
- Authority
- CN
- China
- Prior art keywords
- user
- cluster
- available resources
- characteristic
- distribution situation
- 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.)
- Granted
Links
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
This specification embodiment, which provides a kind of user's classification, available resources Forecasting Methodology, device and equipment, the user classification method, to be included：Obtain the characteristic of user in available resource management system；The characteristic of each user is used for the variation for reflecting user's available resources；Based on predetermined similarity algorithm, the characteristic of each user is converted to the index for weighing the similarity between each user, and according to the index of conversion gained, user in the available resource management system is classified.Implement this specification embodiment, the similar user of the variation of available resources can be divided into a kind of user.
Description
Technical field
This specification embodiment is related to field of computer technology, more particularly to user's classification, available resources Forecasting Methodology, dress
Put and equipment.
Background technology
With the development of Internet technology, there is provided the system of User Resource Management service is more and more, such as Third-party payment
The available resource management systems such as system, fund service system, treasury management services' system, the storage of available resources is supported, is transferred to or turns
Go out, in actual management available resources, these systems are generally required in forecasting system to lift the service efficiency of available resources
Distribution situation of the available resources in the intended duration in future, such as：In the intended duration in future, in current available resources
Minimum how many resource (minimum available resources) can be stored persistently in system, or in current available resources highest how many
Resource can produce system.Referred to herein as intended duration can be 7 days, 1 month, 3 months, 6 months, 9 months, 12 months etc.
Deng.
Because the variation of the available resources of system user is different, such as：The available resources of certain user remain unchanged for a long period of time, and have
The available resources of a little users frequently change, also the available resources installment of some users.At present in the prediction following expected time of arrival
In limit in system during the distribution situation of available resources, typically using a user in system as a kind of user, to predict every class
Distribution situation of the available resources of user in the intended duration in future, it is then combined with predicting the available resources of all types of user not
The distribution situation in intended duration come, as distribution situation of the available resources in system in the intended duration in future.
But in forecasting system available resources future intended duration in distribution situation when, by one in system
There is relatively large deviation in individual user, such as a kind of user, the result of prediction with actual situation：Minimum available money in long expiration
Source is less than normal, and the minimum available resources in short-term limit are bigger than normal, and these deviations cause system to be difficult to according to prediction result, and effectively lifting can
With the service efficiency of resource.Therefore, how Rational Classification is carried out to user in system, is technical problem urgently to be resolved hurrily.
The content of the invention
In view of this, this specification embodiment provides a kind of user's classification, available resources Forecasting Methodology, device and equipment.
According to the first aspect of this specification embodiment, there is provided a kind of user classification method, including step：
Obtain the characteristic of user in available resource management system；The characteristic of each user is used to reflect the user
The variation of available resources；
Based on predetermined similarity algorithm, the characteristic of each user is converted to the finger for weighing the similarity between each user
Mark, and according to the index of conversion gained, user in the available resource management system is classified.
According to the second aspect of this specification embodiment, there is provided a kind of available resources Forecasting Methodology, including step：
Obtain the characteristic of the user in available resource management system；The characteristic of each user is used to reflect the use
The variation of family available resources；
Based on predetermined similarity algorithm, the characteristic of each user is converted to the finger for weighing the similarity between each user
Mark, and according to the index of conversion gained, the user in the available resource management system is classified；
Distribution situation of the available resources based on each user in past intended duration, predict the available money per class user
Distribution situation of the source in the intended duration in future；
Distribution situation of the available resources based on every class user in the intended duration in future, the available money in forecasting system
Distribution situation of the source in the intended duration in future.
According to the third aspect of this specification embodiment, there is provided a kind of user's sorter, including：
Feature acquisition module, for obtaining the characteristic of user in available resource management system；The feature of each user
Data are used for the variation for reflecting user's available resources；
User's sort module, for based on predetermined similarity algorithm, it is each that the characteristic of each user to be converted into measurement
The index of similarity between user, and according to the index of conversion gained, user in the available resource management system is divided
Class.
According to the fourth aspect of this specification embodiment, there is provided a kind of available resources prediction meanss, including：
Feature acquisition module, for obtaining the characteristic of the user in available resource management system；The spy of each user
Sign data are used for the variation for reflecting user's available resources；
User's sort module, for based on predetermined similarity algorithm, it is each that the characteristic of each user to be converted into measurement
The index of similarity between user, and according to the index of conversion gained, the user in the available resource management system is carried out
Classification；
User resources prediction module, for distribution shape of the available resources based on each user in past intended duration
Condition, predict distribution situation of the available resources per class user in the intended duration in future；
System resource prediction module, for distribution shape of the available resources based on every class user in the intended duration in future
Condition, distribution situation of the available resources in the intended duration in future in forecasting system.
According to the 5th of this specification embodiment the aspect, there is provided a kind of computer equipment, including：
Processor；
Store the memory of processor-executable instruction；
Wherein, the processor is coupled in the memory, for reading the programmed instruction of the memory storage, and makees
For response, following operation is performed：
Obtain the characteristic of user in available resource management system；The characteristic of each user is used to reflect the user
The variation of available resources；
Based on predetermined similarity algorithm, the characteristic of each user is converted to the finger for weighing the similarity between each user
Mark, and according to the index of conversion gained, user in the available resource management system is classified.
According to the 6th of this specification embodiment the aspect, there is provided a kind of computer equipment, including：
Processor；
Store the memory of processor-executable instruction；
Wherein, the processor is coupled in the memory, for reading the programmed instruction of the memory storage, and makees
For response, following operation is performed：
Obtain the characteristic of the user in available resource management system；The characteristic of each user is used to reflect the use
The variation of family available resources；
Based on predetermined similarity algorithm, the characteristic of each user is converted to the finger for weighing the similarity between each user
Mark, and according to the index of conversion gained, the user in the available resource management system is classified；
Distribution situation of the available resources based on each user in past intended duration, predict the available money per class user
Distribution situation of the source in the intended duration in future；
Distribution situation of the available resources based on every class user in the intended duration in future, the available money in forecasting system
Distribution situation of the source in the intended duration in future.
Implement the embodiment that this specification provides, by reflecting that the characteristic of variation of user's available resources obtains
The index of the similarity between each user is weighed, then user in the available resource management system is carried out according to the index of gained
Classification.Accordingly, it is capable to the similar user of the variation of available resources is divided into a kind of user, and then according to each classification after classification
Each user distribution situation of the available resources in past intended duration, predict the available resources per class user in future
Distribution situation in intended duration, then distribution situation of the available resources in the intended duration in future in forecasting system.Due to
The available resources variation of fellow users is similar, so obtain prediction result and reality situation deviation it is smaller, be easy to more
Effectively system resource is effectively utilized, and all improved a lot without manual intervention, forecasting efficiency and accuracy rate.
Brief description of the drawings
Fig. 1 a are the flow charts of the user classification method shown in the exemplary embodiment of this specification one；
Fig. 1 b are the classification schematic diagrames shown in the exemplary embodiment of this specification one；
Fig. 2 is the flow chart of the available resources Forecasting Methodology shown in this specification another exemplary embodiment；
Fig. 3 is the logic diagram of user's sorter shown in the exemplary embodiment of this specification one；
Fig. 4 is the logic diagram of the available resources prediction meanss shown in the exemplary embodiment of this specification one；
Fig. 5 is the hardware structure diagram of the computer equipment shown in the exemplary embodiment of this specification one.
Embodiment
Here exemplary embodiment will be illustrated in detail, its example is illustrated in the accompanying drawings.Following description is related to
During accompanying drawing, unless otherwise indicated, the same numbers in different accompanying drawings represent same or analogous key element.Following exemplary embodiment
Described in embodiment do not represent all embodiments consistent with this specification embodiment.On the contrary, they are only
The example of the apparatus and method consistent with some aspects being described in detail in such as appended claims, this specification embodiment
Son.
It is only merely for the purpose of description specific embodiment in the term that this specification embodiment uses, and is not intended to be limiting
This specification embodiment." one kind ", " institute of singulative used in this specification embodiment and appended claims
State " and "the" be also intended to including most forms, unless context clearly shows that other implications.It is also understood that make herein
Term "and/or" refers to and any or all may be combined comprising the associated list items purpose of one or more.
It will be appreciated that though various letters may be described using term first, second, third, etc. in this specification embodiment
Breath, but these information should not necessarily be limited by these terms.These terms are only used for same type of information being distinguished from each other out.For example,
In the case where not departing from this specification scope of embodiments, the first information can also be referred to as the second information, similarly, the second letter
Breath can also be referred to as the first information.Depending on linguistic context, word as used in this " if " can be construed to
" ... when " or " when ... " or " in response to determining ".
The available resources that this specification embodiment is related to, the assets in the various electronic accounts of user can be referred to, such as：The
Assets in the electronic accounts of system such as tripartite's payment system, fund service system, treasury management services' system, banking system, at it
In his example, the available resources of other forms can also be referred to, this specification embodiment is without limitation.
In actual management available resources, these systems generally require prediction to lift the service efficiency of available resources
Distribution situation of the available resources in the intended duration in future in system, such as：In the intended duration in future, current is available
Minimum how many resource (minimum available resources) can be stored persistently in system in resource, or highest in current available resources
How many resource can produce system.It is how pre- under introduction below by taking the assets in the electronic account of third-party payment system as an example
Current available resources in examining system, how many minimum resource can be stored persistently in system within a week in future：
Because the variation of assets is different in each electronic account in third-party payment system, in the prediction following week
, can be using an electronic account (representing a user) in system as an electron-like account in interior system during the distribution situation of assets
Family, the minimum amount of assets that assets are daily within a past week in each electronic account is calculated, then more daily assets
Amount, the amount of assets for choosing minimum is the minimum amount of assets in each electronic account following week, and each electronic account is existed
Minimum amount of assets in following week is added, and obtains minimum assets of the third-party payment system within a week in future
Amount.But the changing condition of the assets in each electronic account is different, occurring the minimum probability of amount of assets on the same day
It is relatively low, so, the prediction result obtained by above-mentioned Forecasting Methodology is less than normal compared with actual state, and with third-party payment system
Contained account number increases, and deviation is more obvious, can have a strong impact on the service efficiency of assets in system, and then influences system and provide
Side and the interests of user.
And it is each electronic account in third-party payment system to cause above-mentioned prediction result and actual state reason devious
The variation of interior assets is different, it is difficult to which the variation that user in system (corresponding with electronic account) is divided into assets is identical
Homogeneous asset, the scheme of this specification embodiment proposes solution for how user in system being carried out into Rational Classification.
The embodiment of this specification, in forecasting system before distribution situation of the available resources in the intended duration in future,
The characteristic for the variation that can react user's available resources can be predefined, then by reflecting user's available resources
The characteristic of variation obtains the index for weighing the similarity between each user, is then used according to the index of gained described
User is classified in resource management system.Accordingly, it is capable to the similar user of the variation of available resources is divided into a kind of user,
And then the distribution situation according to the available resources of each user of each classification after classification in past intended duration, predict per class
Distribution situation of the available resources of user in the intended duration in future, then available resources in forecasting system are predetermined in future
Distribution situation in time limit.Because the available resources variation of fellow users is similar, prediction result and reality are so obtained
The deviation of situation is smaller, is easy to more effectively effectively utilize system resource, and without manual intervention, forecasting efficiency
All improved a lot with accuracy rate.User's assorting process of this specification embodiment is elaborated below in conjunction with accompanying drawing, can use money
Predict process in source.
Referring to Fig. 1 a, Fig. 1 a are the flow charts of the user classification method shown in the exemplary embodiment of this specification one, the reality
The method for applying example, step S102-S104 can be included：
Step S102, the characteristic of user in available resource management system is obtained；The characteristic of each user is used for
Reflect the variation of user's available resources.
Step S104, based on predetermined similarity algorithm, the characteristic of each user is converted to and weighed between each user
The index of similarity, and according to the index of conversion gained, user in the available resource management system is classified.
In this specification embodiment, available resource management system refers to the system for providing User Resource Management service, such as the
The available resource management systems such as tripartite's payment system, fund service system, treasury management services' system, available resources refer to user at these
The resource of the other forms such as the assets in the electronic account of system registry.
Need the similar user of available resources variation being divided into a kind of user in view of this specification embodiment, that
, it is necessary to predefine the characteristic for the variation that can react user's available resources., can be directly in practical application
The characteristic of the variation of available resources is reacted, such as the average and standard deviation of available resources in scheduled time slot, this specification
The designer of scheme can also replace standard deviation with the variance of available resources in scheduled time slot, in other examples, this explanation
The designer of book scheme can also obtain the characteristic of other variations that can directly react available resources, this specification
Embodiment is without limitation.
In addition to directly the characteristic of variation of available resources is reacted, it can also determine that indirect reaction user can use money
The characteristic of the variation in source, such as：Age of user, sex, educational background, constellation, occupation, resident address, accounts information etc.
User property, accounts information can include be the no real-name authentication of account, the real-name authentication grade of account, member's grade of account,
The liveness of account, Account Type etc., by taking Alipay system as an example, Account Type can be divided into Taobao's seller's account, wash in a pan again
Precious C2C seller's account, day cat seller account, Alipay interior employee's account etc..
Different mode classifications can be taken for different characteristics, is exemplified below several：
Mode one：Characteristic is the average and standard deviation of available resources in scheduled time slot, and predetermined similarity algorithm is
K-means clustering algorithms.Referred to herein as scheduled time slot can be one month, three months, half a year or 1 year, by this specification side
The designer of case makes a reservation for according to concrete scene and required precision of prediction.In other examples, K-means can also be selected to cluster
Other clustering algorithms outside algorithm are similarity algorithm.
In practical application, using K-means clustering algorithms by the average and standard deviation of available resources in scheduled time slot, conversion
To weigh the index of the similarity between each user.
During specific conversion, average and standard deviation can be converted to by Euclidean distance between different user by following operation,
To classify to the user in system：
K user is randomly selected from each user, the center as k cluster；K is the integer more than or equal to 2, specifically
Numerical value can be determined by the designer of this specification scheme according to the application scenarios of reality.
Using the characteristic of each user as respective coordinate, and the coordinate based on each user performs following cluster iteration behaviour
Make：
Each user is calculated to the Euclidean distance at the center of each cluster.
Based on the Euclidean distance for calculating gained, each user is divided into the cluster minimum with its Euclidean distance.
Based on division result, the coordinate average value of each user in each cluster is updated to the coordinate at the center of the cluster.
It is criterion functional value by the Coordinate Conversion of the coordinate at the center of each cluster and each user.Wherein, criterion function
For error sum of squares criterion function, each user putting down to the range difference at center in each cluster can be specifically obtained during conversion
Side, criterion function value can be both got by below equation：
Wherein, SSE refers to criterion function value, and i is the numbering of cluster, and k is the number of cluster, CiRefer to ith cluster.
If criterion function value restrains, the cluster iterative operation is terminated, if criterion function value does not restrain, again
Perform the cluster iterative operation.Referred to herein as convergence refer to the gap of criterion function value and some numerical value in preset range
It is interior, or successive ignition operation criterion function value gap in the range of predetermined value.
After the cluster iterative operation is terminated, by the user at the center for being chosen for each cluster and it is divided into this and gathers
User in class completes the classification to user as a kind of user.
Mode two：Characteristic includes the average of available resources and the age of standard deviation and user in scheduled time slot, used
The sex at family, the occupation of user, user educational background and user accounts information at least two user properties, processing average and
The similarity algorithm of standard deviation is K-means clustering algorithms, handles the age of user, the sex of user, the occupation of user, user
Educational background and user accounts information in the similarity algorithm of at least two be decision Tree algorithms.
In practical application, K-means clustering algorithms can be first used by the average and standard of available resources in scheduled time slot
Difference, be converted to the index for weighing the similarity between each user.
During specific conversion, can by following operation, first average and standard deviation are converted between different user it is European away from
From, the user in system is carried out just to classify, then using decision tree, based on the user property of first sorted each user,
To first sorted each user subseries again：
K user is randomly selected from each user, the center as k cluster；K is the integer more than or equal to 2, specifically
Numerical value can be determined by the designer of this specification scheme according to the application scenarios of reality.
Using the characteristic of each user as respective coordinate, and the coordinate based on each user performs following cluster iteration behaviour
Make：
Each user is calculated to the Euclidean distance at the center of each cluster.
Based on the Euclidean distance for calculating gained, each user is divided into the cluster minimum with its Euclidean distance.
Based on division result, the coordinate average value of each user in each cluster is updated to the coordinate at the center of the cluster.
It is criterion functional value by the Coordinate Conversion of the coordinate at the center of each cluster and each user.
If criterion function value restrains, the cluster iterative operation is terminated, if criterion function value does not restrain, again
Perform the cluster iterative operation.
After the cluster iterative operation is terminated, using all users in available resource management system as same sample set
In user.
Following classification iterative operation is performed for every user property of each user in each sample set：
It is testing attribute to select the maximum user property of the information gain of each sample set.
The value identical user of testing attribute in each sample set is divided into same group.
Judge whether the user in each group belongs to same cluster.
When user in any group belongs to same cluster, the user in the group is divided into a kind of user.
When user in any group is not belonging to same cluster, using the user in the group as the use in same sample set
Family, and perform the classification iterative operation.
Wherein, information gain exactly sees a feature, when system has it and not it for feature one by one
Information content be respectively how many, both differences are exactly the information content that this feature is brought to system, i.e. information gain.
Assuming that an attribute A (one in user property) has k different value { a1, a2..., ak, utilize attribute A
Set S is divided into k subset { S1, S2..., SkWherein SjThe sample of attribute A values in set S is contained, while sets sijIt is
Subset SjMiddle cluster is CiSample number (number of users), according to attribute A divide sample comentropy be：
Wherein, pijIt is the probability of the sample that classification is in subset.
It is with the information gain obtained by after attribute A division sample sets S：
Gain (A)=I (s1, s2..., sm)-E(S/A)；
Bigger in view of information gain, the information for selecting attribute A to provide classification is bigger, selects the uncertain journey of A classification
Spend it is smaller, so, the general user property for selecting the information gain of each sample set maximum of choosing is testing attribute during classification, so
The value identical user of testing attribute in each sample set is divided into same group afterwards.
The user in system is divided into by testing attribute it is multigroup after, whole group user can be belonged to same cluster
One group of user is finally divided into a class, and the group of same cluster is not belonging to for whole group user, can be using whole group user as one
Individual sample set, repeat it is above-mentioned with decision tree classification process, until same group of user belongs to same cluster.
One, which is enumerated, below in conjunction with Fig. 1 b uses decision Tree algorithms, it is right based on the user property of first sorted each user
The example of first sorted each user subseries again：
In this example, for ease of example, just only include two clusters after classification (may gather in practical application including dozens of
Class), cluster A and cluster B, user property is respectively the occupation of user, age, educational background, member's grade of account.
For first sorted user, in iterative operation of classifying first, the maximum user property of information gain is user
Occupation, with occupation can carry out first time packet to the user in system.
After being grouped for the first time, the occupation of first group of user is profession, and the occupation of second group of user is white collar, the 3rd group
The occupation of user is worker, and the user of existing cluster A in each group is found after judgement, the user for also having cluster B, therefore, by each group
User performs classification iterative operation, in the second subseries iterative operation, information gain is maximum again respectively as a set
User property be user age, carries out second packet to the user in each set with the age.
After second is grouped, occupation belongs to cluster A for this group of user of profession and age higher than 25, can will
This group of user is divided into a classification, and it is 1 to mark classification, to distinguish different classes of user；Occupation for profession and
This group user of the age below 25 belongs to cluster B, and this group of user can be divided into a classification, and it is 2 to mark classification；
Occupation belongs to cluster A for the one group of user of white collar and age higher than 35, this group of user can be divided into a classification, and mark
It is 3 to remember classification；Occupation belongs to cluster A for the one group of user of worker and age higher than 40, this group of user can be divided into one
Individual classification, and it is 5 to mark classification.Respective existing cluster A user in the user of other groups, also there is cluster B user, therefore,
Using this remaining each group user as a set, third time classification iterative operation need to be performed.
In third time classifies iterative operation, occupation is maximum for the information gain of this set of white collar and age below 35
User property be user account member's grade, third time packet is carried out to the user in the set with member's grade, packet
Afterwards, occupation for white collar, the age below 35 and below member's grade LV8 one group of user belong to cluster A, can be by this group of user
A classification is divided into, and it is 5 to mark classification；Occupation is white collar, the age is below 35 and member's grade is higher than LV8 one group of user
Cluster B is belonged to, this group of user can be divided into a classification, and it is 6 to mark classification.
In addition, the user property that occupation is the information gain maximum of this set of worker and age below 40 is user's
Educational background, third time packet is carried out to the user in the set with educational background, after packet, occupation is white collar, the age is below 40 and academic
One group of user below undergraduate course belongs to cluster A, and this group of user can be divided into a classification, and it is 7 to mark classification；Occupation
Below 40 and educational background this one group of user above section level belongs to cluster B, this group of user can be divided into one for white collar, age
Classification, and it is 8 to mark classification.
From the foregoing, it will be observed that the user property based on first sorted each user can will cluster A and B to user's subseries again
The subscriber segmentation of two classifications is 8 classifications, and the variation of the available resources of same class user is more like after subdivision, is predicting
In system during available resources, prediction result is closer with actual conditions.
In other embodiments, using decision Tree algorithms, based on the user property of first sorted each user, divide first
Each user after class before subseries, can extract a certain proportion of user as classification samples from each cluster, pass through respectively again
Classification samples carry out classification based training to predetermined decision Tree algorithms, then again by the decision Tree algorithms after training in system
User is classified.
Mode three：Characteristic includes the average of available resources and the age of standard deviation and user in scheduled time slot, used
At least two user properties in the sex at family, the occupation of user, the educational background of user and the accounts information of user, predetermined is similar
Degree algorithm is K-means clustering algorithms.For the ease of calculating Euclidean distance, can be assigned in advance for every user property different
Variate-value.
In other embodiments, other similarity algorithms can also be taken.The variation of user's available resources will be reflected
Characteristic be converted to index of similarity, the user in system is classified, this specification embodiment is without limitation.
After class has been divided to user, classification results can be based on, to the available resources of system in the intended duration in future
Distribution situation detected, specifically may refer to Fig. 2, Fig. 2 is the available resources shown in the exemplary embodiment of this specification one
The flow chart of Forecasting Methodology, the method for the embodiment can include step S202-S208：
Step S202, the characteristic of the user in available resource management system is obtained；The characteristic of each user is used
In the variation for reflecting user's available resources.
Step S204, based on predetermined similarity algorithm, the characteristic of each user is converted to and weighed between each user
The index of similarity, and according to the index of conversion gained, the user in the available resource management system is classified.
Step S206, distribution situation of the available resources in past intended duration based on each user, predict and used per class
Distribution situation of the available resources at family in the intended duration in future.
Step S208, distribution situation of the available resources in the intended duration in future based on every class user, forecasting system
In available resources future intended duration in distribution situation.
Technology contents involved by the step S202 to S204 of this specification embodiment, the step S102 to S104 with Fig. 1 a
Involved technology contents are corresponding, will not be repeated here.
For step S206, the distribution situation of the available resources of each user in past intended duration, user can be referred to
Unit time of the available resources in past intended duration total amount, the available resources of user can also be referred to past
The variable quantity of unit time in intended duration, in other examples, other distribution situations can also be referred to, this specification is implemented
Example is without limitation.
In one example, can be by following available resources of the operation based on each user in past intended duration
Distribution situation, predict distribution situation of the available resources per class user in the intended duration in future:
Available resources based on each user in the unit time of past intended duration, obtain per a kind of each user
In the summation of the available resources of identical unit interval, such user is formed in the unit time of past intended duration
Available resources total amount.Referred to herein as unit interval can be set according to the time measurement unit of intended duration, such as the time limit be 7
My god, 30 days etc., the unit interval be 1 day.
Compare the magnitude relationship of available resources total amount of every class user in unit time, and according to comparative result, choose
Minimum available resources total amount characterizes distribution situation of the available resources of such user in the intended duration in future.
, can be with when predicting distribution situation of the available resources per class user in the intended duration in future in practical application
The account of these users in systems will be integrated into a unified account, then calculated per class user as an entirety
The available resources gone out in the unified account calculate each list of gained in the stock number of the unit time of past intended duration
The stock number of position time, select point of available resources of the minimum stock number as such user in the intended duration in future
Cloth amount, to characterize distribution situation.
Illustrate how that the stock number for selecting minimum characterizes distribution situation below in conjunction with instantiation：
Intended duration is 7 days, and the unit interval is day, in past seven days, all users in P classes are whole as one
Body, the account of these users in systems is integrated into a unified account, the unified account is daily in 7 days to be can use
Resource is：100000,50,000,60,000,80,000,90,000,80,000, after comparing, the available resources of such user are characterized not with minimum value 50,000
The distribution situation in intended duration come, i.e., P classes user is minimum has 50,000 assets to retain in systems.
Before prediction, the designer of this specification scheme can be as needed, sets multiple intended durations, is predicting
After the minimum stock number for characterizing the distribution situation of each intended duration, histogram can be drawn with every class user, intuitively table
Show distribution situation of the available resources of such user in each intended duration.
In other examples, other stock numbers can also be calculated according to other demands to characterize the available resources of all types of user
Distribution situation in the intended duration in future, for example, at most how many resource produces etc., this specification embodiment to this not
It is limited.
For step S208, distribution situation of the available resources in the intended duration in future in system, refer generally to all kinds of
The summation of distribution situation of the available resources of user in the intended duration in future.
In one example, can be by following available resources of the operation based on every class user in the intended duration in future
Distribution situation, distribution situation of the available resources in the intended duration in future in forecasting system：
The summation of the available resources total amount of the distribution situation of computational representation all types of user, the available resources in sign system exist
Distribution situation in following intended duration.
In practical application, system does not only need to know point of the available resources in single intended duration when managing resource
Cloth situation, it is also necessary to know available resources in system different maturity periods between abundance, available resources in composition system
Term structure.
Such as：The distribution situation in distribution situation and 30 days in 7 days, it is known that system needs to know 7 days to 30 days this
The distribution situation of section.During calculating, can based on the available resources in system future each intended duration in distribution situation, meter
Abundance of the available resources between adjacent intended duration in calculation system.
If available resources in sign system are in 7 days of future, (three months) (in 1 month), in 90 days in 30 days,
In 180 days (in six months), in 270 days (in 9 months), in 365 days the distribution situation of (1 year) stock number：balceIn 7 days、
balanceIn 30 days、balanceIn 90 days、balanceIn 180 days、balanceIn 270 days、balanceIn 365 days, then the resource in system exists
Abundance between each intended duration can be obtained by below equation：
Balance [more than 1 year]=balanceWithin 365 days
Balance [9-12M]=balanceWithin 270 days-balanceWithin 365 days
Balance [6-9M]=balanceWithin 180 days-balanceWithin 270 days
Balance [3-6M]=balanceWithin 90 days-balanceWithin 180 days
Balance [1-3M]=balanceWithin 30 days-balanceWithin 90 days
Balance [7-30d]=balanceWithin 7 days-balanceWithin 30 days
Balance [0-7d]=balance_total-balance [7-30d]-balance [1-3M]
- balance [3-6M]-balance [6-9M]-balance [9-12M]-balance [more than 1 year]
When characterizing distribution situation with the minimum flow of the available resources in minimum system, the available resources in system are each predetermined
Abundance between time limit is minimum abundance, when implementing the scheme of this specification embodiment, can use R softwares and SQL
Forecasting efficiency can be improved by carrying out calculating processing.
From above-described embodiment, this specification embodiment is successively by directly affecting the average of available resources variation
With variance, the user property of indirect available resources variation, using Clustering Model and decision-tree model, available money can be based on
The general character that source changes carries out user's classification, is to calculate further according to minimum available volume of resources of all types of user in intended duration
The term structure of resource in system.
For the current assets in third-party payment system, using the scheme of this specification embodiment, it can calculate and be
The term structure of total assets in system, then can monthly the term structure of the current surplus of the total assets be given in more new system
Asset management personnel, it is easy to improve the utilization rate of assets.
Corresponding with the embodiment of preceding method, this specification embodiment additionally provides the embodiment of device.
Referring to Fig. 3, Fig. 3 is the logic diagram of user's sorter shown in the exemplary embodiment of this specification one, the dress
Putting to include：Feature acquisition module 310 and user's sort module 320.
Wherein, feature acquisition module 310, for obtaining the characteristic of user in available resource management system；It is each to use
The characteristic at family is used for the variation for reflecting user's available resources；
User's sort module 320, for based on predetermined similarity algorithm, the characteristic of each user to be converted into measurement
The index of similarity between each user, and according to the index of conversion gained, user in the available resource management system is carried out
Classification.
In some examples, the similarity algorithm includes K-means clustering algorithms and/or decision Tree algorithms.
As an example, the characteristic includes the average and standard deviation of available resources in scheduled time slot, user's classification mould
Block 320 can include：
Module is randomly selected, for randomly selecting k user from each user, the center as k cluster；K be more than
Or the integer equal to 2；
Iteration module is clustered, for using the characteristic of each user as respective coordinate, and based on the coordinate of each user
Perform following cluster iterative operation：
Each user is calculated to the Euclidean distance at the center of each cluster；
Based on the Euclidean distance for calculating gained, each user is divided into the cluster minimum with its Euclidean distance；
Based on division result, the coordinate average value of each user in each cluster is updated to the coordinate at the center of the cluster；
It is criterion functional value by the Coordinate Conversion of the coordinate at the center of each cluster and each user；
When criterion function value restrains, the cluster iterative operation is terminated, when criterion function value does not restrain, is performed again
The cluster iterative operation.
As an example, the characteristic also includes following at least two user properties：
The age of user, the sex of user, the occupation of user, the educational background of user, the accounts information of user；
User's sorter of this specification embodiment can also include：
Sample determining module, for using all users in available resource management system as the use in same sample set
Family；
Classification iteration module, for performing following classification iteration for every user property of each user in each sample set
Operation：
It is testing attribute to select the maximum user property of the information gain of each sample set；
The value identical user of testing attribute in each sample set is divided into same group；
Judge whether the user in each group belongs to same cluster；
When user in any group belongs to same cluster, the user in the group is divided into a kind of user；
When user in any group is not belonging to same cluster, using the user in the group as the use in same sample set
Family, and perform the classification iterative operation.
Referring to Fig. 4, Fig. 4 is the logic diagram of the available resources prediction meanss shown in the exemplary embodiment of this specification one,
The device can include：Feature acquisition module 410, user's sort module 420, user resources prediction module 430 and system resource
Prediction module 440.
Wherein, feature acquisition module 410, for obtaining the characteristic of the user in available resource management system；Each
The characteristic of user is used for the variation for reflecting user's available resources；
User's sort module 420, for based on predetermined similarity algorithm, the characteristic of each user to be converted into measurement
The index of similarity between each user, and according to the index of conversion gained, the user in the available resource management system is entered
Row classification；
User resources prediction module 430, for distribution of the available resources based on each user in past intended duration
Situation, predict distribution situation of the available resources per class user in the intended duration in future；
System resource prediction module 440, for point of the available resources based on every class user in the intended duration in future
Cloth situation, distribution situation of the available resources in the intended duration in future in forecasting system.
In some examples, the similarity algorithm includes K-means clustering algorithms and/or decision Tree algorithms.
As an example, the characteristic includes the average and standard deviation of available resources in scheduled time slot, user's classification mould
Block 420 can include：
Module is randomly selected, for randomly selecting k user from each user, the center as k cluster；K be more than
Or the integer equal to 2；
Iteration module is clustered, for using the characteristic of each user as respective coordinate, and based on the coordinate of each user
Perform following cluster iterative operation：
Each user is calculated to the Euclidean distance at the center of each cluster；
Based on the Euclidean distance for calculating gained, each user is divided into the cluster minimum with its Euclidean distance；
Based on division result, the coordinate average value of each user in each cluster is updated to the coordinate at the center of the cluster；
It is criterion functional value by the Coordinate Conversion of the coordinate at the center of each cluster and each user；
It is to terminate the cluster iterative operation in the convergence of criterion function value, when criterion function value does not restrain, performs again
The cluster iterative operation.
As an example, the characteristic also includes following at least two user properties：
The age of user, the sex of user, the occupation of user, the educational background of user, the accounts information of user；
The available resources prediction meanss of the present embodiment also include：
Sample determining module, for it is described cluster iteration module terminate it is described cluster iterative operation after, by available resources
All users in management system are as the user in same sample set；
Classification iteration module, for performing following classification iteration for every user property of each user in each sample set
Operation：
It is testing attribute to select the maximum user property of the information gain of each sample set；
The value identical user of testing attribute in each sample set is divided into same group；
Judge whether the user in each group belongs to same cluster；
When user in any group belongs to same cluster, the user in the group is divided into a kind of user；
When user in any group is not belonging to same cluster, using the user in the group as the use in same sample set
Family, and perform the classification iterative operation.
In other examples, user resources prediction module 430 can include:
Unit resource acquisition module, for the available money based on each user in the unit time of past intended duration
Source, obtain per a kind of each user in the summation of the available resources of identical unit interval, form such user past pre-
Available resources total amount in the unit time periodically limited；
Least resource acquisition module, closed for the size relatively per available resources total amount of the class user in unit time
System, and according to comparative result, the available resources total amount for choosing minimum characterizes intended duration of the available resources in future of such user
Interior distribution situation.
As an example, system resource prediction module 440 can include：
Total resources computing module, for the summation of the available resources total amount of the distribution situation of computational representation all types of user, table
Distribution situation of the available resources in the intended duration in future in sign system.
As an example, the user resources prediction meanss of this specification embodiment can also include：
Time limit prediction module, it is each pre- in future based on the available resources in system for when intended duration has multiple
The periodically distribution situation in limit, abundance of the available resources between adjacent intended duration in computing system.
The function of unit (or module) and the implementation process of effect specifically refer to right in the above method in said apparatus
The implementation process of step is answered, will not be repeated here.
For device embodiment, because it corresponds essentially to embodiment of the method, so related part is real referring to method
Apply the part explanation of example.Device embodiment described above is only schematical, wherein described be used as separating component
The unit or module of explanation can be or may not be physically separate, and the part shown as unit or module can be with
It is or may not be physical location or module, you can with positioned at a place, or multiple network lists can also be distributed to
In member or module.Some or all of module therein can be selected to realize this specification embodiment side according to the actual needs
The purpose of case.Those of ordinary skill in the art are without creative efforts, you can to understand and implement.
The embodiment of this specification embodiment user sorter/user resources prediction meanss can be applied to be set in computer
It is standby upper.It can specifically be realized by computer chip or entity, or be realized by the product with certain function.It is a kind of typical
In realization, computer equipment is computer, and the concrete form of computer can be personal computer, laptop computer, individual
It is any several in digital assistants, media player, tablet PC, internet television, other smart machines or these equipment
The combination of kind equipment.
Device embodiment can be realized by software, can also be realized by way of hardware or software and hardware combining.With
Exemplified by software is realized, as the device on a logical meaning, being will be non-volatile by the processor of computer equipment where it
Property the computer-readable recording medium such as memory in corresponding computer program instructions read what operation in internal memory was formed.From hardware view
Speech, as shown in figure 5, one for computer equipment where this specification embodiment user sorter/user resources prediction meanss
Kind hardware structure diagram, in addition to the processor shown in Fig. 5, internal memory, network interface and nonvolatile memory, embodiment
Computer equipment where middle device can also include other hardware, to this generally according to the actual functional capability of the computer equipment
Repeat no more.
In one embodiment, the memory of computer equipment can store processor executable program instructions；Processor
Following operation for reading the programmed instruction of the memory storage, and as response, can be performed with coupled memory：
Obtain the characteristic of user in available resource management system；The characteristic of each user is used to reflect the user
The variation of available resources；
Based on predetermined similarity algorithm, the characteristic of each user is converted to the finger for weighing the similarity between each user
Mark, and according to the index of conversion gained, user in the available resource management system is classified.
In one embodiment, the memory of computer equipment can store processor executable program instructions；Processor
Following operation for reading the programmed instruction of the memory storage, and as response, can be performed with coupled memory：
Obtain the characteristic of the user in available resource management system；The characteristic of each user is used to reflect the use
The variation of family available resources；
Based on predetermined similarity algorithm, the characteristic of each user is converted to the finger for weighing the similarity between each user
Mark, and according to the index of conversion gained, the user in the available resource management system is classified；
Distribution situation of the available resources based on each user in past intended duration, predict the available money per class user
Distribution situation of the source in the intended duration in future；
Distribution situation of the available resources based on every class user in the intended duration in future, the available money in forecasting system
Distribution situation of the source in the intended duration in future.
In other embodiments, the operation performed by processor may be referred to description related in embodiment of the method above,
It will not be described here.
It is above-mentioned that this specification specific embodiment is described.Other embodiments are in the scope of the appended claims
It is interior.In some cases, the action recorded in detail in the claims or step can be come according to different from the order in embodiment
Perform and still can realize desired result.In addition, the process described in the accompanying drawings not necessarily require show it is specific suitable
Sequence or consecutive order could realize desired result.In some embodiments, multitasking and parallel processing be also can
With or be probably favourable.
The preferred embodiment of this specification is the foregoing is only, it is all in this explanation not to limit this specification
Within the spirit and principle of book, any modification, equivalent substitution and improvements done etc., the model of this specification protection should be included in
Within enclosing.
Claims (24)
1. a kind of user classification method, including step：
Obtain the characteristic of user in available resource management system；The characteristic of each user is used to reflect that the user can use
The variation of resource；
Based on predetermined similarity algorithm, the characteristic of each user is converted to the index for weighing the similarity between each user,
And according to the index of conversion gained, user in the available resource management system is classified.
2. according to the method for claim 1, the similarity algorithm includes K-means clustering algorithms and/or decision tree is calculated
Method.
3. according to the method for claim 2, the characteristic includes the average and standard of available resources in scheduled time slot
Difference, it is described based on predetermined similarity algorithm, the characteristic of each user is converted to the finger for weighing the similarity between each user
Mark, and the index according to obtained by conversion, classify to user in the available resource management system, including：
K user is randomly selected from each user, the center as k cluster；K is the integer more than or equal to 2；
Using the characteristic of each user as respective coordinate, and the coordinate based on each user performs following cluster iterative operation：
Each user is calculated to the Euclidean distance at the center of each cluster；
Based on the Euclidean distance for calculating gained, each user is divided into the cluster minimum with its Euclidean distance；
Based on division result, the coordinate average value of each user in each cluster is updated to the coordinate at the center of the cluster；
It is criterion functional value by the Coordinate Conversion of the coordinate at the center of each cluster and each user；
If criterion function value restrains, the cluster iterative operation is terminated, if criterion function value does not restrain, is performed again
The cluster iterative operation.
4. according to the method for claim 3, the characteristic also includes following at least two user properties：
The age of user, the sex of user, the occupation of user, the educational background of user, the accounts information of user；
After the cluster iterative operation is terminated, methods described is further comprising the steps of：
Using all users in available resource management system as the user in same sample set；
Following classification iterative operation is performed for every user property of each user in each sample set：
It is testing attribute to select the maximum user property of the information gain of each sample set；
The value identical user of testing attribute in each sample set is divided into same group；
Judge whether the user in each group belongs to same cluster；
When user in any group belongs to same cluster, the user in the group is divided into a kind of user；
When user in any group is not belonging to same cluster, using the user in the group as the user in same sample set, and
Perform the classification iterative operation.
5. a kind of available resources Forecasting Methodology, including step：
Obtain the characteristic of the user in available resource management system；The characteristic of each user is used to reflect that the user can
With the variation of resource；
Based on predetermined similarity algorithm, the characteristic of each user is converted to the index for weighing the similarity between each user,
And according to the index of conversion gained, the user in the available resource management system is classified；
Distribution situation of the available resources based on each user in past intended duration, predict and exist per the available resources of class user
Distribution situation in following intended duration；
Distribution situation of the available resources based on every class user in the intended duration in future, the available resources in forecasting system exist
Distribution situation in following intended duration.
6. according to the method for claim 5, the similarity algorithm includes K-means clustering algorithms and/or decision tree is calculated
Method.
7. according to the method for claim 6, the characteristic includes the average and standard of available resources in scheduled time slot
Difference, it is described based on predetermined similarity algorithm, the characteristic of each user is converted to the finger for weighing the similarity between each user
Mark, and the index according to obtained by conversion, classify to user in the available resource management system, including：
K user is randomly selected from each user, the center as k cluster；K is the integer more than or equal to 2；
Using the characteristic of each user as respective coordinate, and the coordinate based on each user performs following cluster iterative operation：
Each user is calculated to the Euclidean distance at the center of each cluster；
Based on the Euclidean distance for calculating gained, each user is divided into the cluster minimum with its Euclidean distance；
Based on division result, the coordinate average value of each user in each cluster is updated to the coordinate at the center of the cluster；
It is criterion functional value by the Coordinate Conversion of the coordinate at the center of each cluster and each user；
If criterion function value restrains, the cluster iterative operation is terminated, if criterion function value does not restrain, is performed again
The cluster iterative operation.
8. according to the method for claim 7, the characteristic also includes following at least two user properties：
The age of user, the sex of user, the occupation of user, the educational background of user, the accounts information of user；
After the cluster iterative operation is terminated, methods described is further comprising the steps of：
Using all users in available resource management system as the user in same sample set；
Following classification iterative operation is performed for every user property of each user in each sample set：
It is testing attribute to select the maximum user property of the information gain of each sample set；
The value identical user of testing attribute in each sample set is divided into same group；
Judge whether the user in each group belongs to same cluster；
When user in any group belongs to same cluster, the user in the group is divided into a kind of user；
When user in any group is not belonging to same cluster, using the user in the group as the user in same sample set, and
Perform the classification iterative operation.
9. the method according to any one of claim 5 to 8, the available resources based on each user are past predetermined
Distribution situation in time limit, distribution situation of the available resources per class user in the intended duration in future is predicted, including:
Available resources based on each user in the unit time of past intended duration, obtain per a kind of each user in phase
The summation of the available resources of same unit interval, it is available in the unit time of past intended duration to form such user
Total resources；
Compare the magnitude relationship of available resources total amount of every class user in unit time, and according to comparative result, choose minimum
Available resources total amount characterize the distribution situation of the available resources of such user in the intended duration in future.
10. according to the method for claim 9, the available resources based on every class user are in the intended duration in future
Distribution situation, distribution situation of the available resources in the intended duration in future in forecasting system, including：
The summation of the available resources total amount of the distribution situation of computational representation all types of user, the available resources in sign system will be in future
Intended duration in distribution situation.
11. according to the method for claim 10, when intended duration has multiple, methods described is further comprising the steps of：
Distribution situation based on the available resources in system in each intended duration in future, the available resources in computing system exist
Abundance between adjacent intended duration.
12. a kind of user's sorter, including：
Feature acquisition module, for obtaining the characteristic of user in available resource management system；The characteristic of each user
For reflecting the variation of user's available resources；
User's sort module, for based on predetermined similarity algorithm, the characteristic of each user being converted to and weighs each user
Between similarity index, and according to conversion gained index, user in the available resource management system is classified.
13. device according to claim 12, the similarity algorithm includes K-means clustering algorithms and/or decision tree
Algorithm.
14. device according to claim 13, the characteristic includes the average and mark of available resources in scheduled time slot
Accurate poor, user's sort module includes：
Module is randomly selected, for randomly selecting k user from each user, the center as k cluster；K is to be more than or wait
In 2 integer；
Iteration module is clustered, is performed for the coordinate using the characteristic of each user as respective coordinate, and based on each user
Following cluster iterative operation：
Each user is calculated to the Euclidean distance at the center of each cluster；
Based on the Euclidean distance for calculating gained, each user is divided into the cluster minimum with its Euclidean distance；
Based on division result, the coordinate average value of each user in each cluster is updated to the coordinate at the center of the cluster；
It is criterion functional value by the Coordinate Conversion of the coordinate at the center of each cluster and each user；
When criterion function value restrains, the cluster iterative operation is terminated, when criterion function value does not restrain, again described in execution
Cluster iterative operation.
15. device according to claim 14, the characteristic also includes following at least two user properties：
The age of user, the sex of user, the occupation of user, the educational background of user, the accounts information of user；
Described device also includes：
Sample determining module, for using all users in available resource management system as the user in same sample set；
Classification iteration module, for performing following classification iteration behaviour for every user property of each user in each sample set
Make：
It is testing attribute to select the maximum user property of the information gain of each sample set；
The value identical user of testing attribute in each sample set is divided into same group；
Judge whether the user in each group belongs to same cluster；
When user in any group belongs to same cluster, the user in the group is divided into a kind of user；
When user in any group is not belonging to same cluster, using the user in the group as the user in same sample set, and
Perform the classification iterative operation.
16. a kind of available resources prediction meanss, including：
Feature acquisition module, for obtaining the characteristic of the user in available resource management system；The characteristic of each user
According to the variation for reflecting user's available resources；
User's sort module, for based on predetermined similarity algorithm, the characteristic of each user being converted to and weighs each user
Between similarity index, and according to conversion gained index, the user in the available resource management system is classified；
User resources prediction module, for distribution situation of the available resources based on each user in past intended duration, in advance
Survey distribution situation of the available resources per class user in the intended duration in future；
System resource prediction module, for the available resources based on every class user future intended duration in distribution situation,
Distribution situation of the available resources in the intended duration in future in forecasting system.
17. device according to claim 16, the similarity algorithm includes K-means clustering algorithms and/or decision tree
Algorithm.
18. device according to claim 17, the characteristic includes the average and mark of available resources in scheduled time slot
Accurate poor, user's sort module includes：
Module is randomly selected, for randomly selecting k user from each user, the center as k cluster；K is to be more than or wait
In 2 integer；
Iteration module is clustered, is performed for the coordinate using the characteristic of each user as respective coordinate, and based on each user
Following cluster iterative operation：
Each user is calculated to the Euclidean distance at the center of each cluster；
Based on the Euclidean distance for calculating gained, each user is divided into the cluster minimum with its Euclidean distance；
Based on division result, the coordinate average value of each user in each cluster is updated to the coordinate at the center of the cluster；
It is criterion functional value by the Coordinate Conversion of the coordinate at the center of each cluster and each user；
It is to terminate the cluster iterative operation in the convergence of criterion function value, when criterion function value does not restrain, again described in execution
Cluster iterative operation.
19. device according to claim 18, the characteristic also includes following at least two user properties：
The age of user, the sex of user, the occupation of user, the educational background of user, the accounts information of user；
Described device also includes：
Sample determining module, for after the cluster iteration module terminates the cluster iterative operation, available resources to be managed
All users in system are as the user in same sample set；
Classification iteration module, for performing following classification iteration behaviour for every user property of each user in each sample set
Make：
It is testing attribute to select the maximum user property of the information gain of each sample set；
The value identical user of testing attribute in each sample set is divided into same group；
Judge whether the user in each group belongs to same cluster；
When user in any group belongs to same cluster, the user in the group is divided into a kind of user；
When user in any group is not belonging to same cluster, using the user in the group as the user in same sample set, and
Perform the classification iterative operation.
20. the device according to any one of claim 16 to 19, the user resources prediction module includes:
Unit resource acquisition module, for the available resources based on each user in the unit time of past intended duration,
Obtain per a kind of each user in the summation of the available resources of identical unit interval, form such user in the past expected time of arrival
Available resources total amount in the unit time of limit；
Least resource acquisition module, for comparing the magnitude relationship per available resources total amount of the class user in unit time,
And according to comparative result, the available resources total amount for choosing minimum characterizes the available resources of such user in the intended duration in future
Distribution situation.
21. device according to claim 20, the system resource prediction module includes：
Total resources computing module, for the summation of the available resources total amount of the distribution situation of computational representation all types of user, characterize system
Distribution situation of the available resources in the intended duration in future in system.
22. device according to claim 21, described device also include：
Time limit prediction module, for when intended duration has multiple, based on the available resources in system future each expected time of arrival
Distribution situation in limit, abundance of the available resources between adjacent intended duration in computing system.
23. a kind of computer equipment, including：
Processor；
Store the memory of processor-executable instruction；
Wherein, the processor is coupled in the memory, for reading the programmed instruction of the memory storage, and as sound
Should, perform following operation：
Obtain the characteristic of user in available resource management system；The characteristic of each user is used to reflect that the user can use
The variation of resource；
Based on predetermined similarity algorithm, the characteristic of each user is converted to the index for weighing the similarity between each user,
And according to the index of conversion gained, user in the available resource management system is classified.
24. a kind of computer equipment, including：
Processor；
Store the memory of processor-executable instruction；
Wherein, the processor is coupled in the memory, for reading the programmed instruction of the memory storage, and as sound
Should, perform following operation：
Obtain the characteristic of the user in available resource management system；The characteristic of each user is used to reflect that the user can
With the variation of resource；
Based on predetermined similarity algorithm, the characteristic of each user is converted to the index for weighing the similarity between each user,
And according to the index of conversion gained, the user in the available resource management system is classified；
Distribution situation of the available resources based on each user in past intended duration, predict and exist per the available resources of class user
Distribution situation in following intended duration；
Distribution situation of the available resources based on every class user in the intended duration in future, the available resources in forecasting system exist
Distribution situation in following intended duration.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710823670.7A CN107622326B (en) | 2017-09-13 | 2017-09-13 | User classification and available resource prediction method, device and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710823670.7A CN107622326B (en) | 2017-09-13 | 2017-09-13 | User classification and available resource prediction method, device and equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107622326A true CN107622326A (en) | 2018-01-23 |
| CN107622326B CN107622326B (en) | 2021-02-09 |
Family
ID=61089852
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710823670.7A Active CN107622326B (en) | 2017-09-13 | 2017-09-13 | User classification and available resource prediction method, device and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107622326B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108564380A (en) * | 2018-04-11 | 2018-09-21 | 重庆大学 | A kind of telecommunication user sorting technique based on iteration decision tree |
| CN109857816A (en) * | 2019-01-11 | 2019-06-07 | 平安科技（深圳）有限公司 | Choosing method and device, storage medium, the electronic equipment of test sample |
| CN110163460A (en) * | 2018-03-30 | 2019-08-23 | 腾讯科技（深圳）有限公司 | A kind of method and apparatus determined using score value |
| CN110308995A (en) * | 2019-07-08 | 2019-10-08 | 童晓雯 | A kind of edge cloud computing service system edges cloud node deployment device |
| CN110858313A (en) * | 2018-08-24 | 2020-03-03 | 国信优易数据有限公司 | Crowd classification method and crowd classification system |
| JP2020086708A (en) * | 2018-11-20 | 2020-06-04 | 株式会社三菱Ｕｆｊ銀行 | Use amount prediction method and program |
| CN112232643A (en) * | 2020-09-25 | 2021-01-15 | 上海淇毓信息科技有限公司 | Method and device for managing business strategy and electronic equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080086345A1 (en) * | 2006-09-15 | 2008-04-10 | Electronic Data Systems Corporation | Asset Data Collection, Presentation, and Management |
| CN105335886A (en) * | 2014-05-28 | 2016-02-17 | 华为技术有限公司 | Method and device for processing financial data |
| CN105760957A (en) * | 2016-02-23 | 2016-07-13 | 国元证券股份有限公司 | Securities soft lost customer prediction method |
| CN106599436A (en) * | 2016-12-08 | 2017-04-26 | 湖南大学 | User in-room behavior prediction method for office building |
| CN107122390A (en) * | 2017-03-04 | 2017-09-01 | 华数传媒网络有限公司 | Recommendation system building method based on groups of users |
-
2017
- 2017-09-13 CN CN201710823670.7A patent/CN107622326B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080086345A1 (en) * | 2006-09-15 | 2008-04-10 | Electronic Data Systems Corporation | Asset Data Collection, Presentation, and Management |
| CN105335886A (en) * | 2014-05-28 | 2016-02-17 | 华为技术有限公司 | Method and device for processing financial data |
| CN105760957A (en) * | 2016-02-23 | 2016-07-13 | 国元证券股份有限公司 | Securities soft lost customer prediction method |
| CN106599436A (en) * | 2016-12-08 | 2017-04-26 | 湖南大学 | User in-room behavior prediction method for office building |
| CN107122390A (en) * | 2017-03-04 | 2017-09-01 | 华数传媒网络有限公司 | Recommendation system building method based on groups of users |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110163460A (en) * | 2018-03-30 | 2019-08-23 | 腾讯科技（深圳）有限公司 | A kind of method and apparatus determined using score value |
| CN110163460B (en) * | 2018-03-30 | 2023-09-19 | 腾讯科技（深圳）有限公司 | Method and equipment for determining application score |
| CN108564380A (en) * | 2018-04-11 | 2018-09-21 | 重庆大学 | A kind of telecommunication user sorting technique based on iteration decision tree |
| CN108564380B (en) * | 2018-04-11 | 2021-07-20 | 重庆大学 | Telecommunication user classification method based on iterative decision tree |
| CN110858313A (en) * | 2018-08-24 | 2020-03-03 | 国信优易数据有限公司 | Crowd classification method and crowd classification system |
| CN110858313B (en) * | 2018-08-24 | 2023-01-31 | 国信优易数据股份有限公司 | Crowd classification method and crowd classification system |
| JP7236256B2 (en) | 2018-11-20 | 2023-03-09 | 株式会社三菱Ｕｆｊ銀行 | Usage forecast method and program |
| JP2020086708A (en) * | 2018-11-20 | 2020-06-04 | 株式会社三菱Ｕｆｊ銀行 | Use amount prediction method and program |
| CN109857816A (en) * | 2019-01-11 | 2019-06-07 | 平安科技（深圳）有限公司 | Choosing method and device, storage medium, the electronic equipment of test sample |
| CN109857816B (en) * | 2019-01-11 | 2024-05-28 | 平安科技（深圳）有限公司 | Test sample selection method and device, storage medium and electronic equipment |
| CN110308995A (en) * | 2019-07-08 | 2019-10-08 | 童晓雯 | A kind of edge cloud computing service system edges cloud node deployment device |
| CN110308995B (en) * | 2019-07-08 | 2021-11-16 | 童晓雯 | Edge cloud node deployment device of edge cloud computing service system |
| CN112232643A (en) * | 2020-09-25 | 2021-01-15 | 上海淇毓信息科技有限公司 | Method and device for managing business strategy and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107622326B (en) | 2021-02-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107622326A (en) | User's classification, available resources Forecasting Methodology, device and equipment | |
| EP3985578A1 (en) | Method and system for automatically training machine learning model | |
| CN111738628A (en) | Risk group identification method and device | |
| CN110163647A (en) | A kind of data processing method and device | |
| CN109739844B (en) | Data classification method based on attenuation weight | |
| CN110457577B (en) | Data processing method, device, equipment and computer storage medium | |
| CN106844407B (en) | Tag network generation method and system based on data set correlation | |
| CN105225135B (en) | Potential customer identification method and device | |
| CN114510735B (en) | Role management-based intelligent shared financial management method and platform | |
| CN113656699B (en) | User feature vector determining method, related equipment and medium | |
| CN112396428B (en) | User portrait data-based customer group classification management method and device | |
| CN112232944B (en) | Method and device for creating scoring card and electronic equipment | |
| CN110276382A (en) | Listener clustering method, apparatus and medium based on spectral clustering | |
| CN112257818A (en) | Sample data processing method and device and computer equipment | |
| CN115545103A (en) | Abnormal data identification method, label identification method and abnormal data identification device | |
| CN113435900A (en) | Transaction risk determination method and device and server | |
| CN110619564B (en) | Anti-fraud feature generation method and device | |
| CN116821759A (en) | Identification prediction method and device for category labels, processor and electronic equipment | |
| Lejeune et al. | Optimization for simulation: LAD accelerator | |
| CN114021716A (en) | Model training method and system and electronic equipment | |
| CN113688206A (en) | Text recognition-based trend analysis method, device, equipment and medium | |
| CN114066513A (en) | User classification method and device | |
| CN113191570A (en) | Fund planning recommendation method, device and equipment based on deep learning | |
| CN112732891A (en) | Office course recommendation method and device, electronic equipment and medium | |
| CN113627653B (en) | Method and device for determining activity prediction strategy of mobile banking user |
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: 20200922 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: 20200922 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |