CN109447461A - User credit appraisal procedure and device, electronic equipment, storage medium - Google Patents
User credit appraisal procedure and device, electronic equipment, storage medium Download PDFInfo
- Publication number
- CN109447461A CN109447461A CN201811260889.1A CN201811260889A CN109447461A CN 109447461 A CN109447461 A CN 109447461A CN 201811260889 A CN201811260889 A CN 201811260889A CN 109447461 A CN109447461 A CN 109447461A
- Authority
- CN
- China
- Prior art keywords
- parameter
- branch mailbox
- target component
- class
- class parameter
- 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
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
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0609—Buyer or seller confidence or verification
-
- 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
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
-
- 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
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/06—Asset management; Financial planning or analysis
Abstract
The disclosure is directed to a kind of user credit appraisal procedure and device, electronic equipment, storage mediums, are related to Internet technical field, this method comprises: obtaining multiple characteristic informations of target user, the multiple characteristic information includes first kind parameter and the second class parameter;The first kind parameter and the second class parameter are pre-processed;The pretreated first kind parameter is converted, target component is generated;The pretreated second class parameter and the target component are inputted into machine learning model, obtain the credit evaluation result of the target user;Wherein, IV value is the first kind parameter lower than the characteristic information of preset threshold, and the characteristic information that IV value is higher than the preset threshold is the second class parameter.The disclosure can more accurately determine user credit assessment result, accurately identify credit risk.
Description
Technical field
This disclosure relates to which Internet technical field, is commented in particular to a kind of user credit appraisal procedure, user credit
Estimate device, electronic equipment and computer readable storage medium.
Background technique
Credit evaluation card mold type is financial field one of the most common type risk score model, this model in interpretation and
Algorithm complexity obtains balance.
In the related technology, typically using the Default Probability of the parameter prediction user of strong Financial Attribute.But it is big susceptible
The user data obtained under condition does not have very strong Financial Attribute.Therefore, the data volume for the strong Financial Attribute parameter that can be used
Less, the credit evaluation result that may cause prediction is inaccurate and application range is restricted, and cannot accurately measure user's wind
Danger.
It should be noted that information is only used for reinforcing the reason to the background of the disclosure disclosed in above-mentioned background technology part
Solution, therefore may include the information not constituted to the prior art known to persons of ordinary skill in the art.
Summary of the invention
The disclosure is designed to provide a kind of user credit appraisal procedure and device, electronic equipment, storage medium, in turn
Asking for consumer's risk cannot be accurately measured caused by overcoming the limitation and defect due to the relevant technologies at least to a certain extent
Topic.
Other characteristics and advantages of the disclosure will be apparent from by the following detailed description, or partially by the disclosure
Practice and acquistion.
According to one aspect of the disclosure, a kind of user credit appraisal procedure is provided, comprising: obtain the multiple of target user
Characteristic information, the multiple characteristic information include first kind parameter and the second class parameter;To the first kind parameter and described
Two class parameters are pre-processed;The pretreated first kind parameter is converted, target component is generated;After pre-processing
The second class parameter and the target component input machine learning model, obtain the credit evaluation knot of the target user
Fruit;Wherein, IV value is the first kind parameter lower than the characteristic information of preset threshold, and IV value is higher than the feature of the preset threshold
Information is the second class parameter.
In a kind of exemplary embodiment of the disclosure, the first kind parameter and the second class parameter are located in advance
Reason includes: to carry out branch mailbox processing respectively to the first kind parameter and the second class parameter by weight evidence weight values, after obtaining branch mailbox
First kind parameter and branch mailbox after the second class parameter.
In a kind of exemplary embodiment of the disclosure, the pretreated first kind parameter is converted, is generated
Target component includes: to carry out feature group to the first kind parameter after the associated branch mailbox of each theme using linear discriminent algorithm
It closes, generates the target component.
In a kind of exemplary embodiment of the disclosure, the method also includes: the target component is divided again
Case, and the target component after branch mailbox again is put into candidate parameter pond;The second class parameter after branch mailbox is put into the candidate change
Measure pond.
In a kind of exemplary embodiment of the disclosure, by the pretreated second class parameter and the target component
Input machine learning model includes: to reject the second class parameter in the candidate variables pond after branch mailbox and the again target after branch mailbox
Multicollinearity between parameter, obtains rest parameter;The rest parameter is inputted into the machine learning model.
In a kind of exemplary embodiment of the disclosure, reject the second class parameter in the candidate variables pond after branch mailbox and
Multicollinearity between target component after branch mailbox again, obtaining rest parameter includes: to reject from the candidate variables pond
Weight evidence weight values target component after branch mailbox less than the second class parameter after the branch mailbox of preset value and again, obtains the residue
Parameter.
In a kind of exemplary embodiment of the disclosure, weight evidence weight values are rejected from the candidate variables pond less than default
The second class parameter after the branch mailbox of value and the again target component after branch mailbox include: according to the weight evidence weight values from small to large
Put in order, the second class parameter and the mesh after branch mailbox again after the weight evidence weight values are less than with the branch mailbox of the preset value
Mark parameter is rejected;Recalculate the second class parameter after the branch mailbox of rejecting and the again card of the target component after branch mailbox
According to weighted value;According to the weight evidence weight values putting in order from small to large, the weight evidence weight values recalculated are less than described
The second class parameter after the branch mailbox of the preset value and target component after branch mailbox is rejected again, until having rejected the card
Until the second class parameter and target component that are less than the preset value according to weighted value.
According to one aspect of the disclosure, a kind of user credit assessment device is provided, comprising: feature obtains module, is used for
Multiple characteristic informations of target user are obtained, the multiple characteristic information includes first kind parameter and the second class parameter;Parameter is pre-
Processing module, for being pre-processed to the first kind parameter and the second class parameter;Target component generation module, is used for
The pretreated first kind parameter is converted, target component is generated;Assessment result determining module, for that will pre-process
The second class parameter and the target component afterwards inputs machine learning model, obtains the credit evaluation knot of the target user
Fruit;Wherein, IV value is the first kind parameter lower than the characteristic information of preset threshold, and IV value is higher than the feature of the preset threshold
Information is the second class parameter.
According to one aspect of the disclosure, a kind of electronic equipment is provided, comprising: processor;And memory, for storing
The executable instruction of the processor;Wherein, the processor is configured to above-mentioned to execute via the executable instruction is executed
User credit appraisal procedure described in any one.
According to one aspect of the disclosure, a kind of computer readable storage medium is provided, computer program is stored thereon with,
The computer program realizes user credit appraisal procedure described in above-mentioned any one when being executed by processor.
A kind of user credit appraisal procedure for being there is provided in disclosure exemplary embodiment, user credit assessment device, electronics
In equipment and computer readable storage medium, on the one hand, by carrying out being converted into mesh to pretreated first kind parameter
Parameter is marked, the target component that first kind parameter can be made to be converted into is used for credit evaluation, avoids in the related technology only by the
Two class parameters are used for the problem that data volume is insufficient and application range is small caused by credit evaluation, increase data volume and application
Range;On the other hand, by the way that the target component generated after pretreated second class parameter and conversion is inputted machine learning
Model increases data volume, so as to be obtained based on the target component generated after pretreated second class parameter and conversion
To accurate credit evaluation as a result, accurately measuring user credit risk.
It should be understood that above general description and following detailed description be only it is exemplary and explanatory, not
The disclosure can be limited.
Detailed description of the invention
The drawings herein are incorporated into the specification and forms part of this specification, and shows the implementation for meeting the disclosure
Example, and together with specification for explaining the principles of this disclosure.It should be evident that the accompanying drawings in the following description is only the disclosure
Some embodiments for those of ordinary skill in the art without creative efforts, can also basis
These attached drawings obtain other attached drawings.
Fig. 1 schematically shows a kind of user credit appraisal procedure schematic diagram in disclosure exemplary embodiment;
Fig. 2 schematically shows the specific flow chart that user credit is assessed in disclosure exemplary embodiment;
Fig. 3 schematically shows a kind of block diagram of user credit assessment device in disclosure exemplary embodiment;
Fig. 4 schematically shows the block diagram of a kind of electronic equipment in disclosure exemplary embodiment;
Fig. 5 schematically shows a kind of program product in disclosure exemplary embodiment.
Specific embodiment
Example embodiment is described more fully with reference to the drawings.However, example embodiment can be with a variety of shapes
Formula is implemented, and is not understood as limited to example set forth herein;On the contrary, thesing embodiments are provided so that the disclosure will more
Fully and completely, and by the design of example embodiment comprehensively it is communicated to those skilled in the art.Described feature, knot
Structure or characteristic can be incorporated in any suitable manner in one or more embodiments.In the following description, it provides perhaps
More details fully understand embodiment of the present disclosure to provide.It will be appreciated, however, by one skilled in the art that can
It is omitted with technical solution of the disclosure one or more in the specific detail, or others side can be used
Method, constituent element, device, step etc..In other cases, be not shown in detail or describe known solution to avoid a presumptuous guest usurps the role of the host and
So that all aspects of this disclosure thicken.
In addition, attached drawing is only the schematic illustrations of the disclosure, it is not necessarily drawn to scale.Identical attached drawing mark in figure
Note indicates same or similar part, thus will omit repetition thereof.Some block diagrams shown in the drawings are function
Energy entity, not necessarily must be corresponding with physically or logically independent entity.These function can be realized using software form
Energy entity, or these functional entitys are realized in one or more hardware modules or integrated circuit, or at heterogeneous networks and/or place
These functional entitys are realized in reason device device and/or microcontroller device.
A kind of user credit appraisal procedure is provided firstly in this example embodiment, refering to what is shown in Fig. 1, to user credit
Appraisal procedure is described in detail.
In step s 110, multiple characteristic informations of target user are obtained, the multiple characteristic information includes first kind ginseng
Several and the second class parameter.
In the present exemplary embodiment, characteristic information refers to the corresponding data characteristics of the historical data of target user, specifically
It may include first kind parameter and the second class parameter.Wherein, IV value is the first kind parameter lower than the characteristic information of preset threshold,
The characteristic information that IV value is higher than the preset threshold is the second class parameter.Characteristic information can be the corresponding ginseng of each theme
Number, and each theme may include multiple characteristic informations.For example, characteristic information include but is not limited to the age, income, consumption data,
Browsing time, browsing time etc..
After obtaining multiple characteristic informations, characteristic information can be divided into first kind parameter and the second class parameter, specifically
Characteristic information can be divided according to IV value.IV (information value) value is referred to as the value of information, refers to
When constructing model by model methods such as logistic regression, decision trees, fine or not client is distinguished for measuring some characteristic information
IV value can be used to screen and classify all characteristic informations for one index of ability.In general, IV value is bigger, table
The value of information of bright characteristic information is bigger, therefore can be using the big characteristic information of IV value as being put into model, to intend model
Close training.
For example, it is assumed that in a classification problem, the classification of characteristic information has two classes: Y1, Y2.For one to pre-
The individual A of survey will judge that individual A belongs to Y1 or Y2, then need certain information.Assuming that this informational capacity is I, and these
Required information just lies in all characteristic information C1, C2, C3 ..., in Cn, then, for one of special
Reference ceases for Ci, and the information contained is more, then this feature information is got over for judging that A belongs to the contribution of Y1 or Y2
Greatly, the value of information of Ci is bigger, and the IV value of Ci is bigger, illustrates that the separating capacity of this feature information is better, then can be used
Characteristic information Ci establishes model.
First kind parameter includes weak parameter, such as weak Financial Attribute parameter;Second class parameter includes strong parameter, such as strong gold
Melt property parameters.In the present exemplary embodiment, preset threshold can be set to 0.2, can be low by the value of information in all characteristic informations
Characteristic information in 0.2 is determined as first kind parameter, and the characteristic information by the value of information in all characteristic informations greater than 0.2 determines
For the second class parameter.But preset threshold is not limited to above-mentioned numerical value, can also be configured according to actual needs.Except this it
Outside, can also using the value of information less than 0.05 characteristic information as extremely weak parameter.For extremely weak parameter, due to this feature information
Separating capacity it is very poor, therefore directly extremely weak parameter can be filtered out, the influence to avoid extremely weak parameter to entire assessment result.
In the step s 120, the first kind parameter and the second class parameter are pre-processed.
In the present exemplary embodiment, pretreatment refers to that branch mailbox is handled.Branch mailbox processing is packet transaction, and referring to will be continuous
Multi-mode discrete parameter is either merged into the discrete parameter of few state by parameter discretization.In the present exemplary embodiment, tool
Body can be used WOE algorithm and carry out branch mailbox processing.WOE (Weight of Evidence, evidence weight) is referred to initial parameter
A kind of coding form.Before carrying out WOE coding to a parameter, needs that this parameter is first carried out branch mailbox processing, specifically may be used
Including equidistant branch mailbox, etc. the modes such as deep branch mailbox, optimal branch mailbox.After carrying out branch mailbox, it can be calculated by formula (1) for i-th group
Weight evidence weight values WOE:
Wherein, pyiIt is the ratio that customer in response accounts for all customer in response in all samples in i-th group, pniBe in i-th group not
Customer in response accounts for the ratio of all non-customer in response in sample, #yiIt is the quantity of customer in response in i-th group, #niBe in i-th group not
The quantity of customer in response, #yTIt is the quantity of all customer in response in sample, #nTIt is the quantity of all non-customer in response in sample.
Customer in response herein refers to that Model Parameter value is 1 individual.
It follows that weight evidence weight values WOE expression is that " customer in response accounts for the ratio of all customer in response in current group
The difference of example " and " ratio for the client that the client Zhan Suoyou not responded in current group is not responded to ".Formula (1) is carried out
Transformation, can be obtained formula (2):
Wherein, a possibility that weight evidence weight values WOE is bigger, and difference is bigger, then shows sample responses in this grouping is bigger.
By carrying out branch mailbox processing to first kind parameter and the second class parameter, can to avoid parameter random error either
Anomaly parameter improves processing speed and efficiency to denoise to parameter.
In step s 130, the pretreated first kind parameter is converted, generates target component.
In the present exemplary embodiment, since first kind parameter can not be directly used in assessment credit evaluation, lead to a part
Characteristic information, which is unable to get, to be made full use of.In characteristic information there is no belong to strong Financial Attribute the second class parameter when, then without
Method carries out credit evaluation.In order to avoid the above problem, all first kind parameters can be converted, join the original first kind
Number is converted into target component.The detailed process of conversion refers to that feature combines, and target component refers to and the second class parameter association
Parameter, i.e., with strong Financial Attribute parameter.The parameter of the second class parameter association herein refers to and the second class parameter class
The identical parameter of type, for example, by multiple weak parameter combinations of script be strong parameter, so as to according to strong parameter and by weak parameter
The strong parameter of conversion carries out credit evaluation.It should be noted that multiple first kind parameters can be obtained by step S110, wherein often
A theme may also comprise multiple first kind parameters, such as 10,50 etc., not make special limit in the present exemplary embodiment to this
It is fixed.But pass through the feature combination in step S130, for each theme, it can only obtain a target component.Citing and
Speech, weak parameter 1 corresponding to theme 1, weak parameter 2, weak parameter 3, weak parameter 4 carry out feature combination, obtain the strong parameter of theme 1
1。
When carrying out feature combination, after linear discriminent algorithm all branch mailbox associated to each theme specifically can be used
First kind parameter carries out feature combination, to obtain the parameter of a second class parameter association.Theme for example may include trading, being clear
Look at, predict etc..For each theme, including first kind parameter and the second class parameter may be all different.For
All weak parameters are made to meet operational interpretation, it is therefore desirable to for the associated all first kind ginsengs of each theme
Number carries out feature combination.For example, the corresponding all first kind parameters of subject of transaction are combined, by the corresponding institute of browse themes
There is first kind parameter to be combined.Carry out feature combination by first kind parameter to each theme, avoid different themes it
Between parameter influence each other, so that the efficiency and accuracy rate of feature combination can be improved.
Linear discriminent algorithm LDA refers to that multiple weak parameters after branch mailbox are carried out linear combinations in assorting process,
To form a linear representation.It obtains including each weak parameter specifically, carrying out linear combination for multiple weak parameters
Linear representation, in assorting process on the feature space where linear representation by target component around different angular turns,
A best angle is obtained during rotation using linear discriminent algorithm, so that point of the target component on the best angle
Class gesture is maximum, so as to obtain strong parameter according to the classification maximum target component of gesture.One is obtained according to best angle in other words
A optimal linear combination, so that the classification gesture that optimum linear combines corresponding target component is maximum, so as to according to classification gesture
Maximum target component obtains strong parameter.Wherein, target component refers to any one in all weak parameters.Classification gesture refers to
It is the gesture for classification.
For example, for subject of transaction, weak parameter includes browsing time x1, evaluation y1, then can will browse secondary
Number x1 and evaluation y1 carries out linear combination and obtains linear representation Ax1+By1, and then can be in the feature where linear representation
The maximum best angle of gesture so that target component is classified spatially is obtained according to linear discriminent algorithm, and then by browsing time
X1, evaluation y1 group are combined into a strong parameter associated with browsing time and evaluation.It through the above way can will be under each theme
Weak parameter combination be strong parameter, thus using conversion strong parameter carry out model construction.In this way, relative to the relevant technologies
In for, the weak parameter building stronger risk score model of interpretation can be fully utilized, and can be included in it is more,
The more parameters of type.
After generating target component, target component and the second class parameter can be put into candidate variables pond.It is put into the time of model
The parameter for selecting first kind parameter and first kind parameter association in variable pond is all parameter after branch mailbox.It is wanted to meet this
It asks, after converting target component for first kind parameter, also needs to carry out branch mailbox again to these target components, and in branch mailbox again
Target component is put into candidate variables pond afterwards.For example may include in candidate variables pond strong parameter 1 after branch mailbox, after branch mailbox by weak
The strong parameter 4 etc. that parameter 2 and weak parameter 3 are combined into.
Next, the pretreated second class parameter and the target component are inputted machine in step S140
Learning model obtains the credit evaluation result of the target user.
In the present exemplary embodiment, machine learning model can be trained machine learning model, such as convolutional Neural
Network algorithm, deep learning algorithm etc., are illustrated by taking convolutional neural networks model as an example in the present exemplary embodiment.Convolution
Neural network model generally comprises input layer, mapping layer and output layer.
When by the pretreated second class parameter and target component input machine learning model, in order to guarantee
As a result accuracy, can the second class parameter to the target component in candidate variables pond and after branch mailbox screen.That is, can be to time
It selects all parameters in variable pond to be screened to obtain rest parameter, inputs machine learning for rest parameter as input parameter
Model, the output of the output layer of machine learning model can belong to the probability in some section for user credit, thus according to belonging to
The size of the probability in some section determines user credit assessment result, such as when the maximum probability for belonging to the good section of credit
When, determine that credit evaluation result is good.
Specifically, in order to guarantee the accuracy of result, the second class parameter and the target component in candidate variables pond can be rejected
Between multicollinearity.So-called multicollinearity refers between the parameter in linear regression model (LRM) due to there is accurate related pass
System or highly relevant relationship and make model estimation distortion or be difficult to estimate it is accurate.Weight evidence weight values generally should all be positive value, such as
There is negative value in the weight evidence weight values that fruit calculates, then it is contemplated that whether carrying out the influence of autoregressive parameter multicollinearity.It is based on
This, can determine between parameter according to the size relation between weight evidence weight values and preset value with the presence or absence of multicollinearity.It is default
Value for example can be 0, if weight evidence weight values are less than 0 (as negative value), then it is assumed that there are multicollinearities between parameter, thus
Can these parameters successively by weight evidence weight values less than 0 weed out, obtain rest parameter.
It, can be suitable according to the arrangement of the weight evidence weight values from small to large when rejecting the multicollinearity between parameter
Sequence, to the weight evidence weight values be less than the preset value branch mailbox after the second class parameter and again the target component after branch mailbox into
Row is rejected;Recalculate the second class parameter after the branch mailbox of rejecting and the again evidence weight of the target component after branch mailbox
Value;According to the weight evidence weight values putting in order from small to large, the preset value is less than to the weight evidence weight values recalculated
Branch mailbox after the second class parameter and the target component after branch mailbox is rejected again, until rejected the evidence weight
Until value is less than the second class parameter and target component of the preset value.
Such as it is -1 and the weight evidence weight values of parameter 3 that the weight evidence weight values of parameter 1, which are the weight evidence weight values of -3, parameter 2,
It is 1, then can weeds out the parameter 1 that weight evidence weight values are -3 for the first time, while linear combination is carried out to parameter 2 and parameter 3 again,
And recalculate the weight evidence weight values of each parameter.Next the parameter that weight evidence weight values are negative minimum can be weeded out.It removes
Except this, multicollinearity can also be rejected according to other algorithms.By successively rejecting a weight evidence weight values less than preset value
Parameter, weight evidence weight values then are recalculated to remaining all parameters, can more accurately reject all weight evidence weight values
Less than the parameter of preset value, so that more accurate rest parameter is obtained, to input trained machine learning by rest parameter
Model obtains more accurate credit evaluation as a result, to accurately measure consumer's risk.
The specific flow chart of determining credit evaluation result is shown in Fig. 2.Wherein:
In step s 201, feature extraction is carried out to data in modeling layer, obtains feature and summarize wide table, including multiple
Characteristic information.
In step S202, Feature Selection is carried out based on modeling layer and value of information IV, by multiple characteristic informations according to IV value
It is divided into strong parameter, weak parameter and extremely weak parameter, and extremely weak parameter is directly filtered out.
In step S203, branch mailbox processing is carried out to strong parameter and weak parameter using WOE algorithm, obtains strong parameter WOE points
Case and weak parameter WOE branch mailbox.
In step S204, feature combination is carried out to weak parameter, with specific reference to LDA linear discriminent algorithm to each theme
Corresponding multiple weak parameters carry out linear combination, obtain the corresponding strong parameter of each theme.For example, obtaining the LDA of browse themes
Combination, the LDA combination of subject of transaction, the combination of transaction credit card theme.It is further corresponding to each theme to be converted by weak parameter
At strong parameter carry out branch mailbox again, the WOE branch mailbox after obtaining LDA combination.
In step S205, LDA combination producing is passed through to the strong parameter after WOE branch mailbox and the weak parameter after WOE branch mailbox
Strong parameter rejects multicollinearity and obtains rest parameter, and rest parameter input machine learning model is obtained user credit assessment knot
Fruit.
Next, can be monitored to credit evaluation result.Such as ROC curve can be used, Gini coefficient AR value, distinguish energy
The modes such as power index KS value either Lorentz curve are monitored;In addition to this, PSI index also can be used to be monitored, with
Monitor the accuracy of assessment result.
By the step in Fig. 2, the weak parameter attribute group of each theme can be combined into strong parameter, thus based on strong parameter into
Row credit evaluation.Weak parameter can be fully utilized in this way and carry out the stronger risk score of interpretation, and can be with
More, the more parameters of type are included in, so that parameter is more acurrate, more comprehensively, so that credit evaluation result solves
The property released is stronger, so as to it is more acurrate, measure consumer's risk more in time.
The disclosure additionally provides a kind of user credit assessment device.Refering to what is shown in Fig. 3, the user credit assesses device 300
May include:
Feature obtains module 301, and for obtaining multiple characteristic informations of target user, the multiple characteristic information includes the
A kind of parameter and the second class parameter;
Parameter preprocessing module 302, for being pre-processed to the first kind parameter and the second class parameter;
Target component generation module 303 generates target ginseng for converting the pretreated first kind parameter
Number;
Assessment result determining module 304, for inputting the pretreated second class parameter and the target component
Machine learning model obtains the credit evaluation result of the target user;
Wherein, IV value is the first kind parameter lower than the characteristic information of preset threshold, and IV value is higher than the preset threshold
Characteristic information be the second class parameter.
In a kind of exemplary embodiment of the disclosure, parameter preprocessing module includes: branch mailbox processing module, for passing through
Weight evidence weight values carry out branch mailbox processing to the first kind parameter and the second class parameter respectively, the first kind parameter after obtaining branch mailbox
With the second class parameter after branch mailbox.
In a kind of exemplary embodiment of the disclosure, target component generation module includes: feature combination module, for adopting
Feature combination is carried out to the first kind parameter after the associated branch mailbox of each theme with linear discriminent algorithm, generates the target ginseng
Number.
In a kind of exemplary embodiment of the disclosure, described device further include: the first memory module, for the mesh
It marks parameter and carries out branch mailbox again, and the target component after branch mailbox again is put into candidate parameter pond;Second memory module, being used for will
The second class parameter after branch mailbox is put into the candidate variables pond.
In a kind of exemplary embodiment of the disclosure, assessment result determining module includes: that parameter rejects module, for picking
Except the second class parameter after branch mailbox in the candidate variables pond and again multicollinearity between the target component after branch mailbox, obtains
To rest parameter;Input control module, for the rest parameter to be inputted the machine learning model.
In a kind of exemplary embodiment of the disclosure, it includes: rejecting control module that parameter, which rejects module, is used for from described
The second class parameter and the target after branch mailbox again that weight evidence weight values are less than after the branch mailbox of preset value are rejected in candidate variables pond
Parameter obtains the rest parameter.
In a kind of exemplary embodiment of the disclosure, rejecting control module includes: to reject module for the first time, for according to institute
Weight evidence weight values putting in order from small to large is stated, the second class after the weight evidence weight values are less than with the branch mailbox of the preset value
Target component after parameter and again branch mailbox is rejected;Weight evidence weight values computing module, for recalculating by rejecting
The second class parameter after branch mailbox and again the weight evidence weight values of the target component after branch mailbox;Module is rejected again, for according to institute
Weight evidence weight values putting in order from small to large is stated, after the weight evidence weight values recalculated are less than with the branch mailbox of the preset value
The second class parameter and target component after branch mailbox is rejected again, until having rejected described in the weight evidence weight values are less than
Until the second class parameter and target component of preset value.
It should be noted that the detail of each module is believed in corresponding user in above-mentioned user credit assessment device
It is described in detail in appraisal procedure, therefore details are not described herein again.
It should be noted that although being referred to several modules or list for acting the equipment executed in the above detailed description
Member, but this division is not enforceable.In fact, according to embodiment of the present disclosure, it is above-described two or more
Module or the feature and function of unit can embody in a module or unit.Conversely, an above-described mould
The feature and function of block or unit can be to be embodied by multiple modules or unit with further division.
In addition, although describing each step of method in the disclosure in the accompanying drawings with particular order, this does not really want
These steps must be executed in this particular order by asking or implying, or having to carry out step shown in whole could realize
Desired result.Additional or alternative, it is convenient to omit multiple steps are merged into a step and executed by certain steps, and/
Or a step is decomposed into execution of multiple steps etc..
In an exemplary embodiment of the disclosure, a kind of electronic equipment that can be realized the above method is additionally provided.
Person of ordinary skill in the field it is understood that various aspects of the invention can be implemented as system, method or
Program product.Therefore, various aspects of the invention can be embodied in the following forms, it may be assumed that complete hardware embodiment, complete
The embodiment combined in terms of full Software Implementation (including firmware, microcode etc.) or hardware and software, can unite here
Referred to as circuit, " module " or " system ".
The electronic equipment 400 of this embodiment according to the present invention is described referring to Fig. 4.The electronics that Fig. 4 is shown
Equipment 400 is only an example, should not function to the embodiment of the present invention and use scope bring any restrictions.
As shown in figure 4, electronic equipment 400 is showed in the form of universal computing device.The component of electronic equipment 400 can wrap
It includes but is not limited to: at least one above-mentioned processing unit 410, at least one above-mentioned storage unit 420, the different system components of connection
The bus 430 of (including storage unit 420 and processing unit 410).
Wherein, the storage unit is stored with program code, and said program code can be held by the processing unit 410
Row, so that various according to the present invention described in the execution of the processing unit 410 above-mentioned " illustrative methods " part of this specification
The step of illustrative embodiments.For example, the processing unit 410 can execute step as shown in fig. 1.
Storage unit 420 may include the readable medium of volatile memory cell form, such as Random Access Storage Unit
(RAM) 4201 and/or cache memory unit 4202, it can further include read-only memory unit (ROM) 4203.
Storage unit 420 can also include program/utility with one group of (at least one) program module 4205
4204, such program module 4205 includes but is not limited to: operating system, one or more application program, other program moulds
It may include the realization of network environment in block and program data, each of these examples or certain combination.
Bus 430 can be to indicate one of a few class bus structures or a variety of, including storage unit bus or storage
Cell controller, peripheral bus, graphics acceleration port, processing unit use any bus structures in a variety of bus structures
Local bus.
Display unit 440 can be display having a display function, to pass through the display exhibits by processing unit 410
Execute processing result obtained from the method in the present exemplary embodiment.Display include but is not limited to liquid crystal display either
Other displays.
Electronic equipment 400 can also be with one or more external equipments 600 (such as keyboard, sensing equipment, bluetooth equipment
Deng) communication, can also be enabled a user to one or more equipment interact with the electronic equipment 400 communicate, and/or with make
Any equipment (such as the router, modulation /demodulation that the electronic equipment 400 can be communicated with one or more of the other calculating equipment
Device etc.) communication.This communication can be carried out by input/output (I/O) interface 450.Also, electronic equipment 400 can be with
By network adapter 460 and one or more network (such as local area network (LAN), wide area network (WAN) and/or public network,
Such as internet) communication.As shown, network adapter 460 is communicated by bus 430 with other modules of electronic equipment 400.
It should be understood that although not shown in the drawings, other hardware and/or software module can not used in conjunction with electronic equipment 400, including but not
Be limited to: microcode, device driver, redundant processing unit, external disk drive array, RAID system, tape drive and
Data backup storage system etc..
In an exemplary embodiment of the disclosure, a kind of computer readable storage medium is additionally provided, energy is stored thereon with
Enough realize the program product of this specification above method.In some possible embodiments, various aspects of the invention may be used also
In the form of being embodied as a kind of program product comprising program code, when described program product is run on the terminal device, institute
Program code is stated for executing the terminal device described in above-mentioned " illustrative methods " part of this specification according to this hair
The step of bright various illustrative embodiments.
Refering to what is shown in Fig. 5, describing the program product for realizing the above method of embodiment according to the present invention
500, can using portable compact disc read only memory (CD-ROM) and including program code, and can in terminal device,
Such as it is run on PC.However, program product of the invention is without being limited thereto, in this document, readable storage medium storing program for executing can be with
To be any include or the tangible medium of storage program, the program can be commanded execution system, device or device use or
It is in connection.
Described program product can be using any combination of one or more readable mediums.Readable medium can be readable letter
Number medium or readable storage medium storing program for executing.Readable storage medium storing program for executing for example can be but be not limited to electricity, magnetic, optical, electromagnetic, infrared ray or
System, device or the device of semiconductor, or any above combination.The more specific example of readable storage medium storing program for executing is (non exhaustive
List) include: electrical connection with one or more conducting wires, portable disc, hard disk, random access memory (RAM), read-only
Memory (ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, portable compact disc read only memory
(CD-ROM), light storage device, magnetic memory device or above-mentioned any appropriate combination.
Computer-readable signal media may include in a base band or as carrier wave a part propagate data-signal,
In carry readable program code.The data-signal of this propagation can take various forms, including but not limited to electromagnetic signal,
Optical signal or above-mentioned any appropriate combination.Readable signal medium can also be any readable Jie other than readable storage medium storing program for executing
Matter, the readable medium can send, propagate or transmit for by instruction execution system, device or device use or and its
The program of combined use.
The program code for including on readable medium can transmit with any suitable medium, including but not limited to wirelessly, have
Line, optical cable, RF etc. or above-mentioned any appropriate combination.
The program for executing operation of the present invention can be write with any combination of one or more programming languages
Code, described program design language include object oriented program language-Java, C++ etc., further include conventional
Procedural programming language-such as " C " language or similar programming language.Program code can be fully in user
It calculates and executes in equipment, partly executes on a user device, being executed as an independent software package, partially in user's calculating
Upper side point is executed on a remote computing or is executed in remote computing device or server completely.It is being related to far
Journey calculates in the situation of equipment, and remote computing device can pass through the network of any kind, including local area network (LAN) or wide area network
(WAN), it is connected to user calculating equipment, or, it may be connected to external computing device (such as utilize ISP
To be connected by internet).
In addition, above-mentioned attached drawing is only the schematic theory of processing included by method according to an exemplary embodiment of the present invention
It is bright, rather than limit purpose.It can be readily appreciated that the time that above-mentioned processing shown in the drawings did not indicated or limited these processing is suitable
Sequence.In addition, be also easy to understand, these processing, which can be, for example either synchronously or asynchronously to be executed in multiple modules.
Those skilled in the art after considering the specification and implementing the invention disclosed here, will readily occur to its of the disclosure
His embodiment.This application is intended to cover any variations, uses, or adaptations of the disclosure, these modifications, purposes or
Adaptive change follow the general principles of this disclosure and including the undocumented common knowledge in the art of the disclosure or
Conventional techniques.The description and examples are only to be considered as illustrative, and the true scope and spirit of the disclosure are by claim
It points out.
Claims (10)
1. a kind of user credit appraisal procedure characterized by comprising
Multiple characteristic informations of target user are obtained, the multiple characteristic information includes first kind parameter and the second class parameter;
The first kind parameter and the second class parameter are pre-processed;
The pretreated first kind parameter is converted, target component is generated;
The pretreated second class parameter and the target component are inputted into machine learning model, obtain the target user
Credit evaluation result;
Wherein, IV value is the first kind parameter lower than the characteristic information of preset threshold, and IV value is higher than the spy of the preset threshold
Reference breath is the second class parameter.
2. user credit appraisal procedure according to claim 1, which is characterized in that the first kind parameter and described
Two class parameters carry out pretreatment
Branch mailbox processing is carried out respectively to the first kind parameter and the second class parameter by weight evidence weight values, after obtaining branch mailbox
The second class parameter after a kind of parameter and branch mailbox.
3. user credit appraisal procedure according to claim 2, which is characterized in that join the pretreated first kind
Number is converted, and is generated target component and is included:
Feature combination is carried out to the first kind parameter after the associated branch mailbox of each theme using linear discriminent algorithm, described in generation
Target component.
4. user credit appraisal procedure according to claim 3, which is characterized in that the method also includes:
Branch mailbox again is carried out to the target component, and the target component after branch mailbox again is put into candidate parameter pond;
The second class parameter after branch mailbox is put into the candidate variables pond.
5. user credit appraisal procedure according to claim 4, which is characterized in that join pretreated second class
The several and target component inputs machine learning model
Reject the second class parameter in the candidate variables pond after branch mailbox and multiple total between the target component after branch mailbox again
Linearly, rest parameter is obtained;
The rest parameter is inputted into the machine learning model.
6. user credit appraisal procedure according to claim 5, which is characterized in that reject branch mailbox in the candidate variables pond
Rear the second class parameter and the again multicollinearity between the target component after branch mailbox, obtaining rest parameter includes:
The second class parameter after rejecting weight evidence weight values in the candidate variables pond and being less than the branch mailbox of preset value and again divide
Target component after case obtains the rest parameter.
7. user credit appraisal procedure according to claim 6, which is characterized in that reject card from the candidate variables pond
According to weighted value, less than the second class parameter after the branch mailbox of preset value and again, the target component after branch mailbox includes:
According to the weight evidence weight values putting in order from small to large, the weight evidence weight values are less than with the branch mailbox of the preset value
Rear the second class parameter and target component after branch mailbox is rejected again;
Recalculate the second class parameter after the branch mailbox of rejecting and again the weight evidence weight values of the target component after branch mailbox;
According to the weight evidence weight values putting in order from small to large, the preset value is less than to the weight evidence weight values recalculated
Branch mailbox after the second class parameter and the target component after branch mailbox is rejected again, until rejected the evidence weight
Until value is less than the second class parameter and target component of the preset value.
8. a kind of user credit assesses device characterized by comprising
Feature obtains module, and for obtaining multiple characteristic informations of target user, the multiple characteristic information includes first kind ginseng
Several and the second class parameter;
Parameter preprocessing module, for being pre-processed to the first kind parameter and the second class parameter;
Target component generation module generates target component for converting the pretreated first kind parameter;
Assessment result determining module, for the pretreated second class parameter and the target component to be inputted machine learning
Model obtains the credit evaluation result of the target user;
Wherein, IV value is the first kind parameter lower than the characteristic information of preset threshold, and IV value is higher than the spy of the preset threshold
Reference breath is the second class parameter.
9. a kind of electronic equipment characterized by comprising
Processor;And
Memory, for storing the executable instruction of the processor;
Wherein, the processor is configured to come described in perform claim requirement 1-7 any one via the execution executable instruction
User credit appraisal procedure.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
User credit appraisal procedure described in claim 1-7 any one is realized when being executed by processor.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811260889.1A CN109447461B (en) | 2018-10-26 | 2018-10-26 | User credit evaluation method and device, electronic equipment and storage medium |
CA3059937A CA3059937A1 (en) | 2018-10-26 | 2019-10-24 | User credit evaluation method and device, electronic device, storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811260889.1A CN109447461B (en) | 2018-10-26 | 2018-10-26 | User credit evaluation method and device, electronic equipment and storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109447461A true CN109447461A (en) | 2019-03-08 |
CN109447461B CN109447461B (en) | 2022-05-03 |
Family
ID=65548550
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811260889.1A Active CN109447461B (en) | 2018-10-26 | 2018-10-26 | User credit evaluation method and device, electronic equipment and storage medium |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN109447461B (en) |
CA (1) | CA3059937A1 (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110727510A (en) * | 2019-09-25 | 2020-01-24 | 浙江大搜车软件技术有限公司 | User data processing method and device, computer equipment and storage medium |
CN110782349A (en) * | 2019-10-25 | 2020-02-11 | 支付宝(杭州)信息技术有限公司 | Model training method and system |
CN110866696A (en) * | 2019-11-15 | 2020-03-06 | 成都数联铭品科技有限公司 | Method and device for training shop falling risk assessment model |
CN111709826A (en) * | 2020-06-11 | 2020-09-25 | 中国建设银行股份有限公司 | Target information determination method and device |
CN111815457A (en) * | 2020-07-01 | 2020-10-23 | 北京金堤征信服务有限公司 | Target object evaluation method and device |
CN111950889A (en) * | 2020-08-10 | 2020-11-17 | 中国平安人寿保险股份有限公司 | Client risk assessment method and device, readable storage medium and terminal equipment |
CN112529477A (en) * | 2020-12-29 | 2021-03-19 | 平安普惠企业管理有限公司 | Credit evaluation variable screening method, device, computer equipment and storage medium |
CN112734568A (en) * | 2021-01-29 | 2021-04-30 | 深圳前海微众银行股份有限公司 | Credit scoring card model construction method, device, equipment and readable storage medium |
CN112734433A (en) * | 2020-12-10 | 2021-04-30 | 深圳市欢太科技有限公司 | Abnormal user detection method and device, electronic equipment and storage medium |
CN113570066A (en) * | 2021-07-23 | 2021-10-29 | 中国恩菲工程技术有限公司 | Data processing method, system, electronic device and storage medium |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100290700A1 (en) * | 2009-05-13 | 2010-11-18 | Jun Yokono | Information processing device and method, learning device and method, programs, and information processing system |
US20150178754A1 (en) * | 2013-12-19 | 2015-06-25 | Microsoft Corporation | Incentive system for interactive content consumption |
CN104866484A (en) * | 2014-02-21 | 2015-08-26 | 阿里巴巴集团控股有限公司 | Data processing method and device |
US20170337487A1 (en) * | 2014-10-24 | 2017-11-23 | National Ict Australia Limited | Learning with transformed data |
CN108230067A (en) * | 2016-12-14 | 2018-06-29 | 阿里巴巴集团控股有限公司 | The appraisal procedure and device of user credit |
CN108399255A (en) * | 2018-03-06 | 2018-08-14 | 中国银行股份有限公司 | A kind of input data processing method and device of Classification Data Mining model |
-
2018
- 2018-10-26 CN CN201811260889.1A patent/CN109447461B/en active Active
-
2019
- 2019-10-24 CA CA3059937A patent/CA3059937A1/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100290700A1 (en) * | 2009-05-13 | 2010-11-18 | Jun Yokono | Information processing device and method, learning device and method, programs, and information processing system |
US20150178754A1 (en) * | 2013-12-19 | 2015-06-25 | Microsoft Corporation | Incentive system for interactive content consumption |
CN104866484A (en) * | 2014-02-21 | 2015-08-26 | 阿里巴巴集团控股有限公司 | Data processing method and device |
US20170337487A1 (en) * | 2014-10-24 | 2017-11-23 | National Ict Australia Limited | Learning with transformed data |
CN108230067A (en) * | 2016-12-14 | 2018-06-29 | 阿里巴巴集团控股有限公司 | The appraisal procedure and device of user credit |
CN108399255A (en) * | 2018-03-06 | 2018-08-14 | 中国银行股份有限公司 | A kind of input data processing method and device of Classification Data Mining model |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110727510A (en) * | 2019-09-25 | 2020-01-24 | 浙江大搜车软件技术有限公司 | User data processing method and device, computer equipment and storage medium |
CN110782349A (en) * | 2019-10-25 | 2020-02-11 | 支付宝(杭州)信息技术有限公司 | Model training method and system |
CN110866696A (en) * | 2019-11-15 | 2020-03-06 | 成都数联铭品科技有限公司 | Method and device for training shop falling risk assessment model |
CN111709826A (en) * | 2020-06-11 | 2020-09-25 | 中国建设银行股份有限公司 | Target information determination method and device |
CN111815457A (en) * | 2020-07-01 | 2020-10-23 | 北京金堤征信服务有限公司 | Target object evaluation method and device |
CN111950889A (en) * | 2020-08-10 | 2020-11-17 | 中国平安人寿保险股份有限公司 | Client risk assessment method and device, readable storage medium and terminal equipment |
CN112734433A (en) * | 2020-12-10 | 2021-04-30 | 深圳市欢太科技有限公司 | Abnormal user detection method and device, electronic equipment and storage medium |
CN112529477A (en) * | 2020-12-29 | 2021-03-19 | 平安普惠企业管理有限公司 | Credit evaluation variable screening method, device, computer equipment and storage medium |
CN112734568A (en) * | 2021-01-29 | 2021-04-30 | 深圳前海微众银行股份有限公司 | Credit scoring card model construction method, device, equipment and readable storage medium |
CN112734568B (en) * | 2021-01-29 | 2024-01-12 | 深圳前海微众银行股份有限公司 | Credit scoring card model construction method, device, equipment and readable storage medium |
CN113570066A (en) * | 2021-07-23 | 2021-10-29 | 中国恩菲工程技术有限公司 | Data processing method, system, electronic device and storage medium |
CN113570066B (en) * | 2021-07-23 | 2024-03-29 | 中国恩菲工程技术有限公司 | Data processing method, system, electronic device and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CA3059937A1 (en) | 2020-04-26 |
CN109447461B (en) | 2022-05-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109447461A (en) | User credit appraisal procedure and device, electronic equipment, storage medium | |
US10929614B2 (en) | Automated contextual dialog generation for cognitive conversation | |
US20190005586A1 (en) | Prediction algorithm based attribute data processing | |
CN112037760B (en) | Training method and device of voice spectrum generation model and electronic equipment | |
US11874861B2 (en) | Retraining a conversation system based on negative feedback | |
CN110472675A (en) | Image classification method, image classification device, storage medium and electronic equipment | |
US20200151555A1 (en) | Detecting and reducing bias in machine learning models | |
CN109344906A (en) | Consumer's risk classification method, device, medium and equipment based on machine learning | |
US11182447B2 (en) | Customized display of emotionally filtered social media content | |
CN110110062A (en) | Machine intelligence answering method, device and electronic equipment | |
CN105719073A (en) | Enterprise credit evaluation system and method | |
CN109034206A (en) | Image classification recognition methods, device, electronic equipment and computer-readable medium | |
US11237806B2 (en) | Multi objective optimization of applications | |
CN110059112A (en) | Usage mining method and device based on machine learning, electronic equipment, medium | |
US20200366716A1 (en) | Adaptive identity broker for governance of decentralized identities across multiple heterogeneous identity networks | |
US20220147895A1 (en) | Automated data forecasting using machine learning | |
CN115455171B (en) | Text video mutual inspection rope and model training method, device, equipment and medium | |
CN109947924A (en) | Conversational system training data construction method, device, electronic equipment and storage medium | |
CN115759748A (en) | Risk detection model generation method and device and risk individual identification method and device | |
CN109034199A (en) | Data processing method and device, storage medium and electronic equipment | |
US20180349476A1 (en) | Evaluating theses using tree structures | |
CN114708539A (en) | Image type identification method and device, equipment, medium and product thereof | |
CN110009219A (en) | Financial big data processing method, storage medium and equipment | |
CN109635168A (en) | Chart output method and device | |
CN113657538A (en) | Model training method, data classification method, device, equipment, storage medium and product |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |