CN109978302A - A kind of credit-graded approach and equipment - Google Patents
A kind of credit-graded approach and equipment Download PDFInfo
- Publication number
- CN109978302A CN109978302A CN201711458946.2A CN201711458946A CN109978302A CN 109978302 A CN109978302 A CN 109978302A CN 201711458946 A CN201711458946 A CN 201711458946A CN 109978302 A CN109978302 A CN 109978302A
- Authority
- CN
- China
- Prior art keywords
- evaluation index
- data
- user
- score
- credit
- 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
- 238000013459 approach Methods 0.000 title claims abstract description 16
- 238000011156 evaluation Methods 0.000 claims abstract description 254
- 238000000034 method Methods 0.000 claims abstract description 64
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 32
- 230000008569 process Effects 0.000 claims abstract description 14
- 238000004590 computer program Methods 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 6
- 238000009434 installation Methods 0.000 claims description 4
- 238000013481 data capture Methods 0.000 claims description 3
- 239000000203 mixture Substances 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 9
- 238000004364 calculation method Methods 0.000 description 8
- 238000012545 processing Methods 0.000 description 7
- 238000004335 scaling law Methods 0.000 description 7
- 230000002159 abnormal effect Effects 0.000 description 6
- 238000007637 random forest analysis Methods 0.000 description 6
- 239000000725 suspension Substances 0.000 description 6
- 230000006399 behavior Effects 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 235000012054 meals Nutrition 0.000 description 4
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000010219 correlation analysis Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000012217 deletion Methods 0.000 description 2
- 230000037430 deletion Effects 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 206010000117 Abnormal behaviour Diseases 0.000 description 1
- 241001269238 Data Species 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013475 authorization Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000012954 risk control Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000006467 substitution reaction Methods 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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0609—Buyer or seller confidence or verification
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Development Economics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Theoretical Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Educational Administration (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Data Mining & Analysis (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The invention discloses a kind of credit-graded approach and equipment to improve the accuracy of user credit scoring for reducing the subjectivity of user credit scoring.This method comprises: obtaining N number of data sample of N number of user；The data sample includes the data of P evaluation index, and N, P are positive integer；The Score index system based on analytic hierarchy process AHP framework being made of destination layer, rule layer and solution layer is constructed according to N number of data sample；The destination layer is the scoring of user credit, the rule layer includes M element for influencing user credit scoring, and an element in the M element is corresponding at least one evaluation index that the solution layer includes, wherein, evaluation index corresponding to different elements is different, and M is positive integer；The weight of each evaluation index at least one evaluation index corresponding to each element is calculated by sorting algorithm；It is scored according to the weight of each evaluation index and Score index system user credit.
Description
Technical field
The present invention relates to big data processing technology field, in particular to a kind of credit-graded approach and equipment.
Background technique
The development of mobile service is so that the management control to user credit becomes operator's internal control, risk control and industry
One of important process that business is expanded.Each operator would generally according to information such as user's essential characteristic, credit history, consuming capacities,
It scores user credit, thus according to the scoring of user credit under the premise of rationally controlling risk, for different credits
The treatment mechanism of the user of grade implementable differentiation in terms of the mobile services such as authorization, refund, process limited realizes service money
Source is distributed rationally, while also providing the deposit of base power for the expansion of follow-up business.
Currently, mainly passing through analytic hierarchy process (AHP) (Analytic Hierarchy to the method that user credit scores
Process, AHP) credit scoring mechanism under frame.AHP mainly includes establishing hierarchy Model, constructing sentencing in each level
Disconnected matrix and consistency check and etc..Wherein, the judgment matrix in each level is constructed to be mainly used for determining that each level includes
The weight of evaluation index.But the judgment matrix in each level is constructed when carrying out user credit scoring by AHP at present
Process is by the way that evaluation index to be compared two-by-two, and by the importance of 1-9 scaling law mark between any two, and then determination is each
The importance of evaluation index, but the scale of this 1-9 scaling law relatively tends to rely on expertise, is commented by artificial definition
The importance of valence index so that finally obtaining the subjective of the weight of evaluation index, and then can largely drop
The accuracy of low user credit scoring.
Summary of the invention
The embodiment of the present invention provides a kind of credit-graded approach and equipment, for reducing the subjectivity of user credit scoring,
Improve the accuracy of user credit scoring.
In a first aspect, a kind of credit-graded approach is provided, this method comprises:
Obtain N number of data sample of N number of user；The data sample includes the data of P evaluation index, and N, P are positive whole
Number；
According to the N number of data sample building by destination layer, rule layer and solution layer form based on analytic hierarchy process AHP
The Score index system of framework；The destination layer is the scoring of user credit, and the rule layer includes influencing user credit scoring
M element, an element in the M element is corresponding at least one evaluation index that the solution layer includes,
In, evaluation index corresponding to different elements is different, and M is positive integer；
The power of each evaluation index at least one evaluation index corresponding to each element is calculated by sorting algorithm
Weight；
It is scored according to the weight of each evaluation index and the Score index system user credit.
Optionally, after the N number of data sample for obtaining N number of user, the method also includes:
According to the normal distribution law of data by the data of the qualitative evaluation index of continuous type in N number of data sample into
Row grouping；
All groups obtained after grouping of data are indicated by numerical value, wherein the data of a group are with identical
Numerical value is indicated.
Optionally, divided by what destination layer, rule layer and solution layer formed based on level according to N number of data sample building
The Score index system of analysis method AHP framework, comprising:
The degree of correlation of any two evaluation index in N number of evaluation index is calculated, if any two evaluation index
The degree of correlation be more than or equal to default relevance threshold, then reject the number of any one in any two evaluation index
According to；
The data that different degree is more than or equal to the evaluation index of default different degree threshold value are extracted, Q evaluation index is obtained
Data, Q be positive integer and Q≤N；Wherein, the different degree of an evaluation index for characterize one evaluation index to
The influence degree of family credit scoring；
According to the characteristic attribute of the evaluation index itself, the data of the Q evaluation index are divided into the M and are wanted
In one of element in element, to form the Score index system.
Optionally, weight of all evaluation indexes under corresponding element is calculated by sorting algorithm, comprising:
N number of data sample is divided into positive sample set and negative sample set；The number that the positive sample set includes
It is more than or equal to default scoring threshold value according to the credit scoring of sample, the credit for the data sample that the negative sample set includes is commented
Divide and is less than the default scoring threshold value；
Repeatedly classified by the sorting algorithm to N number of data sample, and calculate each time classification results with
The error amount of the result of the positive sample set and negative sample set of division；Wherein, evaluation index used by each subseries
Weighted；
The weight of evaluation index used by minimum error values in the error amount is determined as each described evaluation to refer to
Target weight.
Optionally, user credit is carried out according to the weight of each evaluation index and the Score index system
Scoring, comprising:
Calculate the data of each evaluation index that the data sample of the user includes for the user credit
Point；
The score value of the M element is calculated according to the weight of each evaluation index and the score；
It is scored according to the score value of the M element the user credit.
Optionally, the data sample for calculating the user includes credit of the data for the user of each evaluation index
Score, comprising:
Utilize formulaThe data sample for calculating the user includes just
To the score of evaluation index, and utilize formulaCalculate the number of the user
According to the score for the negative sense evaluation index that sample includes；The numerical value pass directly proportional to user credit scoring of the forward direction evaluation index
System, the numerical value and user credit of the negative sense evaluation index score inversely；
Wherein, x'jThe score for j-th of evaluation index that data sample for the user includes, xjFor the number of the user
According to the data for j-th of evaluation index that sample includes, min { x1j,...xnjRefer to for j-th of evaluation of N number of data sample
Minimum value in target data, max { x1j,...xnjBe N number of data sample j-th of evaluation index data in most
Big value.
Optionally,
The score value of the M element is calculated according to the weight of each evaluation index and the score, comprising:
Utilize formulaCalculate the score value of the M element；Wherein, CmIt wants for m-th for the user
The score value of element, 1≤m≤M, k are the total quantity of the corresponding evaluation index of m-th of element, tjIt is j-th of evaluation index in correspondence
Element under weight；
Then scored according to the score value of the M element the user credit, comprising:
Utilize formulaCalculate the credit scoring value of the user；Wherein, the credit that Z is the user is commented
Score value, pmFor the weight of m-th of element.
Second aspect, provides a kind of credit scoring equipment, which includes:
Data capture unit, for obtaining N number of data sample of N number of user；The data sample includes P evaluation index
Data, N, P be positive integer；
Score index system construction unit, for being constructed according to N number of data sample by destination layer, rule layer and scheme
The Score index system based on analytic hierarchy process AHP framework of layer composition；The destination layer is the scoring of user credit, the standard
Then layer includes M element for influencing user credit scoring, an element in the M element and the solution layer include to
A few evaluation index is corresponding, wherein evaluation index corresponding to different elements is different, and M is positive integer；
Computing unit, for calculating each at least one evaluation index corresponding to each element by sorting algorithm
The weight of evaluation index；
The computing unit is also used to weight and the Score index system pair according to each evaluation index
User credit scores.
Optionally, the device data acquiring unit is also used to N number of data according to the normal distribution laws of data
The data of the qualitative evaluation index of continuous type are grouped in sample；By all groups obtained after grouping of data by numerical value into
Row indicates, wherein the data of a group are indicated with identical numerical value.
Optionally, Score index system construction unit is specifically used for calculating any two evaluation in N number of evaluation index
The degree of correlation of index is rejected if the degree of correlation of any two evaluation index is more than or equal to default relevance threshold
The data of any one in any two evaluation index；It extracts different degree and is more than or equal to commenting for default different degree threshold value
The data of valence index, obtain the data of Q evaluation index, and Q is positive integer and Q≤N；Wherein, the different degree of an evaluation index
The influence degree to score for characterizing one evaluation index user credit；According to the feature category of the evaluation index itself
Property, the data of the Q evaluation index are divided into one of element in the M element, to form the scoring
Index system.
Optionally, the computing unit, specifically for N number of data sample is divided into positive sample set and negative sample
Set；The credit scoring for the data sample that the positive sample set includes is more than or equal to default scoring threshold value, the negative sample
The credit scoring for the data sample that this set includes is less than the default scoring threshold value；By the sorting algorithm to described N number of
Data sample is repeatedly classified, and calculates the result of the positive sample set and negative sample set of classification results and division each time
Error amount；Wherein, the weighted of evaluation index used by each subseries；By minimum error values institute in the error amount
The weight of the evaluation index of use is determined as the weight of each evaluation index.
Optionally, the computing unit is specifically also used to calculate each evaluation that the data sample of the user includes and refers to
Score of the target data for the credit of the user；It is calculated according to the weight of each evaluation index and the score
The score value of the M element；It is scored according to the score value of the M element the user credit.
Optionally, the computing unit is used to calculate the data for each evaluation index that the data sample of the user includes
For the score of the credit of the user, comprising:
The computing unit utilizes formulaCalculate the data of the user
The score for the positive evaluation index that sample includes, and utilize formulaIt calculates
The score for the negative sense evaluation index that the data sample of the user includes；The numerical value of the forward direction evaluation index is commented with user credit
It is divided into proportional relation, the numerical value and user credit of the negative sense evaluation index score inversely；
Wherein, x'jThe score for j-th of evaluation index that data sample for the user includes, xjFor the number of the user
According to the data for j-th of evaluation index that sample includes, min { x1j,...xnjRefer to for j-th of evaluation of N number of data sample
Minimum value in target data, max { x1j,...xnjBe N number of data sample j-th of evaluation index data in most
Big value.
Optionally,
The computing unit calculates the M element according to the weight and the score of each evaluation index
Score value, comprising:
The computing unit utilizes formulaCalculate the score value of the M element；Wherein, Cm is the use
The score value of m-th of element at family, 1≤m≤M, k are the total quantity of the corresponding evaluation index of m-th of element；tjIt is commented for j-th
Weight of the valence index under corresponding element；
The computing unit scores to the user credit according to the score value of the M element, comprising:
The computing unit utilizes formulaCalculate the credit scoring value of the user；Wherein, Z is described
The credit scoring value of user, pmFor the weight of m-th of element.
The third aspect provides a kind of computer installation, and described device includes at least one processor, and the processor is used for
The step of credit-graded approach such as first aspect offer is realized when executing the computer program stored in memory.
Fourth aspect provides a kind of computer readable storage medium, is stored thereon with computer program, the computer journey
The step of credit-graded approach such as first aspect offer is realized when sequence is executed by processor.
In embodiments of the present invention, after establishing Score index system, Score index body is obtained by sorting algorithm
The weight of the corresponding evaluation index of each element in system, and user credit is commented according to the weight and Score index system
Point.Wherein, the embodiment of the present invention substitutes the 1-9 scaling law in existing AHP by sorting algorithm, and then is determining evaluation index
Weight when without according to expertise, reduce the subjectivity of artificial scale, and then improve the accurate of user credit scoring
Property.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, will make below to required in the embodiment of the present invention
Attached drawing is briefly described, it should be apparent that, attached drawing described below is only some embodiments of the present invention, for
For those of ordinary skill in the art, without creative efforts, it can also be obtained according to these attached drawings other
Attached drawing.
Fig. 1 is the flow diagram of credit assessment method provided in an embodiment of the present invention；
Fig. 2 is the schematic diagram of the grouping provided in an embodiment of the present invention obtained according to normal distribution law；
Fig. 3 is the structural schematic diagram of Score index system provided in an embodiment of the present invention；
Fig. 4 is a kind of structural schematic diagram of credit appraisal equipment provided in an embodiment of the present invention；
Fig. 5 is a kind of structural schematic diagram of computer installation equipment provided in an embodiment of the present invention.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described.
The technical background of the embodiment of the present invention is described below.
Currently, constructing the process of the judgment matrix in each level by that will comment when carrying out user credit scoring by AHP
Valence index is compared two-by-two, by the importance of 1-9 scaling law mark between any two, and then determines the weight of each evaluation index
The property wanted, but the scale of this 1-9 scaling law relatively tends to rely on expertise, passes through the important of artificial definition evaluation index
Property, so that finally obtaining the subjective of the weight of evaluation index, and then it can largely reduce user credit and comment
The accuracy divided.
In consideration of it, the embodiment of the present invention provides a kind of credit-graded approach, in the method, Score index system is being established
Later, the weight of the corresponding evaluation index of each element in Score index system is obtained by random forests algorithm, and according to this
Weight and Score index system score to user credit.Wherein, the embodiment of the present invention is substituted by random forests algorithm
1-9 scaling law in existing AHP, so it is artificial without reducing according to expertise in the weight for determining evaluation index
The subjectivity of scale, and then improve the accuracy of user credit scoring.
Technical solution provided in an embodiment of the present invention is introduced with reference to the accompanying drawing.
Referring to Figure 1, one embodiment of the invention provides a kind of credit-graded approach, and this method can an implementation through the invention
The credit scoring equipment that example provides executes, which can for example pass through personal computer (Personal
Computer, PC) or the equipment such as server realize.This method comprises:
Step 101: obtaining N number of data sample of N number of user；The data sample includes the data of P evaluation index, N,
P is positive integer；
Step 102: according to the N number of data sample building by destination layer, rule layer and solution layer form based on level
The Score index system of analytic approach AHP framework；The destination layer is the scoring of user credit, and the rule layer includes believing user
Score influential M element, an element in the M element and the solution layer include at least one
Evaluation index is corresponding, wherein evaluation index corresponding to different elements is different, and M is positive integer；
Step 103: each at least one evaluation index corresponding to each element being calculated by sorting algorithm and is evaluated
The weight of index；
Step 104: according to the weight of each evaluation index and the Score index system to user credit into
Row scoring.
In the embodiment of the present invention, N number of data sample for establishing N number of user of Score index system is obtained first,
In, each data may each comprise the data of P evaluation index.Specifically, the Score index body in order to guarantee subsequent foundation
The quantity needs of the accuracy of system, the data sample of acquisition are enough, and evaluation index included by each data sample
Quantity is also required to enough.For example, stratified sampling, Ke Yixuan can be carried out according to the star ratio of user out of the whole country
Take 200,000 data samples at 20 general-purpose families, wherein data included by each data sample for example can be the user and exist
Average data in six months, and each data sample includes the data of multiple evaluation indexes, evaluation index for example can be
110 evaluation indexes such as age of user, gender, spending amount, the duration of call or ordering products quantity, certainly, the present invention are real
Example is applied not limit the quantity in month corresponding to the quantity of data sample, the quantity of evaluation index and user's average data
System, 200,000 data samples, 110 evaluation indexes and the average data in 6 months are only a kind of more preferred embodiment party
Formula.
After obtaining N number of data sample, due to the useless data that may score user credit in these data samples,
Or including some abnormal datas, therefore, it is also desirable to which useless data and abnormal data are deleted.
Specifically, the data for there was only the evaluation index of single value in all standard diagrams can be deleted.For example, evaluation index
Can be whether to be real-name authentication user, the user due to typically today being not real-name authentication can not carry out the use of business, because
The value of this usually evaluation index is all yes, thus the evaluation index is utterly useless for the scoring of user credit, and can also
Increase the calculation amount of follow-up data processing, therefore can be by the data dump of such evaluation index.
Specifically, the accounting that can also delete missing data is more than or equal to the number of the evaluation index of preset ratio threshold value
According to.For example, the accounting of an evaluation index missing data is greater than 80%, that is to say, that there is no this to comment in most of data sample
The data of valence index, thus, the data of the evaluation index have no reference significance, therefore can be by the number of such evaluation index
According to removing.Wherein, preset ratio threshold value, which can be, be configured when data deletion according to evaluation index, such as can be with
It is set as 80% or 70% etc., the embodiment of the present invention is without limitation.
Specifically, the data sample of data exception can also be deleted.For example, there was only the several of minority in multiple data samples
The data exception of data sample, then the data sample of these data exceptions can be deleted.
In the embodiment of the present invention, other than deleting some useless data, if there is the missing number of some evaluation indexes
According to accounting it is smaller, the data of missing are filled by the data of these evaluation indexes without deleting with default value, are write from memory
Recognizing value can be minimum standard value or mean value.For example, when evaluation index is education degree, if education degree in data sample
Missing is then filled the education degree of these data samples with minimum education degree；Alternatively, evaluation index is average for every user
When taking in (Average Revenue Per User, ARPU), if ARPU value lacks in data sample, then by these data samples
This ARPU value is filled with negative value.
It, can also be according to normal distribution law by the qualitative evaluation of continuous type in N number of data sample in the embodiment of the present invention
The data of index are grouped, and all groups obtained after grouping of data are indicated by numerical value, the data of a group
It is indicated with identical numerical value.Wherein, the qualitative evaluation index of continuous type refers to that the value of the evaluation index is the value of description type,
It is as nonumeric, and the value of the evaluation index has certain continuity.Fig. 2 is referred to, to carry out according to normal distribution law
The schematic diagram of grouping.Wherein, the distributed point of normal distribution as grouping threshold value, such as by -3 σ shown in Fig. 2, -2 σ, -1 σ,
μ, 1 σ, 2 σ and 3 σ are divided into 5 groups as packet threshold, by the index of evaluation index, then carry out table by numerical value for each group
Show, to facilitate the processing for carrying out follow-up data.It is grouped by normal distribution, each group after can making grouping can be with
Including certain customers, and the behavior of the user in each group is more closely, can also guarantee grouping not by normal distribution simultaneously
It is influenced by exceptional values such as maximum or minimums.
For example, evaluation index can be user education degree, such as the evaluation index can be divided into unknown, senior middle school with
Under, special secondary school or senior middle school, junior college, university degree and postgraduate or more this 6 groups, then can enable " unknown "=0, " below senior middle school "=1,
" special secondary school or senior middle school "=2, " junior college "=3, " university degree "=4, " postgraduate or more "=5 so then can will be in data sample
The data of qualitative evaluation index be converted to numerical value, the thus more convenient subsequent training etc. for data sample.
In the embodiment of the present invention, after by deleting the partial data in data sample, data sample includes
The quantity of evaluation index would generally be reduced, for example, the quantity of original evaluation index is 110, then carrying out data
After deletion, 81 evaluation indexes can be only included.But similar finger is also likely to be present in this evaluation index after deleting
Mark, the attribute of similar index is more similar, and identical to the played effect of user credit scoring, and hence it is also possible to institute
There is evaluation index to carry out correlation analysis, that is, calculates the correlation of any two evaluation index included by remaining data sample
Degree can then reject one of them and comment if the degree of correlation in any two index is more than or equal to default relevance threshold
The data of valence index, in this way, also reducing the burden of follow-up data processing.Wherein, default relevance threshold can rule of thumb or
Person is configured according to actual evaluation index, such as can be set to 80% or 90% etc., the embodiment of the present invention to this not
It is limited.
In the embodiment of the present invention, after the data for rejecting the biggish evaluation index of correlation, commented included by data sample
Valence index still may be very much, such as the quantity of evaluation index can be 50, and data volume is still more huge, thus may be used also
To extract after correlation analysis in remaining evaluation index, different degree is more than or equal to the evaluation of default different degree threshold value
The data of index obtain the data of Q evaluation index.Wherein, Q is positive integer and Q≤N, and different degree believes user for characterizing
With the influence degree of scoring.
Specifically, can be by machine algorithm come the different degree of Calculation Estimation index, and then choose wherein different degree and be greater than
Or the data of the evaluation index equal to default different degree threshold value.Wherein, machine algorithm for example can be IV analytic approach, principal component
Method or random forest method scheduling algorithm, in addition, the accuracy in order to guarantee the different degree calculated, can also pass through a variety of machines simultaneously
The different degree of device algorithm Calculation Estimation index.
Wherein, Q evaluation index can be obtained after screening by different degree, such as Q can be 28, i.e., this 28 are commented
The different degree of valence index is more than or equal to default different degree threshold value or this 28 evaluation indexes are that importance sorting is forward
28 evaluation indexes.Certainly, for the quantity of standard diagrams, the embodiment of the present invention is simultaneously not limited.
In the embodiment of the present invention, it can also be constructed according to the data of Q evaluation index by destination layer, rule layer and solution layer group
At the Score index system based on analytic hierarchy process AHP framework.Fig. 3 is referred to, is the structural schematic diagram of Score index system,
Wherein, destination layer is the scoring of user credit, and the rule layer, which includes the scoring to user credit influential M, to be wanted
Element, an element in the M element are corresponding at least one evaluation index that the solution layer includes, wherein different
Evaluation index corresponding to element is different.Wherein, shown in Fig. 3 be 5 elements, i.e., identity characteristic, consumer behavior, consumption energy
This 5 big element of power, credit history and relationship cycle, each element correspond to the evaluation of the part in above-mentioned 28 evaluation indexes and refer to
Mark.Which kind of certainly, the quantity of element and can reasonably be adjusted according to specific application scenarios for element, the present invention is simultaneously
It is not limited.
Specifically, the data of Q evaluation index can be divided into M and wanted according to the characteristic attribute of evaluation index itself
In one of element in element, to form Score index system.It is cell class in evaluation index by taking above-mentioned 5 elements as an example
Type identify when, cell type mark belong to it is related with geographical location locating for user, thus can by cell type identify divide
Into the identity characteristic of user；When evaluation index is the basic set meal amount of money, the basic set meal amount of money is basis selected by user
The amount of money of set meal, thus the basic set meal amount of money can be divided into the consumer behavior of user；It is ordering products in evaluation index
When total, ordering products sum can be used for measuring the consuming capacity of user, thus ordering products sum can be divided into user
Consuming capacity in；When evaluation index is full suspension duration, when full suspension a length of user's full suspension duration wherein,
Full suspension refers to that user can not make a phone call and the use of business to other users, and other users can not also put through this
The phone of user, it is longer that full suspension duration is often as the subscriber arrearage time, thus full suspension duration can be divided
Into user credit history；When evaluation index is caller relationship cycle number, then caller relationship cycle number can be divided into use
In the relationship cycle at family.
In the embodiment of the present invention, obtain include Q evaluation index data, i.e., finally obtained each data sample is equal
After Q evaluation index, then these data samples can be divided into positive sample set and negative sample set.Wherein, positive sample
The credit scoring for the data sample that this set includes is more than or equal to default scoring threshold value, the data sample that negative sample set includes
This credit scoring is less than default scoring threshold value, that is to say, that the corresponding user's letter of data sample included by positive sample set
With preferably, and the corresponding user credit of data sample included by negative sample set is poor.
Specifically, can be according to the respective standard of each operator for the standard for dividing positive sample set and negative sample set
It executes, such as the data sample for the user that spending amount is more than or equal to certain threshold value can be divided into positive sample set,
And the data sample for being less than the user of certain threshold value is divided into negative sample set.Specifically how to divide and do not limited herein,
This is related with the respective standard of each operator, is no longer repeated herein.When determining that a data sample is positive sample or negative
After sample, then can the data sample add for characterize the sample be positive sample or be negative sample label, such as
The label of positive sample is 1, and the label of negative sample is 0；Alternatively, the label of positive sample is 0, the label of negative sample is 1.
It in the division for carrying out positive sample set and negative sample set, can repeatedly be divided, choose division result more
It is a kind of as final division result in similar division result.Wherein, number included by positive sample set and negative sample set
According to sample without coincidence.
In the embodiment of the present invention, after dividing positive sample set and negative sample set, can by sorting algorithm come pair
Data sample is repeatedly classified, wherein the weighted of evaluation index used by each subseries.Passing through sorting algorithm
When being classified, then the classification results and the positive sample set divided above and negative sample set of each subseries can be calculated
As a result error amount.Specifically, when the weight using different evaluation index is as the parameter classified, obtained classification results
May be different, thus then can repeatedly adjust weight and classify, to choose the classification knot of minimum error values in final error value
The weight of evaluation index used by fruit, and using the weight as the power of each evaluation index to score eventually for user credit
Weight.
Specifically, sorting algorithm can for example use decision Tree algorithms, certainly, in order to enable evaluation index can be more quasi-
Really, random forests algorithm can also be further used, alternatively, other possible algorithms, the embodiment of the present invention are without limitation.
As it can be seen that in embodiments of the present invention, the weight of evaluation index is obtained by sorting algorithm, base in the prior art is thereby reduced
The subjectivity of evaluation index is determined in expertise, so that end user's credit scoring is more accurate.
In the embodiment of the present invention, after the weight for obtaining evaluation index, then it can score user credit.Its
In, the sample used that scores here is data sample to be scored.
Specifically, each evaluation index included by the data sample for the user that can first treat scoring scores.Its
In, evaluation index can also be divided into positive evaluation index and negative sense evaluation index, and the numerical value of positive evaluation index and user believe
Directly proportional with scoring, numerical value and the user credit scoring of negative sense evaluation index are inversely proportional.
The score calculation formula of positive evaluation index are as follows:
The score calculation formula of negative sense evaluation index are as follows:
Corresponding calculation formula is respectively adopted when calculating the score of different evaluation indexes.Wherein, x'jWait score
The score for j-th of evaluation index that the data sample of user includes, xjInclude j-th of data sample for user to be scored
The data of evaluation index, min { x1j,...xnjIt is for establishing j-th of evaluation index of N number of data sample of score-system
Minimum value in data, max { x1j,...xnjBe N number of data sample j-th of evaluation index data in maximum value.
In the embodiment of the present invention, for the convenience of calculating, before the score for calculating each evaluation index, can also to
The data that the data sample at family includes are normalized, with the dimension of uniform data.
In the embodiment of the present invention, after the score that each evaluation index is calculated, then each element can be carried out
The calculating of score value.
Specifically, the score value calculation formula of element are as follows:
Wherein, CmFor the score value of m-th of element of user to be scored, 1≤m≤M, k are corresponding for element calculated
Evaluation index quantity；tjFor weight of j-th of evaluation index under corresponding element, i.e., obtained by sorting algorithm each
The weight of a evaluation index.
In the embodiment of the present invention, after the score value that each element is calculated, then user to be scored can be calculated
Credit scoring value.User credit score value calculation formula are as follows:
Wherein, Z is user credit score value to be scored, pmFor the weight of m-th of element.Specifically, the weight of element
Can be based on weighted value used by (FICO) credit scoring system, i.e. the weight of the identity characteristic power that is 15%, consumer behavior
Weight is 25%, the weight of consuming capacity is 20%, the weight of credit history is 35%, the weight of relationship cycle is 5%, certainly,
When practical application, each operator can also voluntarily be adjusted weighted value, more to meet the demand of an operator itself.
It is complete in Score index system construction in order to guarantee that the credit scoring value calculated is reliable enough in the embodiment of the present invention
At and the Weight Acquisition of evaluation index after, also need to verify the Score index system.Specifically, being verified
When, it can be using score value accurate rate and recall ratio as evaluation criterion.
Table 1 is referred to, carries out letter for the credit-graded approach of the embodiment of the present invention obtained by the way of off-line verification
With scoring, the comparison of the result of improved method and existing method progress credit scoring is called in the following text.Wherein, table 1 is with international roaming service
For opening, the threshold value for the user credit score value that selected international roaming service is opened is 300 points, chooses 200,000 data
Sample, wherein 20,000 data samples are as training set, 180,000 data are as verifying collection, finally obtained accurate rate and recall ratio
As shown in table 1.Wherein, accurate rate is to refer to judge correct result institute in result that whether user can launch international roaming service
The ratio accounted for, recall ratio are the ratio of the correct result and practical correct result that are judged by existing method or improved method.
| Method | Accurate rate | Recall ratio |
| Existing method | 53.68% | 73.33% |
| Improved method | 74.53% | 97.84% |
Table 1
It can be seen that improved method is promoted on accurate rate and recall ratio relative to existing method.
Table 2 is referred to, for by the online verification result actually used of improved method.Wherein, the verifying knot of table 2
Scene applied by fruit is that international roaming service is exempted to prestore and activates the service, altogether online verifying two months, altogether to 30972 users'
Data sample scores, and judges whether to launch international roaming service to exempt to prestore and activate the service.Wherein, abnormal user is
There are the quantity of the user of abnormal behaviour in the number of users of judgement, for example, existing method can be opened, what improved method can also be opened
Number of users is 26146, and wherein the quantity of abnormal user is 68, and accounting for existing method altogether can open, the use that improved method can also be opened
The 0.3% of amount.
| Explanation | Number of users | Abnormal user | Accounting |
| Existing method can be opened, and improved method can also be opened | 26146 | 68 | 0.3% |
| Existing method cannot be opened, and improved method can not be opened | 162 | 159 | 98.15% |
| Existing method cannot be opened, and improved method can be opened | 4493 | 1 | 0.02% |
| Existing method can be opened, and improved method cannot be opened | 171 | 170 | 99.42% |
Table 2
As it can be seen that the accounting of the abnormal user for the user that can be opened by improved method judgement is significantly less than existing method
Accounting, thus mutually than existing methods, improved method is more reliable.
In conclusion after establishing Score index system, being obtained by random forests algorithm in the embodiment of the present invention
The weight of the corresponding evaluation index of each element in Score index system, and user is believed according to the weight and Score index system
With scoring.Wherein, the embodiment of the present invention substitutes the 1-9 scaling law in existing AHP, Jin Er by random forests algorithm
Without reducing the subjectivity of artificial scale according to expertise when determining the weight of evaluation index, and then improve user's letter
With the accuracy of scoring.In addition, due to calculating the scale value of weight without artificial adjustment weight or adjustment, and need to only pass through
Sorting algorithm is obtained, and the work load of personnel is more reduced.
Fig. 4 is referred to, based on the same inventive concept, one embodiment of the invention provides a kind of credit scoring equipment 40, this sets
It is standby to include:
Data capture unit 401, for obtaining N number of data sample of N number of user；Data sample includes P evaluation index
Data, N, P be positive integer；
Score index system construction unit 402, for being constructed according to N number of data sample by destination layer, rule layer and scheme
The Score index system based on analytic hierarchy process AHP framework of layer composition；Destination layer is the scoring of user credit, and rule layer includes
Influence M element of user credit scoring, at least one evaluation index phase that an element in M element includes with solution layer
It is corresponding, wherein evaluation index corresponding to different elements is different, and M is positive integer；
Computing unit 403, it is every at least one evaluation index corresponding to each element for being calculated by sorting algorithm
The weight of one evaluation index；
Computing unit 403 is also used to weight and Score index system according to each evaluation index to user credit
It scores.
Optionally, device data acquiring unit 401 is also used to N number of data sample according to the normal distribution laws of data
The data of the qualitative evaluation index of middle continuous type are grouped；All groups obtained after grouping of data are subjected to table by numerical value
Show, wherein the data of a group are indicated with identical numerical value.
Optionally, Score index system construction unit 402 is specifically used for calculating any two evaluation in N number of evaluation index
The degree of correlation of index is rejected any if the degree of correlation of any two evaluation index is more than or equal to default relevance threshold
The data of any one in two evaluation indexes；It extracts different degree and is more than or equal to the evaluation index for presetting different degree threshold value
Data, obtain the data of Q evaluation index, and Q is positive integer and Q≤N；Wherein, the different degree of an evaluation index is for characterizing
The influence degree that one evaluation index scores to user credit；According to the characteristic attribute of evaluation index itself, by Q evaluation index
Data be divided into one of element in M element, to form Score index system.
Optionally, computing unit 403, specifically for N number of data sample is divided into positive sample set and negative sample set；
The credit scoring for the data sample that positive sample set includes is more than or equal to default scoring threshold value, the number that negative sample set includes
It is less than default scoring threshold value according to the credit scoring of sample；Repeatedly classified by sorting algorithm to N number of data sample, and is calculated
The error amount of the result of the positive sample set and negative sample set of classification results and division each time；Wherein, each subseries institute
The weighted of the evaluation index of use；The weight of evaluation index used by minimum error values in error amount is determined as each
The weight of a evaluation index.
Optionally, computing unit 403 specifically are also used to calculate the number for each evaluation index that the data sample of user includes
According to the score of the credit for user；The score value of M element is calculated according to the weight of each evaluation index and score；Root
It scores according to the score value of M element user credit.
Optionally, computing unit 403 be used to calculate the data of each evaluation index that the data sample of user includes for
The score of the credit of user, comprising:
Computing unit 403 utilizes formulaCalculate the data sample packet of user
The score of the positive evaluation index included, and utilize formulaCalculate user's
The score for the negative sense evaluation index that data sample includes；The numerical value of positive evaluation index and user credit scoring are proportional,
The numerical value and user credit of negative sense evaluation index score inversely；
Wherein, x'jThe score for j-th of evaluation index that data sample for user includes, xjFor the data sample packet of user
The data of j-th of the evaluation index included, min { x1j,...xnjIt is in the data of j-th of evaluation index of N number of data sample
Minimum value, max { x1j,…xnjBe N number of data sample j-th of evaluation index data in maximum value.
Optionally,
Computing unit 403 calculates the score value of M element according to the weight and score of each evaluation index, comprising:
Computing unit 403 utilizes formulaCalculate the score value of M element；Wherein, Cm is m-th of user
The score value of element, 1≤m≤M, k are the total quantity of the corresponding evaluation index of m-th of element；tjIt is j-th of evaluation index right
Weight under the element answered；
Computing unit 403 scores to user credit according to the score value of M element, comprising:
Computing unit 403 utilizes formulaCalculate the credit scoring value of user；Wherein, Z is the credit of user
Score value, pmFor the weight of m-th of element.
The equipment can be used for executing method provided by embodiment shown in FIG. 1, therefore, for each function of the equipment
The function etc. that module can be realized can refer to the description of embodiment shown in FIG. 1, seldom repeat.
Fig. 5 is referred to, one embodiment of the invention also provides a kind of computer installation, which includes at least one
Processor 501, at least one processor 501 is for realizing reality shown in FIG. 1 when executing the computer program stored in memory
The step of credit-graded approach of example offer is provided.
Optionally, at least one processor 501 can specifically include central processing unit (CPU), application-specific integrated circuit
(application specific integrated circuit, ASIC) can be one or more and hold for controlling program
Capable integrated circuit can be use site programmable gate array (field programmable gate array, FPGA) and open
The hardware circuit of hair, can be baseband processor.
Optionally, at least one processor 501 may include at least one processing core.
Optionally, which further includes memory 502, and memory 502 may include read-only memory (read
Only memory, ROM), random access memory (random access memory, RAM) and magnetic disk storage.Memory
502 for storing data required when at least one processor 501 is run.The quantity of memory 502 is one or more.Its
In, memory 502 is shown together in Fig. 5, but it is understood that memory 502 is not essential functional module, therefore
It is shown in dotted line in Fig. 5.
One embodiment of the invention also provides a kind of computer readable storage medium, is stored thereon with computer program, described
The step of credit-graded approach that embodiment shown in FIG. 1 provides is realized when computer program is executed by processor.
In embodiments of the present invention, it should be understood that disclosed device and method, it can be real by another way
It is existing.For example, apparatus embodiments described above are merely indicative, for example, the division of the unit or unit, only
A kind of logical function partition, there may be another division manner in actual implementation, for example, multiple units or components can combine or
Person is desirably integrated into another system, or some features can be ignored or not executed.Another point, shown or discussed is mutual
Between coupling, direct-coupling or communication connection can be through some interfaces, the INDIRECT COUPLING or communication link of equipment or unit
It connects, can be electrical or other forms.
Each functional unit in embodiments of the present invention can integrate in one processing unit or each unit can also
To be independent physical module.
If the integrated unit is realized in the form of SFU software functional unit and sells or use as independent product
When, it can store in a computer readable storage medium.Based on this understanding, the technical solution of the embodiment of the present invention
All or part can be embodied in the form of software products, which is stored in a storage medium
In, including some instructions use so that a computer equipment, such as can be personal computer, server or network are set
Standby etc. or processor (processor) performs all or part of the steps of the method described in the various embodiments of the present invention.And it is above-mentioned
Storage medium include: general serial bus USB (universal serial bus flash drive), mobile hard disk,
The various media that can store program code such as ROM, RAM, magnetic or disk.
The above, above embodiments are only described in detail to the technical solution to the application, but the above implementation
The method that the explanation of example is merely used to help understand the embodiment of the present invention, should not be construed as the limitation to the embodiment of the present invention.This
Any changes or substitutions that can be easily thought of by those skilled in the art, should all cover the embodiment of the present invention protection scope it
It is interior.
Claims (10)
1. a kind of credit-graded approach characterized by comprising
Obtain N number of data sample of N number of user；The data sample includes the data of P evaluation index, and N, P are positive integer；
According to the N number of data sample building by destination layer, rule layer and solution layer form based on analytic hierarchy process AHP framework
Score index system；The destination layer is the scoring of user credit, and the rule layer includes M for influencing user credit scoring
Element, an element in the M element are corresponding at least one evaluation index that the solution layer includes, wherein no
With the difference of evaluation index corresponding to element, M is positive integer；
The weight of each evaluation index at least one evaluation index corresponding to each element is calculated by sorting algorithm；
It is scored according to the weight of each evaluation index and the Score index system user credit.
2. the method as described in claim 1, which is characterized in that after the N number of data sample for obtaining N number of user, the side
Method further include:
The data of the qualitative evaluation index of continuous type in N number of data sample are divided according to the normal distribution law of data
Group；
All groups obtained after grouping of data are indicated by numerical value, wherein the identical numerical value of the data of a group
It is indicated.
3. method according to claim 1 or 2, which is characterized in that according to N number of data sample building by destination layer, standard
The then Score index system based on analytic hierarchy process AHP framework of layer and solution layer composition, comprising:
The degree of correlation of any two evaluation index in N number of evaluation index is calculated, if the phase of any two evaluation index
Guan Du is more than or equal to default relevance threshold, then rejects the data of any one in any two evaluation index；
The data that different degree is more than or equal to the evaluation index of default different degree threshold value are extracted, the number of Q evaluation index is obtained
According to Q is positive integer and Q≤N；Wherein, the different degree of an evaluation index believes user for characterizing one evaluation index
With the influence degree of scoring；
According to the characteristic attribute of the evaluation index itself, the data of the Q evaluation index are divided into the M element
One of element in, to form the Score index system.
4. method according to claim 2, which is characterized in that calculate all evaluation indexes by sorting algorithm and wanted corresponding
Weight under element, comprising:
N number of data sample is divided into positive sample set and negative sample set；The data sample that the positive sample set includes
This credit scoring is more than or equal to default scoring threshold value, and the credit scoring for the data sample that the negative sample set includes is small
In the default scoring threshold value；
Repeatedly classified by the sorting algorithm to N number of data sample, and calculates classification results each time and divide
Positive sample set and negative sample set result error amount；Wherein, the weight of evaluation index used by each subseries
It is different；
The weight of evaluation index used by minimum error values in the error amount is determined as each evaluation index
Weight.
5. the method as described in claim 1, which is characterized in that according to the weight of each evaluation index and institute's commentary
Point index system scores to user credit, comprising:
The data of each evaluation index that the data sample of the user includes are calculated for the score of the credit of the user；
The score value of the M element is calculated according to the weight of each evaluation index and the score；
It is scored according to the score value of the M element the user credit.
6. method as claimed in claim 5, which is characterized in that the data sample for calculating the user includes each evaluation index
Data for the user credit score, comprising:
Utilize formulaCalculate the positive evaluation that the data sample of the user includes
The score of index, and utilize formulaCalculate the data sample of the user
Including negative sense evaluation index score；The numerical value of the forward direction evaluation index and user credit scoring are proportional, described
The numerical value and user credit of negative sense evaluation index score inversely；
Wherein, x'jThe score for j-th of evaluation index that data sample for the user includes, xjFor the data sample of the user
Originally the data for j-th of evaluation index for including, min { x1j,...xnjBe N number of data sample j-th of evaluation index
Minimum value in data, max { x1j,...xnjBe N number of data sample j-th of evaluation index data in maximum
Value.
7. method as claimed in claim 5, which is characterized in that
The score value of the M element is calculated according to the weight of each evaluation index and the score, comprising:
Utilize formulaCalculate the score value of the M element；Wherein, CmIt is commented for m-th element of the user
Score value, 1≤m≤M, k are the total quantity of the corresponding evaluation index of m-th of element, tjIt is j-th of evaluation index in corresponding element
Under weight；
Then scored according to the score value of the M element the user credit, comprising:
Utilize formulaCalculate the credit scoring value of the user；Wherein, Z is the credit scoring value of the user,
pmFor the weight of m-th of element.
8. a kind of credit scoring equipment characterized by comprising
Data capture unit, for obtaining N number of data sample of N number of user；The data sample includes the number of P evaluation index
According to N, P are positive integer；
Score index system construction unit, for being constructed according to N number of data sample by destination layer, rule layer and solution layer group
At the Score index system based on analytic hierarchy process AHP framework；The destination layer is the scoring of user credit, the rule layer
The M element including influencing user credit scoring, an element in the M element include with the solution layer at least one
A evaluation index is corresponding, wherein evaluation index corresponding to different elements is different, and M is positive integer；
Computing unit is evaluated for calculating each at least one evaluation index corresponding to each element by sorting algorithm
The weight of index；
The computing unit is also used to weight and the Score index system according to each evaluation index to user
Credit scores.
9. a kind of computer installation, which is characterized in that described device includes processor, and the processor is for executing in memory
It is realized when the computer program of storage such as the step of any one of claim 1-7 the method.
10. a kind of computer readable storage medium, is stored thereon with computer program, it is characterised in that: the computer program
It is realized when being executed by processor such as the step of any one of claim 1-7 the method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711458946.2A CN109978302A (en) | 2017-12-28 | 2017-12-28 | A kind of credit-graded approach and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711458946.2A CN109978302A (en) | 2017-12-28 | 2017-12-28 | A kind of credit-graded approach and equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109978302A true CN109978302A (en) | 2019-07-05 |
Family
ID=67074509
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711458946.2A Pending CN109978302A (en) | 2017-12-28 | 2017-12-28 | A kind of credit-graded approach and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109978302A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110599312A (en) * | 2019-09-29 | 2019-12-20 | 支付宝(杭州)信息技术有限公司 | Credit-based interactive credit evaluation method and device |
| CN111626640A (en) * | 2020-06-08 | 2020-09-04 | 中煤科工集团重庆研究院有限公司 | Coal mine safety comprehensive risk evaluation method and system based on neutral reference object |
| CN114444576A (en) * | 2021-12-30 | 2022-05-06 | 北京达佳互联信息技术有限公司 | Data sampling method and device, electronic equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104123656A (en) * | 2014-07-14 | 2014-10-29 | 广东省标准化研究院 | Credit evaluation method based on AHP (analytic hierarchy process) |
| CN106897931A (en) * | 2016-06-12 | 2017-06-27 | 阿里巴巴集团控股有限公司 | A kind of recognition methods of abnormal transaction data and device |
-
2017
- 2017-12-28 CN CN201711458946.2A patent/CN109978302A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104123656A (en) * | 2014-07-14 | 2014-10-29 | 广东省标准化研究院 | Credit evaluation method based on AHP (analytic hierarchy process) |
| CN106897931A (en) * | 2016-06-12 | 2017-06-27 | 阿里巴巴集团控股有限公司 | A kind of recognition methods of abnormal transaction data and device |
Non-Patent Citations (2)
| Title |
|---|
| 未瑞等: "基于聚类分析和层次分析法的电力客户信用评价研究", 《现代电力》 * |
| 李金柱等: "一种基于改进AHP的电信企业信用评分模型", 《通信技术》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110599312A (en) * | 2019-09-29 | 2019-12-20 | 支付宝(杭州)信息技术有限公司 | Credit-based interactive credit evaluation method and device |
| CN111626640A (en) * | 2020-06-08 | 2020-09-04 | 中煤科工集团重庆研究院有限公司 | Coal mine safety comprehensive risk evaluation method and system based on neutral reference object |
| CN114444576A (en) * | 2021-12-30 | 2022-05-06 | 北京达佳互联信息技术有限公司 | Data sampling method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110413707A (en) | The excavation of clique's relationship is cheated in internet and checks method and its system | |
| CN109598095A (en) | Method for building up, device, computer equipment and the storage medium of scorecard model | |
| CN109688275A (en) | Harassing call recognition methods, device and storage medium | |
| CN104573304A (en) | User property state assessment method based on information entropy and cluster grouping | |
| CN106095939B (en) | The acquisition methods and device of account authority | |
| CN107644098A (en) | Method, device, equipment and storage medium for identifying fraudulent behavior | |
| CN109978302A (en) | A kind of credit-graded approach and equipment | |
| CN104657466B (en) | A kind of user interest recognition methods and device based on forum postings feature | |
| CN112396428B (en) | User portrait data-based customer group classification management method and device | |
| CN112559900A (en) | Product recommendation method and device, computer equipment and storage medium | |
| CN109840778A (en) | The recognition methods of fraudulent user and device, readable storage medium storing program for executing | |
| CN109063984A (en) | Risk passenger method, apparatus, computer equipment and storage medium | |
| CN105740434B (en) | Network information methods of marking and device | |
| CN110765351A (en) | Target user identification method and device, computer equipment and storage medium | |
| CN108629508A (en) | Credit risk sorting technique, device, computer equipment and storage medium | |
| CN108563786A (en) | Text classification and methods of exhibiting, device, computer equipment and storage medium | |
| CN114782123A (en) | Credit assessment method and system | |
| JP7092194B2 (en) | Information processing equipment, judgment method, and program | |
| CN113724061A (en) | Consumer financial product credit scoring method and device based on customer grouping | |
| CN106997340A (en) | The generation of dictionary and the Document Classification Method and device using dictionary | |
| CN107767278B (en) | Method and device for constructing community hierarchy | |
| CN110428150A (en) | A kind of change index generation method and device | |
| CN110210884A (en) | Determine the method, apparatus, computer equipment and storage medium of user characteristic data | |
| CN111931035B (en) | Service recommendation method, device and equipment | |
| CN114626940A (en) | Data analysis method and device and electronic equipment |
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 | ||
| CB02 | Change of applicant information |
Address after: 518048, 10 floor, Guo Tong Building, 9023 Binhe Road, Futian District, Shenzhen, Guangdong. Applicant after: CHINA MOBILE INFORMATION TECHNOLOGY Co.,Ltd. Applicant after: CHINA MOBILE COMMUNICATIONS GROUP Co.,Ltd. Address before: 518048, 10 floor, Guo Tong Building, 9023 Binhe Road, Futian District, Shenzhen, Guangdong. Applicant before: CHINA MOBILE INFORMATION TECHNOLOGY Co.,Ltd. Applicant before: China Mobile Communications Corp. |
|
| CB02 | Change of applicant information | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200327 Address after: Room 1006, building 16, yard 16, Yingcai North Third Street, future science city, Changping District, Beijing 102209 Applicant after: China Mobile Information Technology Co.,Ltd. Applicant after: CHINA MOBILE COMMUNICATIONS GROUP Co.,Ltd. Address before: 518048, 10 floor, Guo Tong Building, 9023 Binhe Road, Futian District, Shenzhen, Guangdong. Applicant before: CHINA MOBILE INFORMATION TECHNOLOGY Co.,Ltd. Applicant before: CHINA MOBILE COMMUNICATIONS GROUP Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190705 |
|
| RJ01 | Rejection of invention patent application after publication |