CN109598410A - Presell methods of risk assessment, system, computer installation and readable storage medium storing program for executing - Google Patents
Presell methods of risk assessment, system, computer installation and readable storage medium storing program for executing Download PDFInfo
- Publication number
- CN109598410A CN109598410A CN201811289813.1A CN201811289813A CN109598410A CN 109598410 A CN109598410 A CN 109598410A CN 201811289813 A CN201811289813 A CN 201811289813A CN 109598410 A CN109598410 A CN 109598410A
- Authority
- CN
- China
- Prior art keywords
- presell
- information
- assessed value
- trade company
- input
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- 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
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Educational Administration (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention provides a kind of presell methods of risk assessment, system, computer installation and computer readable storage medium.The presell methods of risk assessment includes: to obtain history feature data, and the history feature data are that trade company has initiated presell behavior and the data with presell result；History feature data corresponding to different presell results are separately input into neural network model to be trained to obtain the first risk evaluation model and the second risk evaluation model；A pending movable characteristic of presell is obtained, and the characteristic is separately input into first risk evaluation model and the second risk evaluation model, obtains the first assessed value and the second assessed value；And the movable success rate of pending presell is calculated according to first assessed value and second assessed value.The present invention is based on neural metwork trainings to obtain presell risk evaluation model, and trade company's presell success rate can be calculated according to model, avoid user's economic loss.
Description
Technical field
The present invention relates to data processing field more particularly to a kind of presell methods of risk assessment, system, computer installation and
Computer readable storage medium.
Background technique
Presell refers to the sales behavior carried out before product does not also formally enter market, can understand this kind by presell
Whether product has market, especially for it is some can only be by the product of mass production for, reach a certain amount of by presell
It just puts into production afterwards, has effectively evaded the existing risk of production.Present businessman likes doing presell in advance before commercial articles vending
Marketing activity, to attract customer.All there is certain risk in a kind of either that businessman of product, operation, once businessman passes through
Battalion is not good at or businessman deliberately swindles, it is easy to cause the economic loss of user.
This part intends to provides background for the embodiments of the present invention stated in claims and specific embodiment
Or context.Description herein recognizes it is the prior art not because not being included in this section.
Summary of the invention
In view of above-mentioned, the present invention provides a kind of presell methods of risk assessment, system, computer installation and computer-readable deposits
Storage media may be implemented to carry out risk profile to trade company's presell behavior.
One embodiment of the application provides a kind of presell methods of risk assessment, which comprises
History feature data are obtained, wherein the history feature data are that trade company has initiated presell behavior and had presell
As a result data, the presell result include presell success and presell failure；
Classify according to different presell results to the history feature data；
History feature data corresponding to different presell results are separately input into neural network model to be trained, with
It respectively obtains corresponding to successful first risk evaluation model of presell and corresponding to the second risk evaluation model of presell failure；
A pending movable characteristic of presell is obtained, and the characteristic is input to first risk assessment
Model obtains the first assessed value, and the characteristic is input to second risk evaluation model and obtains the second assessed value；
And
The movable success rate of pending presell is calculated according to first assessed value and second assessed value.
Preferably, the history feature data include multiple trade company's dimensional informations, and multiple trade company's dimensional informations include
But it is not limited to: urban information, shops where Merchants register fund information, shareholder number information, shareholder's credit record information, shops
Location information, the affiliated trade information of trade company, the preferential dynamics information of commodity, grace period information, merchandise cost information, opening time
Information, trade company's profit or loss information, customer quantity information.
Preferably, after the step of acquisition history feature data further include:
Judge whether there is one or more trade company's dimensional information missing；And
When there are one or more trade company's dimensional information missing, one or more trade company's dimensional information of missing is set
It is set to default characteristic.
Preferably, described that history feature data corresponding to different presell results are separately input into neural network model
It is trained, corresponds to successful first risk evaluation model of presell and corresponding to the second risk of presell failure to respectively obtain
The step of assessment models includes:
Presell result is input to first nerves network model for the successful history feature data of presell to be trained, is obtained
First risk evaluation model；
The history feature data that presell result is presell failure are input to nervus opticus network model to be trained, are obtained
Second risk evaluation model；
Wherein, the first nerves network model and the nervus opticus network model are BP neural network model, institute
Stating BP neural network model includes input layer, hidden layer and output layer.
Preferably, the input layer includes n node, and the hidden layer includes m node, the BP neural network model
Are as follows:
Wherein, y is the output valve of the output layer, when the characteristic is input to first risk evaluation model
When, the output valve y of first risk evaluation model is first assessed value, when the characteristic is input to described the
When two risk evaluation models, the output valve y of second risk evaluation model is second assessed value, tiFor the hidden layer
With the connection weight between the output layer, It is described hidden
Hide the input of layer, the output of the as described input layer, WijFor the connection weight between the input layer and the hidden layer；f
(Si) be the BP neural network model in activation primitive,
Preferably, described that the pending presell work is calculated according to first assessed value and second assessed value
The step of dynamic success rate includes:
Obtain corresponding first weight coefficient of first assessed value and the second assessed value corresponding second weight system
Number；And
The product of first assessed value and first weight coefficient is subtracted into second assessed value and described second
The product of weight coefficient obtains the movable success rate of pending presell.
Preferably, described that the pending presell work is calculated according to first assessed value and second assessed value
After the step of dynamic success rate further include:
Judge whether the movable success rate of pending presell is greater than preset threshold；
When the pending movable success rate of presell is greater than the preset threshold, the first prompt information is exported；And
When the pending movable success rate of presell is not more than the preset threshold, the second prompt information is exported.
One embodiment of the application provides a kind of presell risk evaluating system, the system comprises:
First obtains module, is used for history feature data, wherein the history feature data are that trade company has initiated presell
Behavior and the data with presell result, the presell result include presell success and presell failure；
Model training module, for history feature data corresponding to different presell results to be separately input into nerve net
Network model is trained, and corresponds to successful first risk evaluation model of presell and corresponding to the of presell failure to respectively obtain
Two risk evaluation models；
Second obtains module, for obtaining a pending movable characteristic of presell, and the characteristic is inputted
The first assessed value is obtained to first risk evaluation model, and the characteristic is input to the second risk assessment mould
Type obtains the second assessed value；And
Computing module, for the pending presell to be calculated according to first assessed value and second assessed value
Movable success rate.
One embodiment of the application provides a kind of computer installation, and the computer installation includes processor and memory,
Several computer programs are stored on the memory, the processor is for when executing the computer program stored in memory
The step of realizing presell methods of risk assessment as elucidated before.
One embodiment of the application provides a kind of computer readable storage medium, is stored thereon with computer program, described
The step of presell methods of risk assessment as elucidated before is realized when computer program is executed by processor.
Above-mentioned presell methods of risk assessment, system, computer installation and computer readable storage medium are based on machine learning
Establish and train to obtain risk evaluation model with history trade company presell data, and by a pending movable trade company's feature of presell
Data are input to the risk evaluation model and are calculated the movable success rate of trade company's presell, so user can according to this at
Power judges whether to need to participate in the presell activity of the trade company, avoids economic loss.
Detailed description of the invention
It, below will be to required in embodiment description in order to illustrate more clearly of the technical solution of embodiment of the present invention
The attached drawing used is briefly described, it should be apparent that, the accompanying drawings in the following description is 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 step flow chart of presell methods of risk assessment in one embodiment of the invention.
Fig. 2 is the step flow chart of presell methods of risk assessment in another embodiment of the present invention.
Fig. 3 is the functional block diagram of presell risk evaluating system in one embodiment of the invention.
Fig. 4 is computer schematic device in one embodiment of the invention.
Specific embodiment
To better understand the objects, features and advantages of the present invention, with reference to the accompanying drawing and specific real
Applying mode, the present invention will be described in detail.It should be noted that in the absence of conflict, presently filed embodiment and reality
The feature applied in mode can be combined with each other.
In the following description, numerous specific details are set forth in order to facilitate a full understanding of the present invention, described embodiment
Only some embodiments of the invention, rather than whole embodiments.Based on the embodiment in the present invention, this field
Those of ordinary skill's every other embodiment obtained without making creative work, belongs to guarantor of the present invention
The range of shield.
Unless otherwise defined, all technical and scientific terms used herein and belong to technical field of the invention
The normally understood meaning of technical staff is identical.Term as used herein in the specification of the present invention is intended merely to description tool
The purpose of the embodiment of body, it is not intended that in the limitation present invention.
Preferably, presell methods of risk assessment of the invention is applied in one or more computer installation.The meter
Calculation machine device be it is a kind of can be according to the instruction for being previously set or store, automatic progress numerical value calculating and/or information processing are set
Standby, hardware includes but is not limited to microprocessor, specific integrated circuit (Application Specific Integrated
Circuit, ASIC), programmable gate array (Field-Programmable Gate Array, FPGA), digital processing unit
(Digital Signal Processor, DSP), embedded device etc..
The computer installation can be the calculating such as desktop PC, laptop, tablet computer, server and set
It is standby.The computer installation can carry out people by modes such as keyboard, mouse, remote controler, touch tablet or voice-operated devices with user
Machine interaction.
Embodiment one:
Fig. 1 is the step flow chart of presell methods of risk assessment preferred embodiment of the present invention.It is described according to different requirements,
The sequence of step can change in flow chart, and certain steps can be omitted.
As shown in fig.1, the presell methods of risk assessment specifically includes following steps.
Step S11, obtain history feature data, wherein the history feature data be trade company initiated presell behavior and
Data with presell result, the presell result include presell success and presell failure.
In one embodiment, a merchant data sample database can be connected to by access network, so come obtain with
The relevant history feature data of trade company's presell.The merchant data sample database can be collected by big data mode multiple trade companies into
The movable history feature data of presell of going can also pass through the history feature number relevant to trade company's presell of the artificial typing of reception
According to.
In one embodiment, history feature data relevant to trade company's presell include but is not limited to multiple trade company's dimension letters
Breath, multiple trade company's dimensional informations include but is not limited to: Merchants register fund information, shareholder number information, shareholder's credit note
Record information, urban information where shops, shops's location information, the affiliated trade information of trade company, the preferential dynamics information of commodity, it is preferential when
Between information, merchandise cost information, opening temporal information, trade company profit or loss information, customer quantity information.
It in one embodiment, can also be to acquiring after obtaining relevant to trade company's presell history feature data
Characteristic carries out pretreatment operation.It is to be appreciated that may have one or more missings in the history feature data
Trade company's dimensional information of data information can be by corresponding missing data information in order to improve the accuracy of follow-up data processing result
Trade company's dimensional information be set as default default value.For example, for trade company's dimensional information is the preferential dynamics information of commodity, if
The history feature data of one trade company do not have the preferential dynamics information of commodity, and it is preferential to indicate that the trade company does not have when carrying out presell activity,
At this point it is possible to set default default value (assuming that default default value is 0) for the preferential dynamics information of the commodity of the trade company.
It is to be appreciated that history feature data relevant to trade company's presell can be obtained from the merchant data sample database of structuring
It takes, the history feature data acquired at this time can directly carry out subsequent processing.In other embodiments, with trade company's presell
Relevant history feature data are also possible to may be non-structured data, the data being such as stored in text.Thus, right
When the history feature data are pre-processed, corresponding characteristic can be also obtained from non-structured data.
In one embodiment, when the history feature data are non-structured data, it can first define and be tieed up with trade company
Spend the corresponding keyword of information, trade company's dimensional information as described in multiple include: Merchants register fund information, shareholder number information,
Urban information, shops's location information, the affiliated trade information of trade company, the preferential dynamics of commodity where shareholder's credit record information, shops
Information, grace period information, merchandise cost information, opening temporal information；Then keyword corresponding with those trade company's dimensional informations
Are as follows: registered capital, shareholder's number, shareholder's credit record, city, position, industry, preferential dynamics, grace period, cost, opening
Time.For example, non-structured data include " grace period is on January 1st, 2017 ", it is " preferential since it comprises keywords
Time ", thus it is " on January 1st, 2017 " that corresponding characteristic can be obtained according to keyword " grace period ".
Step S12, by history feature data corresponding to different presell results be separately input into neural network model into
Row training is commented with respectively obtaining the second risk for corresponding to successful first risk evaluation model of presell and failing corresponding to presell
Estimate model.
In one embodiment, can by history feature data according to presell result be presell success and presell unsuccessfully come into
Row classification makees the history feature data of presell failure using the successful history feature data of presell as first sample data set
For the second sample data set.
In one embodiment, first risk evaluation model, second risk evaluation model can be based on mind
The model come is trained through network model and first sample data set, the second sample data set.The neural network model can
Human brain neuroid is abstracted from information processing angle, different networks is formed by different connection types, is not necessarily to thing
It first determines the math equation of mapping relations between input and output, only by the training of itself, is obtained most in given input value
Close to the result of desired output.The neural network model includes input layer, hidden layer and output layer.The history feature number
According to the input layer that can be used as neural network model, one is exported by output layer using after the connection of the hidden layer of neural network model
Assessed value.
It in one embodiment, can be by presell result since the presell result includes that presell success and presell fail
First nerves network model is input to for the successful history feature data of presell to be trained, and obtains the first risk assessment mould
The history feature data that presell result is presell failure are input to nervus opticus network model and are trained, obtained described by type
Second risk evaluation model.The neural network model can be BP (Back Propagation, backpropagation) neural network
Model, the BP neural network model are a kind of multilayer feedforword nets by error back propagation training based on gradient descent method
Network, using gradient search technology, to make the real output value of network and the error mean square difference minimum of desired output.At this
In the other embodiments of invention, other kinds of neural network model can also be selected according to actual needs.
In one embodiment, the input layer includes n node, and the hidden layer includes m node, the BP nerve
Network model can indicate are as follows:
Wherein, y is the output valve of the output layer, when the characteristic is input to first risk evaluation model
When, the output valve y of first risk evaluation model is first assessed value, when the characteristic is input to described second
When risk evaluation model, the output valve y of second risk evaluation model is second assessed value, tiFor the hidden layer and institute
The connection weight between output layer is stated, SiIt is described hidden
Hide the input of layer, the output of the as described input layer, WijFor the connection weight between the input layer and the hidden layer, f ()
For the activation primitive in the BP neural network model, when the hidden layer has input, activation primitive is expressed as f (Si).It is described
Activation primitive f (Si) S type function (Sigmoid function), f (S can be usedi) can indicate are as follows:
It is to be appreciated that tiIt can indicate the connection weight between i-th of node of the hidden layer and the output layer,
For example, t11st connection weight between node and the output layer of the as described hidden layer, t2The as described hidden layer
2nd connection weight between node and the output layer, t33rd node of the as described hidden layer and the output layer it
Between connection weight, tmConnection weight between m-th of the node and the output layer of the as described hidden layer.Similarly it is found that
WijConnection weight between i-th of node of the as described hidden layer and j-th of node of the input layer.By to the BP
Neural network model is trained, and can be correspondingly made available tiAnd WijValue, i.e., training obtain each layer of BP neural network model
Parameter, risk evaluation model so can be obtained.
Step S13, a pending movable characteristic of presell is obtained, and the characteristic is input to described first
Risk evaluation model obtains the first assessed value, and the characteristic is input to second risk evaluation model and obtains second
Assessed value.
In one embodiment, when a trade company A prepares to carry out presell activity, the characteristic of available trade company A,
The characteristic of trade company A includes Merchants register fund information, shareholder number information, shareholder's credit record information, shops place
Urban information, shops's location information, the affiliated trade information of trade company, the preferential dynamics information of commodity, grace period information, merchandise cost
Information, opening temporal information, trade company's profit or loss information, customer quantity information.First risk evaluation model is to be based on
The successful history feature data of presell and construct, therefore, the characteristic of trade company A is input to first risk assessment
Model obtains the first assessed value, and first assessed value can be used to predict the successful probability of trade company A presell.Second wind
Dangerous assessment models are constructed based on the history feature data that presell fails, and therefore, the characteristic of trade company A is input to
Second risk evaluation model obtains the second assessed value, and second assessed value can be used to predict that trade company A presell fails
Probability.
Step S14, the pending presell activity is calculated according to first assessed value and second assessed value
Success rate.
In one embodiment, the first power can be respectively set for first assessed value and second assessed value in advance
Weight coefficient and the second weight coefficient, wherein first assessed value corresponds to the first weight coefficient, second assessed value corresponding the
Two weight coefficients, then by the product of first assessed value and first weight coefficient subtract second assessed value with it is described
The product of second weight coefficient obtain the pending movable success rate of presell of trade company A (calculation formula may is that success rate=
First the second weight coefficient of assessed value * the first weight coefficient the-the second assessed value *).
Please refer to Fig. 2, compared with presell methods of risk assessment shown in fig. 1, Fig. 2 shows presell risk assessment side
Method further includes step S15, S16 and S17.
Step S15, judges whether the movable success rate of pending presell is greater than preset threshold；
Step S16, when the pending movable success rate of presell is greater than the preset threshold, output the first prompt letter
Breath；
Step S17, when the pending movable success rate of presell is not more than the preset threshold, the second prompt of output
Information.
In one embodiment, when the pending movable success rate of presell is greater than the preset threshold, show this
The probalility of success of trade company's A presell is greater than failure probability, and trade company's business risk is lower, and the presell that user can participate in trade company A is living
It is dynamic, when the pending movable success rate of presell is not more than the preset threshold, show the probalility of success of trade company A presell
Less than failure probability, trade company's business risk is larger, it is not recommended that user participates in the presell activity of trade company A.The preset threshold can
To be set and be adjusted according to actual use demand, for example, the preset threshold is set as 0.5, if trade company A is calculated
The success rate of presell is greater than 0.5, then it represents that the probalility of success of the trade company A presell is greater than failure probability, output the first prompt letter
It ceases to user, first prompt information may is that presell risk is lower, and the presell activity for participating in trade company A may be selected；If meter
The success rate for obtaining trade company A presell is calculated no more than 0.5, then it represents that the probalility of success of the trade company A presell is less than failure probability,
The second prompt information is exported to user, second prompt information may is that presell risk is higher, it is not recommended that participate in trade company A
Presell activity.
Embodiment two:
Fig. 3 is the functional block diagram of presell risk evaluating system preferred embodiment of the present invention.
As shown in fig.2, the presell risk evaluating system 10 may include the first acquisition module 101, model training mould
Block 102, second obtains module 103, computing module 104 and output module 105.
The acquisition module 101 be used for history feature data, wherein the history feature data be trade company initiated it is pre-
Behavior and the data with presell result are sold, the presell result includes presell success and presell failure.
In one embodiment, the acquisition module 101 can be connected to a merchant data sample by access network
Library, and then to obtain history feature data relevant to trade company's presell.The merchant data sample database can be by big data side
Formula collects multiple trade companies and carried out the movable history feature data of presell, can also be by receiving artificial typing and trade company's presell
Relevant history feature data.
In one embodiment, history feature data relevant to trade company's presell include but is not limited to multiple trade company's dimension letters
Breath, multiple trade company's dimensional informations include but is not limited to: Merchants register fund information, shareholder number information, shareholder's credit note
Record information, urban information where shops, shops's location information, the affiliated trade information of trade company, the preferential dynamics information of commodity, it is preferential when
Between information, merchandise cost information, opening temporal information, trade company profit or loss information, customer quantity information.
In one embodiment, the acquisition module 101 is after obtaining relevant to trade company's presell history feature data,
Pretreatment operation can also be carried out to characteristic is acquired.It is to be appreciated that may be deposited in the history feature data
In trade company's dimensional information of one or more missing data information, in order to improve the accuracy of follow-up data processing result, can incite somebody to action
Trade company's dimensional information of corresponding missing data information is set as default default value.For example, being that commodity are excellent for trade company's dimensional information
For favour dynamics information, if the history feature data of a trade company do not have the preferential dynamics information of commodity, indicate that the trade company is carrying out in advance
Do not have when selling activity it is preferential, at this point it is possible to set default default value (assuming that default for the preferential dynamics information of the commodity of the trade company
0) default value is.
It is to be appreciated that history feature data relevant to trade company's presell can be obtained from the merchant data sample database of structuring
It takes, the history feature data acquired at this time can directly carry out subsequent processing.In other embodiments, with trade company's presell
Relevant history feature data are also possible to may be non-structured data, the data being such as stored in text.Thus, right
When the history feature data are pre-processed, corresponding characteristic can be also obtained from non-structured data.
In one embodiment, when the history feature data are non-structured data, it can first define and be tieed up with trade company
Spend the corresponding keyword of information, trade company's dimensional information as described in multiple include: Merchants register fund information, shareholder number information,
Urban information, shops's location information, the affiliated trade information of trade company, the preferential dynamics of commodity where shareholder's credit record information, shops
Information, grace period information, merchandise cost information, opening temporal information；Then keyword corresponding with those trade company's dimensional informations
Are as follows: registered capital, shareholder's number, shareholder's credit record, city, position, industry, preferential dynamics, grace period, cost, opening
Time.For example, non-structured data include " grace period is on January 1st, 2017 ", it is " preferential since it comprises keywords
Time ", thus it is " on January 1st, 2017 " that corresponding characteristic can be obtained according to keyword " grace period ".
History feature data corresponding to different presell results for being separately input by the model training module 102
Neural network model is trained, and is lost with respectively obtaining to correspond to successful first risk evaluation model of presell and correspond to presell
The second risk evaluation model lost.
In one embodiment, the history feature data can be that presell success and presell unsuccessfully come according to presell result
Classify, using the successful history feature data of presell as first sample data set, by the history feature data of presell failure
As the second sample data set.
In one embodiment, first risk evaluation model, second risk evaluation model can be based on mind
The model come is trained through network model and first sample data set, the second sample data set.The neural network model can
Human brain neuroid is abstracted from information processing angle, different networks is formed by different connection types, is not necessarily to thing
It first determines the math equation of mapping relations between input and output, only by the training of itself, is obtained most in given input value
Close to the result of desired output.The neural network model includes input layer, hidden layer and output layer.The history feature number
According to the input layer that can be used as neural network model, one is exported by output layer using after the connection of the hidden layer of neural network model
Assessed value.
In one embodiment, since the presell result includes presell success and presell failure, the model training mould
Presell result can be input to first nerves network model for the successful history feature data of presell and is trained by block 102, be obtained
To first risk evaluation model, the history feature data that presell result is presell failure are input to nervus opticus network mould
Type is trained, and obtains second risk evaluation model.The neural network model can be BP (Back
Propagation, backpropagation) neural network model, the BP neural network model be it is a kind of based on gradient descent method by
The Multi-layered Feedforward Networks of error back propagation training, using gradient search technology, to make the real output value and expectation of network
The error mean square difference of output valve is minimum.In other embodiments of the invention, other can also be selected according to actual needs
The neural network model of type.
In one embodiment, the input layer includes n node, and the hidden layer includes m node, the BP nerve
Network model can indicate are as follows:
Wherein, y is the output valve of the output layer, when the characteristic is input to first risk evaluation model
When, the output valve y of first risk evaluation model is first assessed value, when the characteristic is input to described second
When risk evaluation model, the output valve y of second risk evaluation model is second assessed value, tiFor the hidden layer and institute
The connection weight between output layer is stated, SiIt is described hidden
Hide the input of layer, the output of the as described input layer, WijFor the connection weight between the input layer and the hidden layer, f ()
For the activation primitive in the BP neural network model, when the hidden layer has input, activation primitive is expressed as f (Si).It is described
Activation primitive f (Si) S type function (Sigmoid function), f (S can be usedi) can indicate are as follows:
It is to be appreciated that tiIt can indicate the connection weight between i-th of node of the hidden layer and the output layer,
For example, t11st connection weight between node and the output layer of the as described hidden layer, t2The as described hidden layer
2nd connection weight between node and the output layer, t33rd node of the as described hidden layer and the output layer it
Between connection weight, tmConnection weight between m-th of the node and the output layer of the as described hidden layer.Similarly it is found that
WijConnection weight between i-th of node of the as described hidden layer and j-th of node of the input layer.By to the BP
Neural network model is trained, and can be correspondingly made available tiAnd WijValue, i.e., training obtain each layer of BP neural network model
Parameter, risk evaluation model so can be obtained.
Described second obtains module 103 for obtaining a pending movable characteristic of presell, and by the characteristic
The first assessed value is obtained according to first risk evaluation model is input to, and the characteristic is input to second risk
Assessment models obtain the second assessed value.
In one embodiment, when a trade company A prepares to carry out presell activity, the second acquisition module 103 is available should
The characteristic of trade company A specifically can be acquisition module 103 and receive the characteristic that user inputs trade company A.The quotient
The characteristic of family A includes city where Merchants register fund information, shareholder number information, shareholder's credit record information, shops
Information, shops's location information, the affiliated trade information of trade company, the preferential dynamics information of commodity, grace period information, merchandise cost letter
Breath, opening temporal information, trade company's profit or loss information, customer quantity information.First risk evaluation model is based on pre-
It sells successful history feature data and constructs, therefore, the characteristic of trade company A is input to the first risk assessment mould
Type obtains the first assessed value, and first assessed value can be used to predict the successful probability of trade company A presell.Second risk
Assessment models are constructed based on the history feature data that presell fails, and therefore, the characteristic of trade company A are input to institute
It states the second risk evaluation model and obtains the second assessed value, second assessed value can be used to predict trade company A presell failure
Probability.
The computing module 104 be used to be calculated according to first assessed value and second assessed value it is described into
The movable success rate of row presell.
In one embodiment, the first power can be respectively set for first assessed value and second assessed value in advance
Weight coefficient and the second weight coefficient, wherein first assessed value corresponds to the first weight coefficient, second assessed value corresponding the
Two weight coefficients, the computing module 104 can be by subtracting first assessed value and the product of first weight coefficient
Second assessed value and the product of second weight coefficient is gone to obtain the pending movable success rate of presell of trade company A
(calculation formula may is that success rate=first the second weight coefficient of assessed value * the first weight coefficient the-the second assessed value *).
The output module 105 is used for when the pending movable success rate of presell is greater than the preset threshold, defeated
First prompt information out, when the pending movable success rate of presell is not more than the preset threshold, the second prompt of output
Information.
In one embodiment, when the pending movable success rate of presell is greater than the preset threshold, show this
The probalility of success of trade company's A presell is greater than failure probability, and trade company's business risk is lower, and the presell that user can participate in trade company A is living
It is dynamic, when the pending movable success rate of presell is not more than the preset threshold, show the probalility of success of trade company A presell
Less than failure probability, trade company's business risk is larger, it is not recommended that user participates in the presell activity of trade company A.The preset threshold can
To be set and be adjusted according to actual use demand, for example, the preset threshold is set as 0.5, if trade company A is calculated
The success rate of presell is greater than 0.5, then it represents that the probalility of success of the trade company A presell is greater than failure probability, output the first prompt letter
It ceases to user, first prompt information may is that presell risk is lower, and the presell activity for participating in trade company A may be selected；If meter
The success rate for obtaining trade company A presell is calculated no more than 0.5, then it represents that the probalility of success of the trade company A presell is less than failure probability,
The second prompt information is exported to user, second prompt information may is that presell risk is higher, it is not recommended that participate in trade company A
Presell activity.
Fig. 4 is the schematic diagram of computer installation preferred embodiment of the present invention.
The computer installation 1 includes memory 20, processor 30 and is stored in the memory 20 and can be in institute
State the computer program 40 run on processor 30, such as presell risk assessment procedures.The processor 30 executes the calculating
The step in above-mentioned presell methods of risk assessment embodiment, such as step S11~S14 shown in FIG. 1, figure are realized when machine program 40
Step S11~S17 shown in 2.Alternatively, the processor 30 realizes that above-mentioned presell risk is commented when executing the computer program 40
Estimate the function of each module in system embodiment, such as the module 101~105 in Fig. 3.
Illustratively, the computer program 40 can be divided into one or more module/units, it is one or
Multiple module/units are stored in the memory 20, and are executed by the processor 30, to complete the present invention.Described one
A or multiple module/units can be the series of computation machine program instruction section that can complete specific function, and described instruction section is used
In implementation procedure of the description computer program 40 in the computer installation 1.For example, the computer program 40 can be with
Be divided into Fig. 3 first obtain module 101, model training module 102, second obtain module 103, computing module 104 and
Output module 105.Each module concrete function is referring to embodiment two.
The computer installation 1 can be the calculating such as desktop PC, notebook, palm PC and cloud server and set
It is standby.It will be understood by those skilled in the art that the schematic diagram is only the example of computer installation 1, do not constitute to computer
The restriction of device 1 may include perhaps combining certain components or different components, example than illustrating more or fewer components
Such as described computer installation 1 can also include input-output equipment, network access equipment, bus.
Alleged processor 30 can be central processing unit (Central Processing Unit, CPU), can also be
Other general processors, digital signal processor (Digital Signal Processor, DSP), specific integrated circuit
(Application Specific Integrated Circuit, ASIC), ready-made programmable gate array (Field-
Programmable Gate Array, FPGA) either other programmable logic device, discrete gate or transistor logic,
Discrete hardware components etc..General processor can be microprocessor or the processor 30 is also possible to any conventional processing
Device etc., the processor 30 are the control centres of the computer installation 1, utilize various interfaces and the entire computer of connection
The various pieces of device 1.
The memory 20 can be used for storing the computer program 40 and/or module/unit, and the processor 30 passes through
Operation executes the computer program and/or module/unit being stored in the memory 20, and calls and be stored in memory
Data in 20 realize the various functions of the computer installation 1.The memory 20 can mainly include storing program area and deposit
Store up data field, wherein storing program area can application program needed for storage program area, at least one function (for example sound is broadcast
Playing function, image player function etc.) etc.；Storage data area, which can be stored, uses created data (ratio according to computer installation 1
Such as audio data, phone directory) etc..In addition, memory 20 may include high-speed random access memory, it can also include non-easy
The property lost memory, such as hard disk, memory, plug-in type hard disk, intelligent memory card (Smart Media Card, SMC), secure digital
(Secure Digital, SD) card, flash card (Flash Card), at least one disk memory, flush memory device or other
Volatile solid-state part.
If the integrated module/unit of the computer installation 1 is realized in the form of SFU software functional unit and as independence
Product when selling or using, can store in a computer readable storage medium.Based on this understanding, of the invention
It realizes all or part of the process in above-described embodiment method, can also instruct relevant hardware come complete by computer program
At the computer program can be stored in a computer readable storage medium, and the computer program is held by processor
When row, it can be achieved that the step of above-mentioned each embodiment of the method.Wherein, the computer program includes computer program code, institute
Stating computer program code can be source code form, object identification code form, executable file or certain intermediate forms etc..It is described
Computer-readable medium may include: any entity or device, recording medium, U that can carry the computer program code
Disk, mobile hard disk, magnetic disk, CD, computer storage, read-only memory (ROM, Read-Only Memory), arbitrary access
Memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium etc..It needs
It is bright, the content that the computer-readable medium includes can according in jurisdiction make laws and patent practice requirement into
Row increase and decrease appropriate, such as do not include electric load according to legislation and patent practice, computer-readable medium in certain jurisdictions
Wave signal and telecommunication signal.
In several embodiments provided by the present invention, it should be understood that disclosed computer installation and method, it can be with
It realizes by another way.For example, computer installation embodiment described above is only schematical, for example, described
The division of unit, only a kind of logical function partition, there may be another division manner in actual implementation.
It, can also be in addition, each functional unit in each embodiment of the present invention can integrate in same treatment unit
It is that each unit physically exists alone, can also be integrated in same unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of hardware adds software function module.
It is obvious to a person skilled in the art that invention is not limited to the details of the above exemplary embodiments, Er Qie
In the case where without departing substantially from spirit or essential attributes of the invention, the present invention can be realized in other specific forms.Therefore, no matter
From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and the scope of the present invention is by appended power
Benefit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent elements of the claims
Variation is included in the present invention.Any reference signs in the claims should not be construed as limiting the involved claims.This
Outside, it is clear that one word of " comprising " does not exclude other units or steps, and odd number is not excluded for plural number.It is stated in computer installation claim
Multiple units or computer installation can also be implemented through software or hardware by the same unit or computer installation.The
One, the second equal words are used to indicate names, and are not indicated any particular order.
Finally it should be noted that the above examples are only used to illustrate the technical scheme of the present invention and are not limiting, although reference
Preferred embodiment describes the invention in detail, those skilled in the art should understand that, it can be to of the invention
Technical solution is modified or equivalent replacement, without departing from the spirit and scope of the technical solution of the present invention.
Claims (10)
1. a kind of presell methods of risk assessment, which is characterized in that the described method includes:
History feature data are obtained, wherein the history feature data are that trade company has initiated presell behavior and had presell result
Data, the presell result include presell success and presell failure；
History feature data corresponding to different presell results are separately input into neural network model to be trained, with respectively
It obtains corresponding to successful first risk evaluation model of presell and the second risk evaluation model corresponding to presell failure；
A pending movable characteristic of presell is obtained, and the characteristic is input to first risk evaluation model
The first assessed value is obtained, and the characteristic is input to second risk evaluation model and obtains the second assessed value；And
The movable success rate of pending presell is calculated according to first assessed value and second assessed value.
2. presell methods of risk assessment as described in claim 1, which is characterized in that the history feature data include multiple quotient
Family dimensional information, multiple trade company's dimensional informations include but is not limited to: Merchants register fund information, shareholder number information, stock
The preferential dynamics letter of urban information, shops's location information, the affiliated trade information of trade company, commodity where eastern credit record information, shops
Breath, grace period information, merchandise cost information, opening temporal information, trade company's profit or loss information, customer quantity information.
3. presell methods of risk assessment as claimed in claim 2, which is characterized in that the step of the acquisition history feature data
Later further include:
Judge whether there is one or more trade company's dimensional information missing；And
When setting one or more trade company's dimensional information of missing to there are one or more trade company's dimensional information missing
Default characteristic.
4. presell methods of risk assessment as claimed in claim 1 or 2, which is characterized in that described by different presell result institutes
Corresponding history feature data are separately input into neural network model and are trained, and correspond to presell successful the to respectively obtain
One risk evaluation model and corresponding to presell failure the second risk evaluation model the step of include:
Presell result is input to first nerves network model for the successful history feature data of presell to be trained, is obtained described
First risk evaluation model；
The history feature data that presell result is presell failure are input to nervus opticus network model to be trained, are obtained described
Second risk evaluation model；
Wherein, the first nerves network model and the nervus opticus network model are BP neural network model, the BP
Neural network model includes input layer, hidden layer and output layer.
5. presell methods of risk assessment as claimed in claim 4, which is characterized in that the input layer includes n node, described
Hidden layer includes m node, the BP neural network model are as follows:
Wherein, y is the output valve of the output layer, when the characteristic is input to first risk evaluation model,
The output valve y of first risk evaluation model is first assessed value, when the characteristic is input to described second
When risk evaluation model, the output valve y of second risk evaluation model is second assessed value, tiFor the hidden layer with
Connection weight between the output layer,(i=1,2,3...m；It j=1,2,3...n is) described hidden
Hide the input of layer, the output of the as described input layer, WijFor the connection weight between the input layer and the hidden layer；f
(Si) be the BP neural network model in activation primitive,
6. presell methods of risk assessment as described in claim 1, which is characterized in that described according to first assessed value and institute
Stating the step of pending presell movable success rate is calculated in the second assessed value includes:
Obtain corresponding first weight coefficient of first assessed value and corresponding second weight coefficient of second assessed value；And
The product of first assessed value and first weight coefficient is subtracted into second assessed value and second weight
The product of coefficient obtains the movable success rate of pending presell.
7. presell methods of risk assessment as claimed in claim 6, which is characterized in that described according to first assessed value and institute
After stating the step of pending presell movable success rate is calculated in the second assessed value further include:
Judge whether the movable success rate of pending presell is greater than preset threshold；
When the pending movable success rate of presell is greater than the preset threshold, the first prompt information is exported；And
When the pending movable success rate of presell is not more than the preset threshold, the second prompt information is exported.
8. a kind of presell risk evaluating system, which is characterized in that the system comprises:
First obtains module, is used for history feature data, wherein the history feature data are that trade company has initiated presell behavior
And the data with presell result, the presell result include presell success and presell failure；
Model training module, for history feature data corresponding to different presell results to be separately input into neural network mould
Type is trained, and corresponds to successful first risk evaluation model of presell and corresponding to the second wind of presell failure to respectively obtain
Dangerous assessment models；
Second obtains module, for obtaining a pending movable characteristic of presell, and the characteristic is input to institute
It states the first risk evaluation model and obtains the first assessed value, and the characteristic is input to second risk evaluation model and is obtained
To the second assessed value；And
Computing module, for the pending presell activity to be calculated according to first assessed value and second assessed value
Success rate.
9. a kind of computer installation, the computer installation includes processor and memory, is stored on the memory several
Computer program, which is characterized in that such as right is realized when the processor is for executing the computer program stored in memory
It is required that described in any one of 1-7 the step of presell methods of risk assessment.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of presell methods of risk assessment as described in any one of claim 1-7 is realized when being executed by processor.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811289813.1A CN109598410A (en) | 2018-10-31 | 2018-10-31 | Presell methods of risk assessment, system, computer installation and readable storage medium storing program for executing |
| PCT/CN2019/077514 WO2020087828A1 (en) | 2018-10-31 | 2019-03-08 | Pre-sales risk assessment method and system, computer apparatus and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811289813.1A CN109598410A (en) | 2018-10-31 | 2018-10-31 | Presell methods of risk assessment, system, computer installation and readable storage medium storing program for executing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109598410A true CN109598410A (en) | 2019-04-09 |
Family
ID=65957263
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811289813.1A Pending CN109598410A (en) | 2018-10-31 | 2018-10-31 | Presell methods of risk assessment, system, computer installation and readable storage medium storing program for executing |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN109598410A (en) |
| WO (1) | WO2020087828A1 (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110599241A (en) * | 2019-08-22 | 2019-12-20 | 深圳市元征科技股份有限公司 | Marketing scheme recommendation method and device |
| CN111125071A (en) * | 2019-11-25 | 2020-05-08 | 珠海格力电器股份有限公司 | Method and device for evaluating real-time data reliability and storage medium |
| CN113763019A (en) * | 2021-01-28 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | User information management method and device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103366190A (en) * | 2013-07-26 | 2013-10-23 | 中国科学院自动化研究所 | Method for identifying traffic sign |
| US20140258096A1 (en) * | 2008-03-28 | 2014-09-11 | American Express Travel Related Services Company, Inc. | Consumer behaviors at lender level |
| CN106960266A (en) * | 2016-01-11 | 2017-07-18 | 阿里巴巴集团控股有限公司 | The operating mode pattern method and device of a kind of presell article |
| CN107025597A (en) * | 2016-07-20 | 2017-08-08 | 平安科技（深圳）有限公司 | Risk Forecast Method and device based on big data |
| CN107679557A (en) * | 2017-09-19 | 2018-02-09 | 平安科技（深圳）有限公司 | Driving model training method, driver's recognition methods, device, equipment and medium |
-
2018
- 2018-10-31 CN CN201811289813.1A patent/CN109598410A/en active Pending
-
2019
- 2019-03-08 WO PCT/CN2019/077514 patent/WO2020087828A1/en active Application Filing
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140258096A1 (en) * | 2008-03-28 | 2014-09-11 | American Express Travel Related Services Company, Inc. | Consumer behaviors at lender level |
| CN103366190A (en) * | 2013-07-26 | 2013-10-23 | 中国科学院自动化研究所 | Method for identifying traffic sign |
| CN106960266A (en) * | 2016-01-11 | 2017-07-18 | 阿里巴巴集团控股有限公司 | The operating mode pattern method and device of a kind of presell article |
| CN107025597A (en) * | 2016-07-20 | 2017-08-08 | 平安科技（深圳）有限公司 | Risk Forecast Method and device based on big data |
| CN107679557A (en) * | 2017-09-19 | 2018-02-09 | 平安科技（深圳）有限公司 | Driving model training method, driver's recognition methods, device, equipment and medium |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110599241A (en) * | 2019-08-22 | 2019-12-20 | 深圳市元征科技股份有限公司 | Marketing scheme recommendation method and device |
| CN111125071A (en) * | 2019-11-25 | 2020-05-08 | 珠海格力电器股份有限公司 | Method and device for evaluating real-time data reliability and storage medium |
| CN113763019A (en) * | 2021-01-28 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | User information management method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2020087828A1 (en) | 2020-05-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Honegger | Shedding light on black box machine learning algorithms: Development of an axiomatic framework to assess the quality of methods that explain individual predictions | |
| Samoilenko et al. | Determining sources of relative inefficiency in heterogeneous samples: Methodology using Cluster Analysis, DEA and Neural Networks | |
| US20200250511A1 (en) | Artist comprehensive ability evaluation and cultivation assistant system based on artificial intelligence | |
| CN108140075A (en) | User behavior is classified as exception | |
| CN114048331A (en) | Knowledge graph recommendation method and system based on improved KGAT model | |
| US11854088B1 (en) | Methods and systems for improving the underwriting process | |
| CN109598410A (en) | Presell methods of risk assessment, system, computer installation and readable storage medium storing program for executing | |
| JP2017199355A (en) | Recommendation generation | |
| Wu et al. | Wavelet fuzzy cognitive maps | |
| CN109615504A (en) | Products Show method, apparatus, electronic equipment and computer readable storage medium | |
| Cui et al. | Targeting high value customers while under resource constraint: partial order constrained optimization with genetic algorithm | |
| CN109426983A (en) | Dodge purchase activity automatic generation method and device, storage medium, electronic equipment | |
| CN113408582B (en) | Training method and device for feature evaluation model | |
| Chishti et al. | Deep neural network a step by step approach to classify credit card default customer | |
| CN109767333A (en) | Select based method, device, electronic equipment and computer readable storage medium | |
| CN112241920A (en) | Investment and financing organization evaluation method, system and equipment based on graph neural network | |
| Kalaycı et al. | Optimal model description of finance and human factor indices | |
| CN111222993A (en) | Fund recommendation method and device | |
| CN113822390B (en) | User portrait construction method and device, electronic equipment and storage medium | |
| Kang et al. | RETRACTED: A hybrid model to develop aesthetic product design of customer satisfaction | |
| CN115439180A (en) | Target object determination method and device, electronic equipment and storage medium | |
| Zhuang et al. | Application of data mining in term deposit marketing | |
| CN111667307A (en) | Method and device for predicting financial product sales volume | |
| Tarutin et al. | Forecasting trends with the presence of neural networks | |
| Despotovic et al. | Leveraging supplementary modalities in automated real estate valuation using comparative judgments and deep learning |
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 |