CN109740865A - Methods of risk assessment, system, equipment and storage medium - Google Patents
Methods of risk assessment, system, equipment and storage medium Download PDFInfo
- Publication number
- CN109740865A CN109740865A CN201811529145.5A CN201811529145A CN109740865A CN 109740865 A CN109740865 A CN 109740865A CN 201811529145 A CN201811529145 A CN 201811529145A CN 109740865 A CN109740865 A CN 109740865A
- Authority
- CN
- China
- Prior art keywords
- risk
- node
- destination node
- coefficient
- 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
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention provides a kind of methods of risk assessment, system, equipment and storage medium, this method comprises: obtaining itself risk and the first anti-risk coefficient of destination node and the destination node that the risk assessment request is directed toward after detecting risk assessment request;All first nodes being directly linked with the destination node are obtained from preset related network, and obtain the first input risk that all first nodes are transmitted to the destination node;Preset association conduction model is inputted using itself risk of the destination node, the first input risk and the first anti-risk coefficient as ginseng is entered, obtains the output risk of the destination node.The present invention is based on Recognition with Recurrent Neural Network to be associated with conduction model, realizes the quantization of the association conduction risk based on enterprise or a human world incidence relation.
Description
Technical field
The present invention relates to technical field of data processing more particularly to a kind of methods of risk assessment, system, equipment and storage to be situated between
Matter.
Background technique
Financial institution borrowed when handling loan transaction before, borrow in, borrow after this entire loan process air control
Management.When to loan enterprises or personal progress risk monitoring and control, need to assess loan enterprises or individual risk's index.
Wherein, the risk index of loan enterprises or individual are influenced by its affiliated enterprise or individual, but this based on association
Enterprise or personal bring risk are not quantized, thus are not often included in loan enterprises or individual risk's Index Assessment system
In, the especially enterprise of interval multilayer incidence relation or the personal influence to the enterprise or individual is not considered substantially, leads to wind
Dangerous assessment errors are larger.
Summary of the invention
The main purpose of the present invention is to provide a kind of methods of risk assessment, it is intended to solve based on affiliated enterprise or personal band
The technical problem that the risk come is not quantized, causes risk assessment error larger.
To achieve the above object, the present invention provides a kind of methods of risk assessment, which is characterized in that the methods of risk assessment
The following steps are included:
After detecting risk assessment request, destination node and the target section that the risk assessment request is directed toward are obtained
Itself risk and the first anti-risk coefficient of point;
All first nodes being directly linked with the destination node are obtained from preset related network, and obtain all the
One node is transmitted to the first input risk of the destination node;
Itself risk of the destination node, the first input risk and the first anti-risk coefficient is defeated as entering to join
Enter preset association conduction model, obtains the output risk of the destination node.
Optionally, described to obtain the step of all first nodes are transmitted to the first input risk of destination node packet
It includes:
The output risk of each first node is obtained, and obtains each first node respectively to the destination node
The risk coefficient of conductivity;
According to the output risk of each first node and the risk coefficient of conductivity, calculates and obtain all first nodes biographies
Lead the first input risk of the destination node.
Optionally, the step of output risk for obtaining each first node includes:
Itself risk and the second anti-risk coefficient of each first node are obtained, is obtained and each first node respectively
The second node of direct correlation, and the second input risk that the second node is transmitted to corresponding first node is obtained respectively;
It calculates and obtains according to itself risk of each first node, the second input risk and the second anti-risk coefficient
The output risk of each first node.
Optionally, the step of risk coefficient of conductivity for obtaining each first node respectively to destination node packet
It includes:
Obtain node type, strength of association and the association type of the destination node and each first node;
The destination node and each described first are determined respectively according to the node type, strength of association and association type
The risk coefficient of conductivity between node.
Optionally, described after detecting risk assessment request, obtain the destination node that the risk assessment request is directed toward
And the destination node itself risk and the first anti-risk coefficient the step of before include:
When the venture influence factor for detecting arbitrary node in preset related network changes, triggering, which generates, is directed toward institute
The risk assessment request of arbitrary node is stated, the venture influence factor includes the arbitrary node itself risk, preset association net
Other nodes are transmitted to the input risk and at least one of anti-risk coefficient of the arbitrary node in network.
Optionally, described after detecting risk assessment request, obtain the destination node that the risk assessment request is directed toward
And the destination node itself risk and the first anti-risk coefficient the step of before include:
Training sample set is obtained, association conduction model is trained using the training sample set, obtains optimal weights matrix
Parameter trains the preset association conduction model of the matrix parameter containing optimal weights.
Optionally, described by itself risk of the destination node, the first input risk and the first anti-risk system
Include: after the step of number inputs preset association conduction model as ginseng is entered, and obtains the output risk of the destination node
The practical risk of the destination node is inputted the association guided modes by the practical risk for obtaining the destination node
Type, so that association conduction model is according to the discrepancy adjustment optimal weights matrix of the practical risk of the destination node and output risk
Parameter.
In addition, to achieve the above object, the present invention also provides a kind of risk evaluating system, the risk evaluating system packet
It includes:
First obtains module, the mesh being directed toward for after detecting risk assessment request, obtaining the risk assessment request
Mark itself risk and the first anti-risk coefficient of node and the destination node;
Second obtains module, for obtaining be directly linked with the destination node all first from preset related network
Node, and obtain the first input risk that all first nodes are transmitted to the destination node;
Computing module, for by itself risk of the destination node, the first input risk and described first anti-risk
Coefficient inputs preset association conduction model as ginseng is entered, and obtains the output risk of the destination node.
In addition, to achieve the above object, the present invention also provides a kind of risk assessment equipment, the risk assessment equipment includes
Processor, memory and the risk assessment procedures that can be executed on the memory and by the processor are stored in, wherein institute
When stating risk assessment procedures and being executed by the processor, realize such as the step of above-mentioned methods of risk assessment.
In addition, to achieve the above object, the present invention also provides a kind of storage medium, being stored on the storage medium risky
Appraisal procedure, wherein realizing when the risk assessment procedures are executed by processor such as the step of above-mentioned methods of risk assessment.
The target that the embodiment of the present invention is directed toward by after detecting risk assessment request, obtaining the risk assessment request
Node is obtained and is directly closed in itself risk, the first anti-risk coefficient and the preset related network of destination node with destination node
All first nodes of connection are transmitted to the first input risk of destination node, according to itself risk of destination node, the first input
Risk and the output risk of the first anti-risk coefficient assessment destination node are, it can be achieved that risk is conducted in the association based on incidence relation
Quantization, that is, realize the amount of the association conduction risk between the enterprise or node in preset related network, with incidence relation
Change.
Detailed description of the invention
Fig. 1 is the risk assessment device structure schematic diagram for the hardware running environment that the embodiment of the present invention is related to;
Fig. 2 is the network structure exemplary diagram of RNN model in the prior art;
Fig. 3 is part of nodes exemplary diagram in the association conduction model of proposition of the embodiment of the present invention;
Fig. 4 is the flow diagram of methods of risk assessment first embodiment of the present invention;
Fig. 5 is another exemplary diagram of part of nodes in the association conduction model of proposition of the embodiment of the present invention;
Fig. 6 is preset one schematic diagram of related network in the embodiment of the present invention;
Fig. 7 is the another exemplary diagram of part of nodes in the association conduction model of proposition of the embodiment of the present invention;
Fig. 8 is the functional block diagram of risk evaluating system first embodiment of the present invention.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific embodiment
It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not intended to limit the present invention.
Referring to Figure 1, Fig. 1 is the hardware structural diagram of risk assessment equipment provided by the present invention.
The risk assessment equipment can be PC, be also possible to smart phone, tablet computer, portable computer, desk-top meter
The equipment equipment having a display function such as calculation machine, optionally, the risk assessment equipment can be server apparatus, and there are risks
The rear end management system of assessment, user are managed risk assessment equipment by the rear end management system.
The risk assessment equipment may include: the components such as processor 101 and memory 201.In the risk assessment
In equipment, the processor 101 is connect with the memory 201, is stored with risk assessment procedures on the memory 201, place
Reason device 101 can call the risk assessment procedures stored in memory 201, and realize such as each embodiment of following methods of risk assessment
The step of.
The memory 201 can be used for storing software program and various data.Memory 201 can mainly include storage
Program area and storage data area, wherein storing program area can application program needed for storage program area, at least one function
(such as risk assessment procedures) etc.;Storage data area may include database, such as the database of memory node information.In addition, storage
Device 201 may include high-speed random access memory, can also include nonvolatile memory, for example, at least a disk storage
Device, flush memory device or other volatile solid-state parts.
Processor 101 is the control centre of risk assessment equipment, utilizes various interfaces and the entire risk assessment of connection
The various pieces of equipment by running or execute the software program and/or module that are stored in memory 201, and are called and are deposited
The data in memory 201 are stored up, the various functions and processing data of risk assessment equipment are executed, thus to risk assessment equipment
Carry out integral monitoring.Processor 101 may include one or more processing units;Optionally, processor 101 can be integrated using processing
Device and modem processor, wherein the main processing operation system of application processor, user interface and application program etc., modulation
Demodulation processor mainly handles wireless communication.It is understood that above-mentioned modem processor can not also be integrated into processing
In device 101.
It will be understood by those skilled in the art that risk assessment device structure shown in Fig. 1 is not constituted to risk assessment
The restriction of equipment may include perhaps combining certain components or different component cloth than illustrating more or fewer components
It sets.
Based on above-mentioned hardware configuration, each embodiment of the method for the present invention is proposed.
The embodiment of the present invention provides a kind of methods of risk assessment, i.e., a kind of method of quantization association conduction risk, main logical
The association conduction model established based on Recognition with Recurrent Neural Network technology is crossed, realizes that the association based on enterprise or a human world incidence relation passes
The quantization of wind-guiding danger.Association guided modes are input to using the risk of enterprise any in related network or individual and its generation as ginseng is entered
Type can get in related network with this enterprise or personal all affiliated enterprises with incidence relation or individual (enterprise and enterprise
Industry, individual and personal and enterprise and individual), and all affiliated enterprises or personal biography are transmitted to by this enterprise or individual
Wind-guiding danger.Wherein, for association conduction model based on related network, related network is by enterprise and enterprise with incidence relation
Therefore industry, the personal relational network with personal and enterprise and personal accomplishment node composition hereinafter, are indicated with " node "
Enterprise or individual in related network.In addition, also comprising the strength of association between associated nodes in related network, i.e., associated nodes it
Between incidence relation power.
For ease of understanding, the association conduction model proposed to the embodiment of the present invention does brief explanation.
Being associated with conduction model is a kind of Recognition with Recurrent Neural Network model.RNN (Recurrent Neuron Network) circulation
Nerual network technique is the neural network of a kind of pair of sequence data modeling, the i.e. output of a sequence current output and front
It is related.The specific form of expression is that network can remember the information of front and be applied in the calculating currently exported, i.e., hidden
Node between hiding layer is no longer connectionless but has connection, and not only the output including input layer is also wrapped for the input of hidden layer
Include the output of last moment hidden layer.
It is illustrated in figure 2 the network structure exemplary diagram of RNN model in the prior art, wherein
xtIt is the input of t moment, stIt is the hidden state of t moment, the hidden state based on last moment and current input obtain:
st=f (Uxt+Wst-1), wherein f is usually nonlinear activation primitive, is calculating s0When, i.e., the hiding stratiform at first moment
State needs to use s-1, but itself and be not present, be generally set to 0 in the implementation;
otIndicate the output of t moment, ot=softmax (Vst);
In RNN, all levels share same parameter, weight matrix parameter U, V, W in example as above.U is input
Layer arrives the weight matrix of hidden layer, and V is weight matrix of the hidden layer to output layer, and weight matrix W is exactly the hidden layer last time
It is worth the weight as input this time.
Fig. 3 is part of nodes exemplary diagram in the association conduction model of proposition of the embodiment of the present invention.
At the time of node is equivalent in above-mentioned example, wherein the output risk of node A is expressed as RISKoutA, node B biography
The defeated input risk to node A is RISKInBA, the input risk of other node-node transmissions to node A are in related network
(RISKIn1A+...+RISKInnA), itself risk of node A is RISKselfAAnd the anti-risk coefficient a of node AA, then
Have:
RISKoutA=f ((RISKIn1A+...+RISKInnA+RISKInBA+RISKselfA),aA)
It will be appreciated by those skilled in the art that the subscript of above-mentioned expression only does example for explaining, it is not used in restriction
This programme.
Before executing each step of the embodiment of the present invention, building association conduction model, i.e., include: to obtain before step S10
Training sample set is taken, association conduction model is trained using the training sample set, optimal weights matrix parameter is obtained, trains
The preset association conduction model of the matrix parameter containing optimal weights.
Enterprise/the individual with incidence relation is obtained, enterprise/individual with incidence relation is divided into an associated group,
Wherein, the enterprise/individual includes enterprise and enterprise, enterprise and individual, and individual and individual;
The risk information of each associated group is obtained respectively, and the risk information includes risk time of origin, risk occurrence cause
And risk of loss etc., the risk information manually can be obtained and be inputted by user, can also grabbed from network, or from phase
Close industrial and commercial data acquisition;
Using the risk information of associated group as training sample set, training association conduction model obtains optimal model parameters,
In, the algorithm of training association conduction model can be back-propagation algorithm or propagated forward algorithm, and related algorithm is the prior art,
It does not repeat herein;By training input parameter (certain enterprise of such as occurrence risk and its risk of generation) input association guided modes
Type obtains prediction result (such as by the enterprise of the business impact of occurrence risk and its by shadow by the operation and processing inside model
Risk after sound), prediction result is compared with actual result, the difference between prediction result and actual result is lost
Weight matrix parameter is adjusted to optimal value by minimizing loss function, to generate square containing optimal weights by function representation
The association conduction model of battle array parameter.Wherein, loss function is the prior art, is not repeated.
Optionally, methods of risk assessment provided in an embodiment of the present invention is applied to risk evaluating system, the risk assessment system
System is configured in risk assessment equipment, can also be individually present, " system " hereinafter occurred refers to risk evaluating system.Risk assessment
The page in system can show the real-time value-at-risk of related network Zhong Ge enterprise (node), or not show the real-time wind of each enterprise
Danger value only calculates in real-time from the background or timing, when any business risk value is greater than preset value, the business risk value is prompted to be greater than
Preset value can prompt user, such as display value-at-risk to be greater than the enterprise of preset value by prompting message or in a manner of highlighting etc.;
Optionally, related network Zhong Ge enterprise can also can be shown with tabular form, not limited herein with map view or netted display
Its display mode.
It is the flow diagram of methods of risk assessment first embodiment of the present invention referring to Fig. 4, Fig. 4.
In the present embodiment, the methods of risk assessment the following steps are included:
Step S10 obtains the destination node and institute that the risk assessment request is directed toward after detecting risk assessment request
State itself risk and the first anti-risk coefficient of destination node;
Risk assessment request, i.e., carry out the request of risk assessment to destination node, and destination node refers in preset related network
The node that risk assessment request is directed toward, destination node can be the arbitrary node in preset related network, wherein can pass through parsing
Risk assessment request obtains destination node.
Risk assessment request, can be triggered by user's operation, can also be by if user clicks default risk assessment control triggering
The triggering of systemic presupposition condition, is such as triggered by timing task.Different triggering modes correspond to different application scenarios, thus can
Can occur that mode is implemented as follows:
One, by user's operation triggering risk assessment request: when the node in preset related network is with map view/net-shaped
When the forms such as formula/tabular form are shown in the human-computer interaction interface of risk assessment equipment, the selected one or more nodes of user, then
Risk assessment control is triggered, generates the risk assessment request for being directed toward user's selected node, wherein the selected operation of user and risk
The modes such as input can be clicked for touch-screen input/voice input/keyboard input/mouse by assessing control trigger action;
Two, by the triggering risk assessment request of systemic presupposition condition: systemic presupposition condition can be clocked flip, i.e., default
Time point or interval preset time period just trigger the risk assessment request for generating and being directed toward one or more default nodes;Systemic presupposition
Condition can also to update triggering, i.e., for the arbitrary node in preset related network, the venture influence for detecting node because
When son changes, the risk assessment request of the variation node is directed toward in triggering, wherein the venture influence factor, which refers to, influences node risk
The parameter of value, the venture influence factor can be transmitted to the node for other nodes in node itself risk, preset related network
Input at least one of risk and anti-risk coefficient (including first anti-risk coefficient);Systemic presupposition condition can also be wind
The funcall of other modules except danger assessment request generation module, i.e. other modules can call risk assessment request to generate mould
The module interface of block generates risk assessment request, such as: risk Reports module need to obtain report institute when generating risk report
The value-at-risk of each risk sources comprising enterprise, such as the conduction risk based on incidence relation, itself risk, policy risk, this
When, risk Reports module calls directly risk assessment request generation module, and triggering generates risk assessment request, to obtain report institute
Risk comprising enterprise.
Destination node can be specified temporarily, if the selected one or more node requests of user carry out risk assessment, then can be touched
It sends out risk assessment equipment and generates the risk assessment request for being directed toward user's selected node;Destination node can also be preset, can be pre-
One or more nodes are first set as destination node, when requesting risk assessment, generates and is directed toward preset destination node
Risk assessment request.
Itself risk refers to enterprise because the operation of itself, investment behavior, behavior of lending and juristic act are (as carried on a shoulder pole
Protect, mortgage) etc. generations risk, or the personal property because of itself dominates the wind of the generations such as behavior, juristic act, behavior of lending
Danger.Itself risk in the present embodiment, is obtained by itself risk evaluation model, for personal node, by the letter of personal node
With information (as individual is overdue, bull is provided a loan, break one's promise, public security is bad, fraud), legal information (such as legal dispute), assets
Information (such as transaction journal) personal information enters ginseng as itself risk evaluation model, exports personal node itself risk, specifically
Ground obtains itself risk assessment after itself risk evaluation model classifies and receives personal information for itself risk evaluation model
The corresponding weight of the preset various types of personal information of model, and it is based on various types of personal information and its corresponding weight calculation
Itself risk of personal node.For enterprise's node, by Industry risk, company's administration of justice risk, abnormal, administrative penalty, equity are managed
The information such as pledge, chattel mortgage, tax arrear bulletin, judicial auction, unfavorable ratings enter ginseng, output enterprise as itself risk evaluation model
Itself risk of industry node specifically for itself risk evaluation model, classifies in itself risk evaluation model and receives company information
Afterwards, obtain the corresponding weight of the preset various types of company information of itself risk evaluation model, and based on various types of company information and
Its corresponding weight calculation enterprise node itself risk.
Anti-risk coefficient can pass through the ability to ward off risks for characterizing the ability to ward off risks of node or absorbing the ability of risk
Model obtains, and is the personal assets Distribution Indexes weights such as personal income level, Assets, based on each for personal node
People's assets index and its corresponding weight calculate the anti-risk coefficient for obtaining personal node;It is enterprise year for enterprise's node
Report, enterprise's inlet and outlet, registered capital, number, business circumstance, law, enterprise qualification, industry, scale, state-run assets background, production warp
It seeks the business indicators such as situation and distributes weight, be based on each business indicators and its corresponding weight, calculate the wind resistance for obtaining enterprise's node
Dangerous coefficient.First anti-risk coefficient refers in particular to the anti-risk coefficient of destination node.
Risk assessment equipment starts risk assessment procedures after detecting risk assessment request, is requested based on risk assessment
Itself of destination node can be directly acquired after determining destination node by obtaining Object node of the destination node as risk assessment
Risk and the first anti-risk coefficient, for calculating the value-at-risk of destination node.
Step S20 obtains all first nodes being directly linked with the destination node from preset related network, and obtains
Obtain the first input risk that all first nodes are transmitted to the destination node;
It is illustrated in figure 5 part of nodes exemplary diagram in related network, for any node, the risk of the node is all come
Itself risk (RISKself) and input risk+other input risks in itself and extraneous two levels, i.e. Fig. 5
(RISKtarget1+...+RISKtargetn), in addition, assessment to node risk, it is also contemplated that the ability to ward off risks of node
(being characterized with the anti-risk coefficient a in Fig. 3), the assessment for node risk (i.e. node exports risk RISKout), has:
RISKout=f ((RISKtarget1+...+RISKtargetn+RISKself), a)
Therefore, after itself risk and the first anti-risk coefficient for obtaining destination node, destination node is further obtained
Extraneous risk, the extraneous risk of destination node are transmitted to the input wind of destination node for other nodes in preset related network
Danger, as the RISKtarget1+...+RISKtargetn in Fig. 5 (only shows two first nodes, may actually have more in figure
A first node).Resist it should be noted that the value-at-risk for calculating destination node needs to obtain itself risk of destination node and first
Other nodes in risk factor and preset related network are transmitted to the input risk of destination node, wherein itself risk,
First anti-risk coefficient and the acquisition for inputting risk have no strict sequence, it can are obtaining itself risk and first
Input risk is obtained after anti-risk coefficient again, itself risk, the first anti-risk coefficient and input risk can also be obtained simultaneously,
The acquisition sequence of itself risk, the first anti-risk coefficient and input risk is not limited herein.Step S10, in step S20
Description needs are limited only to itself risk, the first anti-risk coefficient and input risk acquisition sequence.
Preset related network refers to the meshed network pre-established according to the incidence relation between node, in preset related network extremely
The strength of association between nodal information and node is contained less, wherein nodal information includes but is not limited to nodename, node
One of type, node assets, node operation information (enterprise), node equity information, personal information etc. are a variety of.
In preset related network, for each node, risk conduction can establish centered on the node
Network, having direct correlation relationship with the node is network first tier, in the present embodiment, will have direct correlation to close with destination node
The node of system is known as first node, and referring to the example in Fig. 6, label 1 is first node;There is indirect association relationship with destination node
Node, transmit risk often through first node, therefore, be transmitted to the input risk of destination node calculating other nodes
When, the input risk that first node is transmitted to destination node only need to be calculated, the input risk of acquisition is in preset related network
Every other node is transmitted to the input risk of the node.
In preset related network, destination node may have one or more first nodes being directly linked therefore counting
When calculating other nodes and being transmitted to the input risk of destination node, all first nodes should be obtained, and obtains all first nodes and passes
Lead the first input risk of destination node.
Step S30 makees itself risk of the destination node, the first input risk and the first anti-risk coefficient
Preset association conduction model is inputted to enter ginseng, obtains the output risk of the destination node.
As shown in figure 5, the assessment for node risk (i.e. node exports risk RISKout), has:
RISKout=f ((RISKtarget1+...+RISKtargetn+RISKself), a)
That is, the output risk of destination node need to be based on itself risk of destination node and the first input risk and the first wind resistance
The function of dangerous coefficient is calculated.First anti-risk coefficient characterization node absorbs the ability of risk, therefore, the first anti-risk coefficient
Bigger, the ability for absorbing risk is stronger, and node output risk is smaller, and therefore, the first anti-risk coefficient is with node output risk
Negative correlativing relation.In one embodiment, acquisition node data pre-process collected node data, obtain comprising section
Itself risk, the first input risk, anti-risk coefficient and the sample data and test sample that export risk are put, sample number is based on
Function formula is obtained according to calculating.
