CN109816221A - Decision of Project Risk method, apparatus, computer equipment and storage medium - Google Patents
Decision of Project Risk method, apparatus, computer equipment and storage medium Download PDFInfo
- Publication number
- CN109816221A CN109816221A CN201910012593.6A CN201910012593A CN109816221A CN 109816221 A CN109816221 A CN 109816221A CN 201910012593 A CN201910012593 A CN 201910012593A CN 109816221 A CN109816221 A CN 109816221A
- Authority
- CN
- China
- Prior art keywords
- risk
- project
- decision
- factor
- data
- 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
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000004590 computer program Methods 0.000 claims description 29
- 230000006870 function Effects 0.000 claims description 21
- 230000015654 memory Effects 0.000 claims description 21
- 230000000694 effects Effects 0.000 claims description 16
- 238000005457 optimization Methods 0.000 claims description 14
- 238000012549 training Methods 0.000 claims description 14
- 238000000513 principal component analysis Methods 0.000 claims description 11
- 238000012502 risk assessment Methods 0.000 abstract description 7
- 238000012545 processing Methods 0.000 abstract description 2
- 238000013139 quantization Methods 0.000 description 12
- 230000008859 change Effects 0.000 description 11
- 239000000284 extract Substances 0.000 description 7
- 238000004422 calculation algorithm Methods 0.000 description 6
- 238000012216 screening Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 238000010276 construction Methods 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 239000000203 mixture Substances 0.000 description 3
- 230000000306 recurrent effect Effects 0.000 description 3
- 238000003066 decision tree Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000013058 risk prediction model Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013075 data extraction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
This application involves big data processing field, in particular to a kind of Decision of Project Risk method, apparatus, computer equipment and storage medium.The method includes receiving Decision of Project Risk instruction；Project category and project information data are obtained from Decision of Project Risk instruction；Default risk forecast model corresponding with project category is searched, and obtains the decision parameters in default risk forecast model；Acquire Real-time Decision factor data corresponding with decision parameters；Threshold value is divided according to risk of the Real-time Decision factor data to default risk forecast model to be adjusted to obtain real-time risk forecast model；Project information data are inputted into real-time risk forecast model and obtain assessment risk, Decision of Project Risk suggestion is generated according to assessment risk.Project Risk Assessment accuracy can be improved using this method.
Description
Technical field
This application involves field of computer technology, set more particularly to a kind of Decision of Project Risk method, apparatus, computer
Standby and storage medium.
Background technique
Project investment is the important hand realized the main path of social capital accumulation function, and expand social reproduction
Section.But project investment has the characteristics that investment amount is more, influence time is long, occurrence frequency is low, cashability is poor, and therefore, item
It is very big that mesh invests investment risk.
Therefore, before being invested, carrying out assessment to project risk is particularly important.Currently, in the market to project
Risk when being assessed, often assessed, but be not able to satisfy the market needs of dynamic change, commented according to expertise
It is unsatisfactory to estimate result.
Summary of the invention
Based on this, it is necessary in view of the above technical problems, provide a kind of project wind for improving Project Risk Assessment accuracy
Dangerous decision-making technique, device, computer equipment and storage medium.
A kind of Decision of Project Risk method, which comprises
Receive Decision of Project Risk instruction；
Project category and project information data are obtained from Decision of Project Risk instruction；
Default risk forecast model corresponding with the project category is searched, and is obtained in the default risk forecast model
Decision parameters；
Acquire Real-time Decision factor data corresponding with the decision parameters；
Threshold value is divided according to risk of the Real-time Decision factor data to the default risk forecast model to carry out
Adjustment obtains real-time risk forecast model；
The project information data are inputted into the real-time risk forecast model and obtain assessment risk, according to institute's commentary
Estimate risk and generates Decision of Project Risk suggestion.
The generating mode of risk forecast model is preset in one of the embodiments, comprising:
Obtain quantizing factor data corresponding with the project category and initial risks prediction model；
Decision factor is filtered out from the quantizing factor data；
Project sample data is acquired according to the decision factor；
Optimization is adjusted to the initial risks prediction model according to the project sample data and constructs default risk
Prediction model.
The initial risks prediction model is adjusted according to the project sample data in one of the embodiments,
Optimize and construct default risk forecast model, comprising:
Decision factor sequence and Sample Risk score are extracted from the project sample data；
It obtains initial category and divides threshold value, threshold value is divided according to the initial category and the Sample Risk score sets item
Mesh risk label；
Project sample set is generated according to the decision factor sequence and the project risk class label；
Obtain preset model loss function；
The project sample set is inputted into the initial risks prediction model, is carried out according to the preset model loss function
Model training optimizes and constructs to obtain default risk forecast model.
Real-time Decision factor data corresponding with the decision parameters is acquired in one of the embodiments, comprising:
Project risk coefficient of variation is calculated according to the project sample data；
Search data sample period corresponding with the project risk coefficient of variation；
Obtain the newest decision factor data of real-time update；
Real-time Decision factor data is extracted from the newest decision factor data according to the data sample period.
Decision factor is filtered out from the quantizing factor data in one of the embodiments, comprising:
Calculate the factor contribution rate of each quantizing factor in the quantizing factor data；
Obtain default contribution rate threshold value；
The quantizing factor that the factor contribution rate is greater than the default contribution rate threshold value is screened as decision factor.
Decision factor is filtered out from the quantizing factor data in one of the embodiments, comprising:
Calculate the factor contribution rate of each quantizing factor in the quantizing factor data；
Prescreening is carried out to the quantizing factor according to the factor contribution rate；
Principal component analysis is carried out to the quantizing factor after prescreening and obtains factor principal component expression formula；
Decision factor is constructed according to the factor principal component expression formula.
In one of the embodiments, according to the Real-time Decision factor data to the wind of the default risk forecast model
Dangerous category division threshold value is adjusted to obtain real-time risk forecast model, comprising:
The Real-time Decision factor data is inputted into the default risk forecast model and obtains risk probability；
The risk probability is converted into risk level score；
Dynamic adjusts the risk and divides threshold value, each dynamic adjustment is calculated according to the risk level score described in
Risk divides the corresponding dynamic class label of threshold value；
Obtain the corresponding true class label of the Real-time Decision factor data；
Threshold classification effect curve is calculated according to the dynamic class label and the true class label；
Best category division threshold value is determined according to the threshold classification effect curve, according to the default risk forecast model
Real-time risk forecast model is generated with the best category division threshold value.
A kind of Decision of Project Risk device, described device include:
Command reception module, for receiving Decision of Project Risk instruction；
Project data obtains module, for obtaining project category and project information number from Decision of Project Risk instruction
According to；
Model searching module, for searching default risk forecast model corresponding with the project category, and described in acquisition
Decision parameters in default risk forecast model；
Real-time data acquisition module, for acquiring Real-time Decision factor data corresponding with the decision parameters；
Real-time model generation module, for according to the Real-time Decision factor data to the default risk forecast model
Risk divides threshold value and is adjusted to obtain real-time risk forecast model；
Decision recommendation module obtains assessment wind for the project information data to be inputted the real-time risk forecast model
Dangerous classification generates Decision of Project Risk suggestion according to the assessment risk.
A kind of computer equipment, including memory and processor, the memory are stored with computer program, the processing
The step of device realizes the above method when executing the computer program.
A kind of computer readable storage medium, is stored thereon with computer program, and the computer program is held by processor
The step of above method is realized when row.
Above-mentioned Decision of Project Risk method, apparatus, computer equipment and storage medium, refer to when receiving Decision of Project Risk
After order, risk forecast model corresponding with the project category to decision project is searched, and is acquired and determining in risk forecast model
The corresponding Real-time Decision data of plan parameter divide threshold value to the risk in model according to Real-time Decision data and adjust
It is whole, decision project is treated according to the risk forecast model after real-time adjustment dynamic threshold and carries out risk assessment, so as to basis
The decision data of dynamic change is constantly adjusted risk forecast model, the effect of model risk classification is improved, to project
Risk assessment is more rationally effective.
Detailed description of the invention
Fig. 1 is the application scenario diagram of project Application of risk decision method in one embodiment；
Fig. 2 is the flow diagram of project Application of risk decision method in one embodiment；
Fig. 3 is the flow diagram that the generation method of risk forecast model is preset in one embodiment；
Fig. 4 is the structural block diagram of Decision of Project Risk device in one embodiment；
Fig. 5 is the internal structure chart of computer equipment in one embodiment.
Specific embodiment
It is with reference to the accompanying drawings and embodiments, right in order to which the objects, technical solutions and advantages of the application are more clearly understood
The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, not
For limiting the application.
Decision of Project Risk method provided by the present application, can be applied in application environment as shown in Figure 1.Wherein, eventually
End 102 is communicated with server 104 by network by network.User generates project risk by operation on the terminal 102
Decision of Project Risk instruction is sent to server 104 by decision instruction, terminal 102, and server 104 receives Decision of Project Risk and refers to
It enables；Project category and project information data are obtained from Decision of Project Risk instruction；Search default wind corresponding with project category
Dangerous prediction model, and obtain the decision parameters in default risk forecast model；Acquire corresponding with decision parameters Real-time Decision because
Subdata；Threshold value is divided according to risk of the Real-time Decision factor data to default risk forecast model to be adjusted to obtain reality
When risk forecast model；Project information data are inputted into real-time risk forecast model and obtain assessment risk, according to assessment wind
Dangerous classification generates Decision of Project Risk suggestion.The Decision of Project Risk suggestion of generation is returned to terminal 102 by server 104, with
Decision of Project Risk is carried out for user.Wherein, terminal 102 can be, but not limited to be various personal computers, laptop, intelligence
Energy mobile phone, tablet computer and portable wearable device, server 104 can use independent server either multiple servers
The server cluster of composition is realized.
In one embodiment, as shown in Fig. 2, providing a kind of Decision of Project Risk method, it is applied to Fig. 1 in this way
In server 104 for be illustrated, in other embodiments, this method also can be applied to server, method include with
Lower step:
Step 210, Decision of Project Risk instruction is received.
The investors such as government, enterprise can input investment when carrying out risk assessment decision to investment project at the terminal
The project data of project.Decision of Project Risk miscellaneous function can be provided in terminal, investor user can pass through selection operation
Decision of Project Risk miscellaneous function is provided and generates Decision of Project Risk instruction, the Decision of Project Risk of generation is instructed and sent by terminal
To server, the Decision of Project Risk instruction that server receiving terminal is sent, Decision of Project Risk instruction carries project to be assessed
Project category and project information data.
Step 220, project category and project information data are obtained from Decision of Project Risk instruction.
Server obtains project category and project information data from Decision of Project Risk instruction, and project category is to various
Investment project is classified, and project category may include financial classification, building classification, medical categories and entertainment classification etc..
Essential information of the project information data comprising investment project, such as investment, investment cycle, undertaking enterprise, Target Enterprise information,
It also include Vehicles Collected from Market information, as (consumer price index, resident disappear by GDP (GDP) growth rate, CPI
Take price index) information such as the rate of change.
Step 230, default risk forecast model corresponding with project category is searched, and is obtained in default risk forecast model
Decision parameters.
The default risk forecast model for the setting of various project categories is stored in server, for investment project
Risk is assessed.The input decision parameters and operation method of the default risk forecast model of disparity items classification may be different,
The output of model is assessment risk.Assessment risk can be set according to risk class, as high risk classification, in
Risk and low-risk classification etc. can also be set according to other category division standards.
Server searches corresponding default risk forecast model according to project category, obtains determining for default risk forecast model
Plan parameter, decision parameters are on the influential decision factor of project risk tool.Decision factor includes reaction project essential information
The factors such as the project factor such as project investment volume, the project cycle, decision factor further include the market fluctuation for reacting fluctuation of price situation
The factor, as GDP (GDP) growth rate, CPI (consumer price index, Consumer Prices index) change
Rate etc..
Step 240, Real-time Decision factor data corresponding with decision parameters is acquired.
Collection of server Real-time Decision factor data corresponding with decision parameters, Real-time Decision factor data are by continuous
The newest decision factor data that dynamic updates.Specifically, server can acquire each decision factor according to preset period of time
Latest data, and existing decision factor data are updated according to latest data, preset period of time can be according to need
It is configured, such as can be set to 1 day, 1 week, 1 month.Newest decision factor data include in preset period of time more
The project decision factor of new generic built project and market trend change decision factor.
Step 250, threshold value is divided according to risk of the Real-time Decision factor data to default risk forecast model to carry out
Adjustment obtains real-time risk forecast model.
Risk divides the boundary threshold of the horizontal score of project risk corresponding to the various risk that threshold value is setting
Value.If the corresponding project risk level of high risk classification is scored at 80 points or more, the corresponding project risk water of medium risk classification
Flat to be scored at 60 points to 80/, the corresponding project risk level of low risk is scored at 20 points to 60/, then
The risk of 20 points, 60 points and 80 points respectively basic, normal, high three risk divides threshold value.
Server extracts the risk really delimited from the project data of updated project sample, will be real
When decision factor data input in default risk forecast model and obtain the assessment risk of projects sample, by constantly adjusting
The numerical value of model risk category division threshold value obtains different classification of risks as a result, by according to the risk of each dynamic change
It divides the classification of risks result that threshold value obtains to be compared with real risk classification, it is the smallest optimal to select classification results difference
Risk divides threshold value, is adjusted according to ultimate risk category division threshold value to the correspondence parameter in default risk forecast model
It is whole, obtain the real-time risk forecast model at current time.
Step 260, project information data are inputted into real-time risk forecast model and obtains assessment risk, according to assessment wind
Dangerous classification generates Decision of Project Risk suggestion.
Project information data to be assessed are inputted real-time risk forecast model and obtain the wind of various risks classification by server
The risk of risk maximum probability is set as assessing risk by dangerous class probability.It is stored in server promising each
The Decision of Project Risk suggestion of the various assessment risk setting of kind project category, server are searched and project category and project
Decision in the face of risk suggests corresponding Decision of Project Risk suggestion, and Decision of Project Risk suggestion is returned to terminal, for user into
Row Decision of Project Risk.
Different project categories may correspond to different Decision of Project Risk suggestions, the project of such as financial classification, if assessment
Risk is medium risk, and corresponding Decision of Project Risk suggestion may be to reduce scale of investment, and can set reduction and throw
The percentage of money scale；And the project of medical categories, if assessment risk is medium risk, corresponding Decision of Project Risk is built
View may be it is not recommended that investing, or optimize to project.
Above-mentioned Decision of Project Risk method is searched and the item to decision project after receiving Decision of Project Risk instruction
The corresponding risk forecast model of mesh classification, and acquire Real-time Decision number corresponding with the decision parameters in risk forecast model
According to dividing threshold value to the risk in model according to Real-time Decision data and be adjusted, after real-time adjustment dynamic threshold
Risk forecast model treat decision project carry out risk assessment, so as to according to the decision data of dynamic change constantly to wind
Dangerous prediction model is adjusted, and improves the effect of model risk classification, more rationally effective to the risk assessment of project.
In one embodiment, as shown in figure 3, the generating mode of default risk forecast model may comprise steps of:
Step 310, quantizing factor data corresponding with project category and initial risks prediction model are obtained.
The project of disparity items classification is corresponding with different quantizing factors and initial risks prediction model, project category with measure
The mapping relations for changing the mapping relations of the factor, project category and initial risks prediction model are previously stored in server, are serviced
Device searches quantizing factor corresponding with project category and initial risks prediction model.
Quantizing factor can generally be divided into project for judging investment project necessity, the numeralization index of feasibility
Essential information, necessity, feasibility, the index of affiliated party's four dimensions can specifically include investment, investment cycle, average
The factors such as cost.Collection of server with project category project historical data, and from historical data extract quantization factor pair answer
Quantizing factor data and history item risk level score.
Initial risks prediction model can be for using the obtained disaggregated model of machine learning algorithm training that have supervision, can be with
Attempting Supervised machine learning algorithm includes the CNN in decision-tree model, Ensemble Learning Algorithms and deep learning algorithm
(Convolutional Neural Network, convolutional neural networks), RNN (Recognition with Recurrent Neural Network), LSTM (Long
Short-Term Memory, shot and long term memory network) scheduling algorithm.
Step 320, decision factor is filtered out from quantizing factor data.
Server can calculate each quantization according to the quantizing factor data of each quantizing factor and the risk level score of extraction
The degree of correlation of the factor and project risk screens the higher quantizing factor of the degree of correlation for decision factor.Server can also be to amount
Change factor data to be analyzed, therefrom construct decision factor, project risk prediction is carried out according to the decision factor constructed.
In one embodiment, it may include: to calculate quantizing factor number that decision factor is filtered out from quantizing factor data
The factor contribution rate of each quantizing factor in；Obtain default contribution rate threshold value；Factor contribution rate is greater than default contribution rate threshold value
Quantizing factor screening be decision factor.
Each quantization is calculated according to the quantizing factor data of each quantizing factor and the risk level score of extraction in server
The factor contribution rate of the factor.Specifically, server can generate each quantizing factor according to the quantizing factor data of projects sample
Quantizing factor sequence, according to the risk level score of each sample generate risk level sequence, by quantizing factor sequence and risk
Video sequence inputs random forest decision-tree model, and the factor contribution rate of each quantizing factor is calculated；Server can also be adopted
The factor contribution rate of each quantizing factor is calculated with Logic Regression Models, server carries out the quantizing factor sequence of each quantizing factor
Standardization, according to after standardization quantizing factor sequence and risk level sequence establish Logic Regression Models, extraction logic returns
The absolute value of the variable regression coefficient of each quantizing factor in model, the accounting for calculating each variable regression absolute coefficient is the factor
Contribution rate.Server can also calculate the factor contribution rate of each quantizing factor using other algorithms.
Server obtains default contribution rate threshold value, and default contribution rate threshold value is for dividing contribution degree.Server will
The factor contribution rate of each quantizing factor is compared with default contribution rate threshold value, and factor contribution rate is greater than default contribution rate threshold value
Quantizing factor screening be decision factor.
In one embodiment, filtered out from quantizing factor data decision factor may include: server to quantization because
Subdata carries out principal component analysis and obtains principal component expression formula, constructs decision factor according to principal component expression formula.
In the present embodiment, server using Principal Component Analysis to quantizing factor data and risk level score data into
Row principal component analysis, obtains multiple principal components, and server constructs principal component expression formula, the principal component of building according to multiple principal components
Expression formula can be first principal component, or the linear combination of top n principal component.Server will be each in principal component expression formula
The product of variable and variation coefficient is configured to decision factor.In the present embodiment, the decision factor of building is and original quantization
The different variable of the factor, but can be calculated by original quantizing factor.During the prediction of subsequent project risk, clothes
First decision factor data are calculated according to quantizing factor data according to the composition of decision factor in business device, then by calculated decision
Factor data inputs default risk forecast model and carries out risk profile.
In one embodiment, it may include: to calculate quantizing factor number that decision factor is filtered out from quantizing factor data
The factor contribution rate of each quantizing factor in；Prescreening is carried out to quantizing factor according to factor contribution rate；To the amount after prescreening
Change factor progress principal component analysis and obtains factor principal component expression formula；Decision factor is constructed according to factor principal component expression formula.
In the present embodiment, the method in above-mentioned two embodiment is combined by server, first calculate each quantization because
The factor contribution rate of son, the calculation method of factor contribution rate are referred to above-described embodiment, and details are not described herein.Obtain default tribute
Rate threshold value is offered, prescreening is carried out to quantizing factor according to default contribution rate threshold value, is screened out from it the higher amount of the risk degree of correlation
Change the factor.Principal component analysis is carried out to the quantizing factor after prescreening using Principal Component Analysis again and obtains the expression of factor principal component
The product of variable each in principal component expression formula and variation coefficient is configured to decision factor by formula.
Step 330, project sample data is acquired according to decision factor.
Collection of server with the built project sample of project category history of project data, from the project of projects sample
Decision factor data corresponding with decision factor are extracted or constructed in historical data, and are extracted from projects sample data
The horizontal score of project risk generates project sample data according to the horizontal score of the decision factor data and project risk that extract.
Step 340, optimization is adjusted to initial risks prediction model according to project sample data and constructs default risk
Prediction model.
Server extracted from project sample data projects sample decision factor data and project risk it is horizontal
Point, classified according to the target risk that the horizontal score of project risk sets the sample, one group of decision factor data and corresponding target
Classification of risks constitutes a sample data, and the sample data of multiple project samples forms project sample set.Server obtains basis
Project sample set is trained optimization to the parameter of initial risks prediction model, generates default wind according to the model parameter after optimization
Dangerous prediction model.Further, project sample set can be divided into training set and test set by server, using training set data
Initial risks prediction model is trained, and the initial risks prediction model after training is optimized using test set data
Adjustment.
In one embodiment, optimization is adjusted to initial risks prediction model according to project sample data and constructed pre-
If risk forecast model, comprising: extract decision factor sequence and Sample Risk score from project sample data；Obtain initial classes
Not Hua Fen threshold value, threshold value and Sample Risk score setting item risk label are divided according to initial category；According to decision because
Subsequence and project risk class label generate project sample set；Obtain preset model loss function；Project sample set is inputted
Initial risks prediction model carries out model training optimization and constructs to obtain default risk profile mould according to preset model loss function
Type.
The decision factor numerical value that server extracts projects sample from project sample data constitutes decision factor sequence,
And therefrom extract the Sample Risk score of projects sample.It is various risk initially set that initial category, which divides threshold value,
The demarcation threshold of the horizontal score of corresponding project risk, can according to expertise, the analysis result of history item data into
Row setting.Server divides threshold value according to initial category and converts corresponding project wind for the Sample Risk score of projects sample
Dangerous classification, and generate project risk label as the target of risk forecast model and export result.According to decision factor sequence
The project sample set comprising multiple sample datas is generated with corresponding project risk class label.
Server obtains preset model loss function, and preset model loss function can be using logarithm loss function or square
Error function etc..In the present embodiment, initial risks prediction model uses time recurrent neural networks model, can such as use
The times recurrent neural network such as RNN, LSTM, by information data, temporally axis is added in this class model, is added to memory and is given up
The valve of past information carries out decayingization place to each phase data information inputted on time dimension according to certain information rejection rate
Reason retains a certain proportion of information in the past and is used as model explanation variable, according to project demands progress model function and hyper parameter
Adjustment.
Preferably, initial risks prediction model uses LSTM model, and model core formula is as follows:
ft=σ (Wf*[Ct-1,ht-1,xt]+bf)
it=σ (Wi*[Ct-1,ht-1,xt]+bi)
ot=σ (Wo*[Ct-1,ht-1,xt]+bo)
Wherein ft, itRespectively forget door and input gate, CtFor the merging of two doors, model above formula is each
Input increases a status signal.
Server by project sample set input initial risks prediction model, and according to preset model loss function to model into
Row training, model training process is the process optimized and revised to each parameter in model, constructs model parameter after training
Optimal default risk forecast model.
In one embodiment, acquiring Real-time Decision factor data corresponding with the decision parameters may include: basis
The project sample data calculates project risk coefficient of variation；Search data sample corresponding with the project risk coefficient of variation
Period；Obtain the newest decision factor data of real-time update；According to the data sample period from the newest decision factor number
Real-time Decision factor data is extracted in.
Server extracts the risk level score of all items sample from project sample data, according to the wind extracted
Dangerous book review score calculates the project risk coefficient of variation of the affiliated project category of investment project.Project risk coefficient of variation can be adopted
The index of risk data fluctuation can be reacted with the variance of risk level score, standard deviation etc. to indicate.
Server divides numberical range, the corresponding data sample of each numberical range to project risk coefficient of variation in advance
This period, server determine numberical range belonging to calculated project risk coefficient of variation, and it is corresponding to search the numberical range
Data sample period.Data sample period is the data sampling period for reacting the decision factor of fluctuation of price situation, generally,
The biggish corresponding data sample period of project risk coefficient of variation is smaller, conversely, corresponding data sample period is larger.Service
Device obtains the newest decision factor data of real-time update, and the more appearing method of newest decision factor data is referred to above-mentioned implementation
Example, repeats no more again.Server carries out data according to the data sample period found from newest decision factor data and adopts
Sample, sampling obtain the time data sequence of each decision factor, form Real-time Decision factor data.
In one embodiment, threshold is divided according to risk of the Real-time Decision factor data to default risk forecast model
Value is adjusted to obtain real-time risk forecast model to may include: that Real-time Decision factor data is inputted default risk forecast model
Obtain risk probability；Risk probability is converted into risk level score；Dynamic adjustment risk divides threshold value, root
The corresponding dynamic class label of threshold value is divided according to the risk that risk level score calculates each dynamic adjustment；Obtain Real-time Decision
The corresponding true class label of factor data；It is bent that threshold classification effect is calculated according to dynamic class label and true class label
Line；Best category division threshold value is determined according to threshold classification effect curve, is drawn according to default risk forecast model and best classification
Threshold value is divided to generate real-time risk forecast model.
Collected Real-time Decision factor data is input in default risk forecast model by server, is calculated various
The risk probability of risk is previously set the various risks class probability risk level final with project in server and obtains
The conversion formula divided, is converted to risk level score for risk probability according to conversion formula.For example, can be to different classes of
Risk probability set Risk rated ratio, Risk rated ratio is weighted read group total with corresponding risk probability, is obtained
To risk level score.
Server obtains the initial risks category division threshold value of default risk forecast model, and constantly dynamic linear adjusts threshold
It is worth value, obtains multiple groups risk and divide threshold value, the linear numerical interval between every group of threshold value, which can according to need, to be set
It sets, the precision of calculated result can also correspondingly increase when being spaced smaller.Server divides threshold value according to every group of risk respectively and sets
The corresponding dynamic class label of risk level score of fixed each project sample, obtains the real time class mark of each project sample
Label calculate the class that each group risk divides threshold value according to the comparison result of the dynamic class label of multiple projects and real time class
Not Hua Fen accuracy rate, such as the accuracy rate that label in project sample is set can be set as category division accuracy rate, can also be with
Category division accuracy rate is calculated using other modes.
It is independent variable that server each group risk, which divides threshold value, is counted using calculated category division accuracy rate as dependent variable
Threshold classification effect curve is calculated and drawn, and best category division threshold value is selected according to threshold classification effect curve, for example, clothes
The smallest extreme point of expression numerical value of curve can be chosen for best category division threshold value etc. by business device.Server is according to optimum kind
Not Hua Fen threshold value default risk forecast model is adjusted to obtain real-time risk forecast model, and according to real-time risk profile mould
Type carries out Risk of Investment Projection prediction, so as to adjust category division threshold value according to real-time market fluctuation situation, so that mould
Type is capable of the influence of continuous adaptation to market variations, improves the accuracy of model prediction.
It should be understood that although each step in the flow chart of Fig. 2-3 is successively shown according to the instruction of arrow,
These steps are not that the inevitable sequence according to arrow instruction successively executes.Unless expressly stating otherwise herein, these steps
Execution there is no stringent sequences to limit, these steps can execute in other order.Moreover, at least one in Fig. 2-3
Part steps may include that perhaps these sub-steps of multiple stages or stage are not necessarily in synchronization to multiple sub-steps
Completion is executed, but can be executed at different times, the execution sequence in these sub-steps or stage is also not necessarily successively
It carries out, but can be at least part of the sub-step or stage of other steps or other steps in turn or alternately
It executes.
In one embodiment, as shown in figure 4, providing a kind of Decision of Project Risk device, comprising: command reception module
410, project data obtains module 420, model searching module 430, real-time data acquisition module 440, real-time model generation module
450 and decision recommendation module 460, in which:
Command reception module 410, for receiving Decision of Project Risk instruction.
Project data obtains module 420, for obtaining project category and project information number from Decision of Project Risk instruction
According to.
Model searching module 430 for searching default risk forecast model corresponding with project category, and obtains default wind
Decision parameters in dangerous prediction model.
Real-time data acquisition module 440, for acquiring Real-time Decision factor data corresponding with decision parameters.
Real-time model generation module 450, for the risk according to Real-time Decision factor data to default risk forecast model
Category division threshold value is adjusted to obtain real-time risk forecast model.
Decision recommendation module 460 obtains assessment risk class for project information data to be inputted real-time risk forecast model
Not, Decision of Project Risk suggestion is generated according to assessment risk.
In one embodiment, Decision of Project Risk device can also include:
Initial model obtains module, predicts mould for obtaining quantizing factor data corresponding with project category and initial risks
Type.
Factor screening module, for filtering out decision factor from quantizing factor data.
Data acquisition module, for acquiring project sample data according to decision factor.
Model construction module, for being adjusted optimization to initial risks prediction model according to project sample data and constructing
Default risk forecast model.
In one embodiment, model construction module may include:
Data extracting unit, for extracting decision factor sequence and Sample Risk score from project sample data.
Label setup unit divides threshold value for obtaining initial category, divides threshold value and Sample Risk according to initial category
Score setting item risk label.
Sample set generation unit, for generating project sample set according to decision factor sequence and project risk class label.
Function acquiring unit, for obtaining preset model loss function.
Model training unit loses letter according to preset model for project sample set to be inputted initial risks prediction model
Number carries out model training optimization and constructs to obtain default risk forecast model.
In one embodiment, real-time data acquisition module 440 may include:
Coefficient calculation unit, for calculating project risk coefficient of variation according to project sample data.
Period searching unit, for searching data sample period corresponding with project risk coefficient of variation.
Real time data acquisition unit, for obtaining the newest decision factor data of real-time update.
Real time data extraction unit, for being extracted from newest decision factor data according to data sample period in real time certainly
Plan factor data.
In one embodiment, factor screening module may include:
Contribution rate computing unit, for calculating the factor contribution rate of each quantizing factor in quantization factor data.
Threshold value acquiring unit, for obtaining default contribution rate threshold value.
Compare screening unit, the quantizing factor for factor contribution rate to be greater than default contribution rate threshold value screen for decision because
Son.
In one embodiment, factor screening module may include:
Contribution rate computing unit, for calculating the factor contribution rate of each quantizing factor in quantization factor data；
Prescreening unit, for carrying out prescreening to quantizing factor according to factor contribution rate；
Expression formula generation unit obtains factor principal component table for carrying out principal component analysis to the quantizing factor after prescreening
Up to formula；
Factor construction unit, for constructing decision factor according to factor principal component expression formula.
In one embodiment, real-time model generation module 450 may include:
Probability generation unit, it is general for the default risk forecast model of Real-time Decision factor data input to be obtained risk
Rate.
Score converting unit, for risk probability to be converted to risk level score.
Tag calculation unit divides threshold value for dynamically adjustment risk, calculates each dynamic according to risk level score
The risk of adjustment divides the corresponding dynamic class label of threshold value.
True tag acquiring unit, for obtaining the corresponding true class label of Real-time Decision factor data.
Curve computation unit, for calculating threshold classification effect curve according to dynamic class label and true class label.
Model generation unit, for determining best category division threshold value according to threshold classification effect curve, according to default wind
Dangerous prediction model and best category division threshold value generate real-time risk forecast model.
In one embodiment, a kind of computer equipment is provided, which can be server, internal junction
Composition can be as shown in Figure 5.The computer equipment include by system bus connect processor, memory, network interface and
Database.Wherein, the processor of the computer equipment is for providing calculating and control ability.The memory packet of the computer equipment
Include non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system, computer program and data
Library.The built-in storage provides environment for the operation of operating system and computer program in non-volatile memory medium.The calculating
The database of machine equipment is used for stored items decision in the face of risk related data.The network interface of the computer equipment is used for and outside
Terminal passes through network connection communication.To realize a kind of Decision of Project Risk method when the computer program is executed by processor.
It will be understood by those skilled in the art that structure shown in Fig. 5, only part relevant to application scheme is tied
The block diagram of structure does not constitute the restriction for the computer equipment being applied thereon to application scheme, specific computer equipment
It may include perhaps combining certain components or with different component layouts than more or fewer components as shown in the figure.
In one embodiment, a kind of computer equipment, including memory and processor are provided, which is stored with
Computer program, which performs the steps of when executing computer program receives Decision of Project Risk instruction；From project wind
Project category and project information data are obtained in dangerous decision instruction；Default risk forecast model corresponding with project category is searched,
And obtain the decision parameters in default risk forecast model；Acquire Real-time Decision factor data corresponding with decision parameters；According to
Real-time Decision factor data divides threshold value to the risk of default risk forecast model and is adjusted to obtain real-time risk profile
Model；Project information data are inputted into real-time risk forecast model and obtain assessment risk, are generated according to assessment risk
Decision of Project Risk suggestion.
In one embodiment, acquisition and project category pair are also performed the steps of when processor executes computer program
The quantizing factor data and initial risks prediction model answered；Decision factor is filtered out from quantizing factor data；According to decision because
Sub- acquisition project sample data；Optimization is adjusted to initial risks prediction model according to project sample data and constructs default wind
Dangerous prediction model.
In one embodiment, it is realized when processor executes computer program pre- to initial risks according to project sample data
It surveys model to be adjusted optimization and be also used to when constructing the step of default risk forecast model: extracting and determine from project sample data
Plan factor sequence and Sample Risk score；It obtains initial category and divides threshold value, threshold value and Sample Risk are divided according to initial category
Score setting item risk label；Project sample set is generated according to decision factor sequence and project risk class label；It obtains
Take preset model loss function；Project sample set is inputted into initial risks prediction model, is carried out according to preset model loss function
Model training optimizes and constructs to obtain default risk forecast model.
In one embodiment, it is realized when processor executes computer program and acquires Real-time Decision corresponding with decision parameters
It is also used to when the step of factor data: project risk coefficient of variation is calculated according to project sample data；It searches and project risk wave
The dynamic corresponding data sample period of coefficient；Obtain the newest decision factor data of real-time update；According to data sample period from most
Real-time Decision factor data is extracted in new decision factor data.
In one embodiment, processor execute computer program when realize filtered out from quantizing factor data decision because
It is also used to when the step of son: calculating the factor contribution rate of each quantizing factor in quantization factor data；Obtain default contribution rate threshold value；
The quantizing factor that factor contribution rate is greater than default contribution rate threshold value is screened as decision factor.
In one embodiment, processor execute computer program when realize filtered out from quantizing factor data decision because
It is also used to when the step of son: calculating the factor contribution rate of each quantizing factor in quantization factor data；According to factor contribution rate to amount
Change the factor and carries out prescreening；Principal component analysis is carried out to the quantizing factor after prescreening and obtains factor principal component expression formula；According to
Factor principal component expression formula constructs decision factor.
In one embodiment, it realizes according to Real-time Decision factor data when processor executes computer program to default wind
The risk of dangerous prediction model divides threshold value and is adjusted to be also used to when obtaining the step of real-time risk forecast model: will be real-time
Decision factor data input default risk forecast model and obtain risk probability；Risk probability is converted into risk level
Score；Dynamic adjustment risk divides threshold value, divides threshold according to the risk that risk level score calculates each dynamic adjustment
It is worth corresponding dynamic class label；Obtain the corresponding true class label of Real-time Decision factor data；According to dynamic class label
Threshold classification effect curve is calculated with true class label；Best category division threshold value is determined according to threshold classification effect curve,
Real-time risk forecast model is generated according to default risk forecast model and best category division threshold value.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
Machine program performs the steps of when being executed by processor receives Decision of Project Risk instruction；It is obtained from Decision of Project Risk instruction
Take project category and project information data；Default risk forecast model corresponding with project category is searched, and obtains default risk
Decision parameters in prediction model；Acquire Real-time Decision factor data corresponding with decision parameters；According to Real-time Decision because of subnumber
Threshold value is divided according to the risk to default risk forecast model to be adjusted to obtain real-time risk forecast model；By project information
Data input real-time risk forecast model and obtain assessment risk, generate Decision of Project Risk according to assessment risk and build
View.
In one embodiment, acquisition and project category are also performed the steps of when computer program is executed by processor
Corresponding quantizing factor data and initial risks prediction model；Decision factor is filtered out from quantizing factor data；According to decision
The factor acquires project sample data；Optimization is adjusted to initial risks prediction model according to project sample data and is constructed default
Risk forecast model.
In one embodiment, it realizes according to project sample data when computer program is executed by processor to initial risks
Prediction model is adjusted optimization and is also used to when constructing the step of default risk forecast model: extracting from project sample data
Decision factor sequence and Sample Risk score；It obtains initial category and divides threshold value, threshold value and sample wind are divided according to initial category
Dangerous score setting item risk label；Project sample set is generated according to decision factor sequence and project risk class label；
Obtain preset model loss function；By project sample set input initial risks prediction model, according to preset model loss function into
Row model training optimizes and constructs to obtain default risk forecast model.
In one embodiment, it realizes that acquisition is corresponding with decision parameters when computer program is executed by processor to determine in real time
It is also used to when the step of plan factor data: project risk coefficient of variation is calculated according to project sample data；Lookup and project risk
The corresponding data sample period of coefficient of variation；Obtain the newest decision factor data of real-time update；According to data sample period from
Real-time Decision factor data is extracted in newest decision factor data.
In one embodiment, it is realized when computer program is executed by processor and filters out decision from quantizing factor data
It is also used to when the step of the factor: calculating the factor contribution rate of each quantizing factor in quantization factor data；Obtain default contribution rate threshold
Value；The quantizing factor that factor contribution rate is greater than default contribution rate threshold value is screened as decision factor.
In one embodiment, it is realized when computer program is executed by processor and filters out decision from quantizing factor data
It is also used to when the step of the factor: calculating the factor contribution rate of each quantizing factor in quantization factor data；According to factor contribution rate pair
Quantizing factor carries out prescreening；Principal component analysis is carried out to the quantizing factor after prescreening and obtains factor principal component expression formula；Root
Decision factor is constructed according to factor principal component expression formula.
In one embodiment, it realizes according to Real-time Decision factor data when computer program is executed by processor to default
The risk of risk forecast model divides threshold value and is adjusted to be also used to when obtaining the step of real-time risk forecast model: will be real
When decision factor data input default risk forecast model and obtain risk probability；Risk probability is converted into risk water
Flat score；Dynamic adjustment risk divides threshold value, is divided according to the risk that risk level score calculates each dynamic adjustment
The corresponding dynamic class label of threshold value；Obtain the corresponding true class label of Real-time Decision factor data；According to dynamic class mark
Label and true class label calculate threshold classification effect curve；Best category division threshold is determined according to threshold classification effect curve
Value generates real-time risk forecast model according to default risk forecast model and best category division threshold value.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
To any reference of memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
Each technical characteristic of above embodiments can be combined arbitrarily, for simplicity of description, not to above-described embodiment
In each technical characteristic it is all possible combination be all described, as long as however, the combination of these technical characteristics be not present lance
Shield all should be considered as described in this specification.
The several embodiments of the application above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the concept of this application, various modifications and improvements can be made, these belong to the protection of the application
Range.Therefore, the scope of protection shall be subject to the appended claims for the application patent.
Claims (10)
1. a kind of Decision of Project Risk method, which comprises
Receive Decision of Project Risk instruction；
Project category and project information data are obtained from Decision of Project Risk instruction；
Default risk forecast model corresponding with the project category is searched, and obtains determining in the default risk forecast model
Plan parameter；
Acquire Real-time Decision factor data corresponding with the decision parameters；
Threshold value is divided according to risk of the Real-time Decision factor data to the default risk forecast model to be adjusted
Obtain real-time risk forecast model；
The project information data are inputted into the real-time risk forecast model and obtain assessment risk, according to the assessment wind
Dangerous classification generates Decision of Project Risk suggestion.
2. the method according to claim 1, wherein the generating mode of the default risk forecast model, comprising:
Obtain quantizing factor data corresponding with the project category and initial risks prediction model；
Decision factor is filtered out from the quantizing factor data；
Project sample data is acquired according to the decision factor；
Optimization is adjusted to the initial risks prediction model according to the project sample data and constructs default risk profile
Model.
3. according to the method described in claim 2, it is characterized in that, it is described according to the project sample data to the initial wind
Dangerous prediction model is adjusted optimization and constructs default risk forecast model, comprising:
Decision factor sequence and Sample Risk score are extracted from the project sample data；
It obtains initial category and divides threshold value, threshold value and the Sample Risk score setting item wind are divided according to the initial category
Dangerous class label；
Project sample set is generated according to the decision factor sequence and the project risk class label；
Obtain preset model loss function；
The project sample set is inputted into the initial risks prediction model, model is carried out according to the preset model loss function
Training optimizes and constructs to obtain default risk forecast model.
4. according to the method described in claim 2, it is characterized in that, acquisition Real-time Decision corresponding with the decision parameters
Factor data, comprising:
Project risk coefficient of variation is calculated according to the project sample data；
Search data sample period corresponding with the project risk coefficient of variation；
Obtain the newest decision factor data of real-time update；
Real-time Decision factor data is extracted from the newest decision factor data according to the data sample period.
5. according to the method described in claim 2, it is characterized in that, it is described filtered out from the quantizing factor data decision because
Son, comprising:
Calculate the factor contribution rate of each quantizing factor in the quantizing factor data；
Obtain default contribution rate threshold value；
The quantizing factor that the factor contribution rate is greater than the default contribution rate threshold value is screened as decision factor.
6. according to the method described in claim 2, it is characterized in that, it is described filtered out from the quantizing factor data decision because
Son, comprising:
Calculate the factor contribution rate of each quantizing factor in the quantizing factor data；
Prescreening is carried out to the quantizing factor according to the factor contribution rate；
Principal component analysis is carried out to the quantizing factor after prescreening and obtains factor principal component expression formula；
Decision factor is constructed according to the factor principal component expression formula.
7. the method according to claim 1, wherein it is described according to the Real-time Decision factor data to described pre-
If the risk of risk forecast model divides threshold value and is adjusted to obtain real-time risk forecast model, comprising:
The Real-time Decision factor data is inputted into the default risk forecast model and obtains risk probability；
The risk probability is converted into risk level score；
Dynamic adjusts the risk and divides threshold value, and the risk of each dynamic adjustment is calculated according to the risk level score
The corresponding dynamic class label of category division threshold value；
Obtain the corresponding true class label of the Real-time Decision factor data；
Threshold classification effect curve is calculated according to the dynamic class label and the true class label；
Best category division threshold value is determined according to the threshold classification effect curve, according to the default risk forecast model and institute
It states best category division threshold value and generates real-time risk forecast model.
8. a kind of Decision of Project Risk device, which is characterized in that described device includes:
Command reception module, for receiving Decision of Project Risk instruction；
Project data obtains module, for obtaining project category and project information data from Decision of Project Risk instruction；
Model searching module for searching default risk forecast model corresponding with the project category, and obtains described default
Decision parameters in risk forecast model；
Real-time data acquisition module, for acquiring Real-time Decision factor data corresponding with the decision parameters；
Real-time model generation module, for the risk according to the Real-time Decision factor data to the default risk forecast model
Category division threshold value is adjusted to obtain real-time risk forecast model；
Decision recommendation module obtains assessment risk class for the project information data to be inputted the real-time risk forecast model
Not, Decision of Project Risk suggestion is generated according to the assessment risk.
9. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In the step of processor realizes any one of claims 1 to 7 the method when executing the computer program.
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 method described in any one of claims 1 to 7 is realized when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910012593.6A CN109816221B (en) | 2019-01-07 | 2019-01-07 | Project risk decision method, apparatus, computer device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910012593.6A CN109816221B (en) | 2019-01-07 | 2019-01-07 | Project risk decision method, apparatus, computer device and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109816221A true CN109816221A (en) | 2019-05-28 |
| CN109816221B CN109816221B (en) | 2024-01-02 |
Family
ID=66604016
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910012593.6A Active CN109816221B (en) | 2019-01-07 | 2019-01-07 | Project risk decision method, apparatus, computer device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109816221B (en) |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110399302A (en) * | 2019-07-26 | 2019-11-01 | 中国工商银行股份有限公司 | Risk Identification Method, device, electronic equipment and the medium of Software Testing Project |
| CN110619462A (en) * | 2019-09-10 | 2019-12-27 | 苏州方正璞华信息技术有限公司 | Project quality assessment method based on AI model |
| CN110929879A (en) * | 2019-11-06 | 2020-03-27 | 南方电网财务有限公司 | Business decision logic updating method based on decision engine and model platform |
| CN111080102A (en) * | 2019-12-05 | 2020-04-28 | 上海创能国瑞能源互联网技术有限公司 | Method for predicting and evaluating contract energy management project |
| CN111160701A (en) * | 2019-11-28 | 2020-05-15 | 万翼科技有限公司 | Engineering project risk monitoring method, system, computer equipment and storage medium |
| CN111160662A (en) * | 2019-12-31 | 2020-05-15 | 咪咕文化科技有限公司 | Risk prediction method, electronic equipment and storage medium |
| CN111311309A (en) * | 2020-01-19 | 2020-06-19 | 百度在线网络技术（北京）有限公司 | User satisfaction determining method, device, equipment and medium |
| CN112001629A (en) * | 2020-08-24 | 2020-11-27 | 北京鸿联九五信息产业有限公司 | Risk early warning method, device, system and storage medium |
| CN112308452A (en) * | 2020-11-18 | 2021-02-02 | 莆田学院 | Power grid project auxiliary decision-making method based on machine learning and storage medium |
| CN112308404A (en) * | 2020-10-29 | 2021-02-02 | 广东电网有限责任公司 | Project risk management method and device, electronic equipment and storage medium |
| CN112465011A (en) * | 2020-11-25 | 2021-03-09 | 平安医疗健康管理股份有限公司 | Project risk prediction method and system based on project research and development process |
| CN112613718A (en) * | 2020-12-17 | 2021-04-06 | 武汉达梦数据技术有限公司 | Specific site risk assessment method and device |
| CN113094270A (en) * | 2021-04-13 | 2021-07-09 | 中国工商银行股份有限公司 | Project automation verification method and device |
| CN113256181A (en) * | 2021-07-13 | 2021-08-13 | 平安科技（深圳）有限公司 | Risk factor prediction method, device, equipment and medium |
| CN113537662A (en) * | 2020-04-15 | 2021-10-22 | 上海汽车集团股份有限公司 | Risk detection method and device |
| CN113592538A (en) * | 2021-07-01 | 2021-11-02 | 广州珠合工程技术有限公司 | Method and system for calculating approximate construction cost |
| CN113590691A (en) * | 2021-08-04 | 2021-11-02 | 浙江网商银行股份有限公司 | Target object processing method and device |
| CN113610373A (en) * | 2021-07-28 | 2021-11-05 | 上海德衡数据科技有限公司 | Information decision processing method and system based on intelligent manufacturing |
| CN114037314A (en) * | 2021-11-17 | 2022-02-11 | 北京中百信信息技术股份有限公司 | Information system engineering supervision project risk dynamic analysis method |
| CN114429308A (en) * | 2022-02-08 | 2022-05-03 | 河南鑫安利安全科技股份有限公司 | Enterprise security risk assessment method and system based on big data |
| CN114662824A (en) * | 2020-12-23 | 2022-06-24 | 顺丰恒通支付有限公司 | Wind control strategy switching method and device, computer equipment and storage medium |
| CN114677184A (en) * | 2022-05-25 | 2022-06-28 | 国网浙江省电力有限公司宁波供电公司 | Data processing method and platform based on operation decision auxiliary model |
| CN116302513A (en) * | 2023-02-28 | 2023-06-23 | 易方达基金管理有限公司 | Quantization factor processing method, quantization factor processing device, computer equipment and readable storage medium |
| CN117541044A (en) * | 2023-07-28 | 2024-02-09 | 中科乾唐(北京)科技股份公司 | Project classification method, system, medium and equipment based on project risk analysis |
| CN118229071A (en) * | 2024-03-07 | 2024-06-21 | 杭州小策科技有限公司 | Deep learning-based risk dynamic evolution method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103208039A (en) * | 2012-01-13 | 2013-07-17 | 株式会社日立制作所 | Method and device for evaluating software project risks |
| WO2015137970A1 (en) * | 2014-03-14 | 2015-09-17 | Mun Johnathan | Qualitative and quantitative modeling of enterprise risk management and risk registers |
| CN106469340A (en) * | 2015-08-17 | 2017-03-01 | 苏州优估营网络科技有限公司 | A kind of compound Project Risk Assessment model based on neutral net |
| CN106844663A (en) * | 2017-01-23 | 2017-06-13 | 中国石油大学(华东) | A kind of ship collision methods of risk assessment and system based on data mining |
-
2019
- 2019-01-07 CN CN201910012593.6A patent/CN109816221B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103208039A (en) * | 2012-01-13 | 2013-07-17 | 株式会社日立制作所 | Method and device for evaluating software project risks |
| WO2015137970A1 (en) * | 2014-03-14 | 2015-09-17 | Mun Johnathan | Qualitative and quantitative modeling of enterprise risk management and risk registers |
| CN106469340A (en) * | 2015-08-17 | 2017-03-01 | 苏州优估营网络科技有限公司 | A kind of compound Project Risk Assessment model based on neutral net |
| CN106844663A (en) * | 2017-01-23 | 2017-06-13 | 中国石油大学(华东) | A kind of ship collision methods of risk assessment and system based on data mining |
Cited By (36)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110399302A (en) * | 2019-07-26 | 2019-11-01 | 中国工商银行股份有限公司 | Risk Identification Method, device, electronic equipment and the medium of Software Testing Project |
| CN110619462A (en) * | 2019-09-10 | 2019-12-27 | 苏州方正璞华信息技术有限公司 | Project quality assessment method based on AI model |
| CN110929879A (en) * | 2019-11-06 | 2020-03-27 | 南方电网财务有限公司 | Business decision logic updating method based on decision engine and model platform |
| CN111160701A (en) * | 2019-11-28 | 2020-05-15 | 万翼科技有限公司 | Engineering project risk monitoring method, system, computer equipment and storage medium |
| CN111080102A (en) * | 2019-12-05 | 2020-04-28 | 上海创能国瑞能源互联网技术有限公司 | Method for predicting and evaluating contract energy management project |
| CN111160662A (en) * | 2019-12-31 | 2020-05-15 | 咪咕文化科技有限公司 | Risk prediction method, electronic equipment and storage medium |
| CN111311309A (en) * | 2020-01-19 | 2020-06-19 | 百度在线网络技术（北京）有限公司 | User satisfaction determining method, device, equipment and medium |
| CN113537662A (en) * | 2020-04-15 | 2021-10-22 | 上海汽车集团股份有限公司 | Risk detection method and device |
| CN112001629A (en) * | 2020-08-24 | 2020-11-27 | 北京鸿联九五信息产业有限公司 | Risk early warning method, device, system and storage medium |
| CN112308404A (en) * | 2020-10-29 | 2021-02-02 | 广东电网有限责任公司 | Project risk management method and device, electronic equipment and storage medium |
| CN112308404B (en) * | 2020-10-29 | 2023-02-17 | 广东电网有限责任公司 | Project risk management method and device, electronic equipment and storage medium |
| CN112308452A (en) * | 2020-11-18 | 2021-02-02 | 莆田学院 | Power grid project auxiliary decision-making method based on machine learning and storage medium |
| CN112465011A (en) * | 2020-11-25 | 2021-03-09 | 平安医疗健康管理股份有限公司 | Project risk prediction method and system based on project research and development process |
| CN112465011B (en) * | 2020-11-25 | 2022-08-02 | 深圳平安医疗健康科技服务有限公司 | Project risk prediction method and system based on project research and development process |
| CN112613718A (en) * | 2020-12-17 | 2021-04-06 | 武汉达梦数据技术有限公司 | Specific site risk assessment method and device |
| CN112613718B (en) * | 2020-12-17 | 2023-12-01 | 武汉达梦数据技术有限公司 | Specific place risk assessment method and device |
| CN114662824A (en) * | 2020-12-23 | 2022-06-24 | 顺丰恒通支付有限公司 | Wind control strategy switching method and device, computer equipment and storage medium |
| CN113094270A (en) * | 2021-04-13 | 2021-07-09 | 中国工商银行股份有限公司 | Project automation verification method and device |
| CN113094270B (en) * | 2021-04-13 | 2024-03-26 | 中国工商银行股份有限公司 | Project automatic verification method and device |
| CN113592538A (en) * | 2021-07-01 | 2021-11-02 | 广州珠合工程技术有限公司 | Method and system for calculating approximate construction cost |
| CN113592538B (en) * | 2021-07-01 | 2024-05-31 | 广州珠合工程技术有限公司 | Project cost approximate calculation method and system |
| CN113256181A (en) * | 2021-07-13 | 2021-08-13 | 平安科技（深圳）有限公司 | Risk factor prediction method, device, equipment and medium |
| CN113610373B (en) * | 2021-07-28 | 2024-01-26 | 上海德衡数据科技有限公司 | Information decision processing method and system based on intelligent manufacturing |
| CN113610373A (en) * | 2021-07-28 | 2021-11-05 | 上海德衡数据科技有限公司 | Information decision processing method and system based on intelligent manufacturing |
| CN113590691B (en) * | 2021-08-04 | 2024-06-28 | 浙江网商银行股份有限公司 | Target object processing method and device |
| CN113590691A (en) * | 2021-08-04 | 2021-11-02 | 浙江网商银行股份有限公司 | Target object processing method and device |
| CN114037314A (en) * | 2021-11-17 | 2022-02-11 | 北京中百信信息技术股份有限公司 | Information system engineering supervision project risk dynamic analysis method |
| CN114429308A (en) * | 2022-02-08 | 2022-05-03 | 河南鑫安利安全科技股份有限公司 | Enterprise security risk assessment method and system based on big data |
| CN114429308B (en) * | 2022-02-08 | 2024-06-07 | 河南鑫安利安全科技股份有限公司 | Enterprise safety risk assessment method and system based on big data |
| CN114677184B (en) * | 2022-05-25 | 2022-08-26 | 国网浙江省电力有限公司宁波供电公司 | Data processing method and platform based on operation decision auxiliary model |
| CN114677184A (en) * | 2022-05-25 | 2022-06-28 | 国网浙江省电力有限公司宁波供电公司 | Data processing method and platform based on operation decision auxiliary model |
| CN116302513A (en) * | 2023-02-28 | 2023-06-23 | 易方达基金管理有限公司 | Quantization factor processing method, quantization factor processing device, computer equipment and readable storage medium |
| CN117541044A (en) * | 2023-07-28 | 2024-02-09 | 中科乾唐(北京)科技股份公司 | Project classification method, system, medium and equipment based on project risk analysis |
| CN117541044B (en) * | 2023-07-28 | 2024-06-11 | 中科乾唐(北京)科技股份公司 | Project classification method, system, medium and equipment based on project risk analysis |
| CN118229071A (en) * | 2024-03-07 | 2024-06-21 | 杭州小策科技有限公司 | Deep learning-based risk dynamic evolution method and system |
| CN118229071B (en) * | 2024-03-07 | 2024-09-20 | 杭州小策科技有限公司 | Deep learning-based risk dynamic evolution method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109816221B (en) | 2024-01-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109816221A (en) | Decision of Project Risk method, apparatus, computer equipment and storage medium | |
| Benidis et al. | Deep learning for time series forecasting: Tutorial and literature survey | |
| KR20210017342A (en) | Time series prediction method and apparatus based on past prediction data | |
| Zhang et al. | Deep learning models for price forecasting of financial time series: A review of recent advancements: 2020–2022 | |
| Llanos et al. | Causal treatment effects in time series: CO2 emissions and energy consumption effect on GDP | |
| CN112634992A (en) | Molecular property prediction method, training method of model thereof, and related device and equipment | |
| Zhang et al. | DoubleEnsemble: A new ensemble method based on sample reweighting and feature selection for financial data analysis | |
| Xing et al. | Predicting NVIDIA's Next-Day Stock Price: A Comparative Analysis of LSTM, MLP, ARIMA, and ARIMA-GARCH Models | |
| Guida et al. | Ensemble learning applied to quant equity: gradient boosting in a multifactor framework | |
| CN117540336A (en) | Time sequence prediction method and device and electronic equipment | |
| CN111815458A (en) | Dynamic investment portfolio configuration method based on fine-grained quantitative marking and integration method | |
| Mahar et al. | Bitcoin price prediction app using deep learning algorithm | |
| Gurav et al. | RNNLBL: A Recurrent Neural Network and Log Bilinear based Efficient Stock Forecasting Model | |
| Fior et al. | Exploring the use of data at multiple granularity levels in machine learning-based stock trading | |
| CN113793220A (en) | Stock market investment decision method based on artificial intelligence model and related equipment | |
| EP3739517A1 (en) | Image processing | |
| Nishitha et al. | Stock price prognosticator using machine learning techniques | |
| Ansary | Machine Learning for Predicting the Stock Price Direction with Trading Indicators | |
| CN110796006A (en) | Stock K-line image identification method and device, intelligent terminal and storage medium | |
| Uma et al. | EVALUATION OF ENSEMBLE LEARNING APPROACH FOR OPTIMISED STOCK PREDICTION | |
| CN114549213B (en) | Analysis management method, device and system for foundation products | |
| Koohfar et al. | Adjustable context-aware transformer | |
| Pratama | Multilayer Perceptron and Long Short-Term Memory for Predicting Indonesian Composite Stock Price Index Using Macroeconomic Factors | |
| US20240256981A1 (en) | Self-adaptive multi-model approach in representation feature space for propensity to action | |
| Victor | Analyzing Stock Market Trends using Social Media user Moods and Social Influence using Lightweight BERT-BiGRU-A Model |
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 |