The target section that the present embodiment is directed toward by after detecting risk assessment request, obtaining the risk assessment request
Point is obtained in itself risk, the first anti-risk coefficient and the preset related network of destination node and is directly linked with destination node
All first nodes be transmitted to destination node first input risk, according to itself risk of destination node, first input wind
The output risk of danger and the first anti-risk coefficient assessment destination node is, it can be achieved that risk is conducted in the association based on incidence relation
Quantization realizes that the quantization of risk is conducted in the association between enterprise or node with incidence relation in preset related network.
It include: the practical risk for obtaining the destination node after the step S30, by the practical wind of the destination node
Danger inputs the association conduction model, so that association conduction model according to the practical risk of the destination node and exports risk
Discrepancy adjustment optimal weights matrix parameter.
Being associated with conduction model can be according to true external adjustment model parameter, by training input parameter (such as occurrence risk
Certain enterprise and its generation risk) input association conduction model obtains prediction result by the operation and processing inside model
(such as by the enterprise of the business impact of occurrence risk and its it is impacted after risk), prediction result and actual result are compared
Compared with the difference between prediction result and actual result being indicated with loss function, by minimizing loss function for weight square
Battle array parameter is adjusted to optimal value, to generate the association conduction model of the matrix parameter containing optimal weights.Wherein, loss function is existing
There is technology, does not repeat.
Optionally, may also include that before step S10
Step S11, when the venture influence factor for detecting arbitrary node in preset related network changes, triggering life
At the risk assessment request for being directed toward the arbitrary node, the venture influence factor includes the arbitrary node itself risk, pre-
Set input risk and at least one of anti-risk coefficient that other nodes in related network are transmitted to the arbitrary node.
The venture influence factor refers to the parameter for influencing node value-at-risk, and in the present embodiment, the venture influence factor can be section
Point itself risk, other nodes are transmitted in the input risk and anti-risk coefficient of the arbitrary node in preset related network
At least one.
In one embodiment, the venture influence factor includes node itself risk, then for arbitrary node in preset related network,
When itself risk of the arbitrary node changes, triggering generates risk assessment request.I.e.: in real time/preset pass of timing acquisition
Itself risk of networking Luo Zhong enterprise or individual;Itself risk of any enterprise or individual hair in detecting preset related network
When changing, the risk assessment request for being directed toward any enterprise or individual is generated, and executes the risk assessment step of step S10-S30
Suddenly;
In one embodiment, the venture influence factor includes anti-risk coefficient, then for arbitrary node in preset related network,
When the anti-risk coefficient of the arbitrary node changes, triggering generates risk assessment request.I.e.: in real time/preset pass of timing acquisition
The anti-risk coefficient of networking Luo Zhong enterprise or individual;Any enterprise or the anti-risk system of individual in detecting preset related network
When number changes, the risk assessment request for being directed toward any enterprise or individual is generated, and the risk for executing step S10-S30 is commented
Estimate step;
In one embodiment, the venture influence factor include in preset related network other nodes be transmitted to the arbitrary node
Input risk, then for arbitrary node in preset related network, when the input risk of the arbitrary node changes, triggering life
It is requested at risk assessment.The input risk of enterprise or individual in i.e.: in real time/preset related network of timing acquisition;It is pre- detecting
When setting the input risk of any enterprise in related network or individual and changing, the risk for being directed toward any enterprise or individual is generated
Assessment request, and execute the risk assessment step of step S10-S30.
For convenient for illustrating the risk conduction in preset related network, by itself risk and/or anti-risk coefficient and/or
The input changed arbitrary node of risk is named as first kind node, when first kind node changes, is based on RISKout
=f ((RISKtarget1+...+RISKtargetn+RISKself), a), it is known that the output risk pole of the first kind node can
It can also change, the node being directly linked with first kind node is named as the second class node, because of the output of first kind node
Risk and the input risk of the second class node have larger association, so the input risk of the second class node also most probably becomes
Change, then for inputting the changed second class node of risk, can also trigger and generate risk assessment request, to the second class node weight
It is new to carry out risk assessment, because in preset related network, the second class node also has the third class node of direct correlation, the second class section
The variation of point output risk is likely to cause the variation of third class node input risk, then the third class node changed can also trigger
Risk assessment request is generated, third class node also has the 4th class node ... of direct correlation so to recycle, and realizes that risk exists
Conduction in preset related network.(the first, second, third and fourth class node therein is only to explain convenient, no particular meaning)
Optionally, it is configured in the risk evaluating system in risk assessment equipment, in the one or more selected nodes of change
The venture influence factor after, the associated nodes of system automatic identification and selected node reappraise the associated nodes of selected node
Risk, and export reappraise after select node associated nodes and its risk, thus realize take up an official post in preset related network
When node occurrence risk of anticipating, which is quantified to other node bring venture entrepreneurs on preset related network.
The embodiment of the present invention is when the venture influence factor of arbitrary node changes in detecting preset related network, touching
The risk assessment request for being directed toward the arbitrary node occurs into, it can be achieved that when the value-at-risk of arbitrary node may change,
Risk assessment is all carried out to it again, and then realizes the real time monitoring to preset related network interior joint risk, meanwhile, because in advance
The output risk for setting arbitrary node in related network may be the input risk of other nodes, so any in preset related network
The variation of the output risk of node all may cause the change for other node venture influence factors being directly linked with the arbitrary node
Change, and then trigger risk assessment request again, so circulation triggering is, it can be achieved that wind occurs for arbitrary node on preset related network
When dangerous, quantify being transmitted to the venture entrepreneur of other associated nodes of the arbitrary node.
Further, in methods of risk assessment second embodiment of the present invention, all first segments of acquisition described in step S20
Point be transmitted to the destination node first input risk the step of include:
Step S21, obtains the output risk of each first node, and obtains each first node respectively to the mesh
Mark the risk coefficient of conductivity of node;
Risk conduction between associated nodes, it may be possible to which absolutely risk is conducted, it is also possible to be had in conductive process
Risk enhancing weakens.Referring to Fig. 7, if the risk conduction of node B to node A has RISKout1- absolutely to conduct
1=RISKIn2-1, i.e. the output risk of node B are equal to the input risk that node B is transmitted to node A;If node B is to node A's
Risk conduction is enhancing or weakens, then may have RISKout1-1=n*RISKIn2-1 (n ≠ 1 and n > 0).
In the present embodiment, it is transmitted to the output risk of risk coefficient of conductivity characterization arbitrary node and is directly linked node
When, which exports weakening/enhancing degree of risk, i.e., understands that how many risk is transmitted to this for calculating the arbitrary node
The corresponding direct correlation node of arbitrary node.Wherein, the risk conduction between node has directionality, it may be assumed that node A to node B
Risk coefficient of conductivity RA- > B it is identical as risk coefficient of conductivity RB- > A possibility of node B to node A, it is also possible to it is not identical.
In the present embodiment, for the risk for assessing destination node, other nodes only calculated in preset related network are transmitted to
The risk of destination node, i.e. calculating first node are transmitted to the first input risk of destination node, thus calculate the first input wind
The risk coefficient of conductivity used in danger is the risk coefficient of conductivity (R first node -> target section of the first node to the destination node
Point), and then when calculating the output risk of first node, and calculate and other nodes of first node direct correlation to first
The risk coefficient of conductivity of node, if by second node is named as with the node that first node is directly linked, it is straight with second node
Connect associated node and be named as third node, with third node be directly linked node be named as fourth node ... .. and the n-th section
The node that point is directly linked is named as the (n+1)th node, then in the output wind for calculating separately second and third, four ... n, n+1 nodes
When dangerous, the risk coefficient of conductivity being related to is respectively R third node -> second node, R fourth node -> third node ... R
N+1 node -> the n-th node.
Further, the step of output risk of each first node of acquisition described in step S21 includes:
Step S211 obtains itself risk and the second anti-risk coefficient of each first node, and obtain respectively with it is each
All second nodes that the first node is directly linked are transmitted to the second input risk of each first node;
Step S212, according to itself risk of each first node, the second input risk and the second anti-risk system
Number calculates the output risk for obtaining each first node.
To obtain the first input risk that all first nodes are transmitted to destination node, each first node conduction need to be calculated
To the input risk of destination node, the output risk of each first node need to be further calculated.Second anti-risk coefficient refers in particular to
The anti-risk coefficient of one node.In preset related network, the output Risk Calculation step of each node is almost the same, in the present embodiment
First node calculating step (step S211, step S212) and destination node calculating step (step S10, S20, S30)
Corresponding consistent, relevant explanation explanation is also corresponding consistent, does not repeat herein.
Further, the risk for obtaining each first node to the destination node described in step S21 respectively is conducted
The step of coefficient includes:
Step S213 obtains node type, strength of association and the association of the destination node and each first node
Type;
The risk coefficient of conductivity between two nodes is related with the multiple parameters of two nodes, and parameter includes: two nodes
Strength of association (Relation weight), association type (relation type), two nodes node type (node
Type), there is R=f (relation type, Relation weight, node type).
Node type includes the classification of personal node and enterprise's node, the relationship type between association type finger joint point, such as
Parent and child relationship, pair bond, siblings' relationship, friend classmate and fellow-villager's relationship between personal node etc., such as
Guarantee relationship, parent-subsidiary relationship between enterprise's node etc..
Strength of association can according to the share-holding amount of money or shareholding ratio, personal node the tenure type of enterprise's node, tenure and
Invest many-sided assessment such as duration and affiliated party's quantity.
Node type, association type and strength of association all can be obtained directly from the nodal information in preset related network.
Step S214, according to the node type, strength of association and association type determine respectively the destination node and
The risk coefficient of conductivity between each first node.
Based on the sample of preset node type, strength of association, association type and value-at-risk, calculated by iterative algorithm
The optimal model parameters of coefficient of conductivity model out train the coefficient of conductivity model containing optimal model parameters.Respectively by target section
The node type of point and each first node, strength of association and association type input coefficient of conductivity model as ginseng is entered, point
It Huo get not the risk coefficient of conductivity between destination node and each first node.
Step S22 is calculated and is obtained all the according to the output risk of each first node and the risk coefficient of conductivity
One node is transmitted to the first input risk of the destination node.
According to the output risk of each first node and the corresponding risk coefficient of conductivity, calculates each first node of acquisition and be transmitted to
The input risk of destination node, it is that all first nodes are transmitted to that each first node, which is transmitted to the sum of input risk of destination node,
First input risk of the destination node;Optionally, the output risk of each first node and the corresponding risk coefficient of conductivity it
Product is the input risk that each first node is transmitted to destination node.
The present embodiment obtains each first node to institute by obtaining the output risk of each first node respectively
State the risk coefficient of conductivity of destination node;According to the output risk of each first node and the risk coefficient of conductivity, calculate
The first input risk that all first nodes are transmitted to the destination node is obtained, so as to the meter of succeeding target node output risk
It calculates, wherein the calculating factor of the risk coefficient of conductivity considers node type, strength of association and association type, can get more quasi-
The true risk coefficient of conductivity, and then obtain the output risk of more accurate first input risk and destination node.
In addition, the present invention also provides a kind of risk evaluating systems corresponding with each step of above-mentioned methods of risk assessment.
It is the functional block diagram of risk evaluating system first embodiment of the present invention referring to Fig. 8, Fig. 8.
In the present embodiment, risk evaluating system of the present invention includes:
First obtains module 10, for after detecting risk assessment request, obtaining the risk assessment request direction
Itself risk and the first anti-risk coefficient of destination node and the destination node;
Second obtains module 20, for obtaining be directly linked with the destination node all the from preset related network
One node, and obtain the first input risk that all first nodes are transmitted to the destination node;
Computing module 30, for itself risk of the destination node, first to be inputted risk and first wind resistance
Dangerous coefficient inputs preset association conduction model as ginseng is entered, and obtains the output risk of the destination node.
Further, the second acquisition module 20 is also used to obtain the output risk of each first node, and respectively
Obtain each first node to the destination node the risk coefficient of conductivity;According to the output risk of each first node and
The risk coefficient of conductivity calculates and obtains the first input risk that all first nodes are transmitted to the destination node.
Further, the second acquisition module 20 is also used to obtain itself risk of each first node and second and resists
Risk factor obtains the second node being directly linked with each first node respectively, and obtains the second node respectively and pass
Lead the second input risk of corresponding first node;According to itself risk of each first node, the second input risk and institute
It states the second anti-risk coefficient and calculates the output risk for obtaining each first node.
Further, the second acquisition module 20 is also used to obtain the section of the destination node and each first node
Vertex type, strength of association and association type;Described in being determined respectively according to the node type, strength of association and association type
The risk coefficient of conductivity between destination node and each first node.
Further, the risk evaluating system further include:
Trigger module is detected, for changing when the venture influence factor for detecting arbitrary node in preset related network
When, triggering generates the risk assessment request for being directed toward the arbitrary node, the venture influence factor include the arbitrary node from
Other nodes are transmitted in the input risk and anti-risk coefficient of the arbitrary node extremely in body risk, preset related network
It is one few.
Further, the risk evaluating system further include:
Model training module trains association conduction model using the training sample set for obtaining training sample set,
Optimal weights matrix parameter is obtained, the preset association conduction model of the matrix parameter containing optimal weights is trained.
Further, the risk evaluating system further include:
Model adjusts module, for obtaining the practical risk of the destination node, by the practical risk of the destination node
The association conduction model is inputted, so that association conduction model is according to the difference of the practical risk of the destination node and output risk
Different adjustment optimal weights matrix parameter.
The present invention also proposes a kind of storage medium, is stored thereon with computer program.The storage medium can be Fig. 1's
Memory 201 in risk assessment equipment is also possible to such as ROM (Read-Only Memory, read-only memory)/RAM
At least one of (Random Access Memory, random access memory), magnetic disk, CD, the storage medium includes
Some instructions use so that one with processor equipment equipment (can be mobile phone, computer, server, the network equipment or
Risk assessment equipment in the embodiment of the present invention etc.) execute method described in each embodiment of the present invention.
It should be noted that, in this document, the terms "include", "comprise" or its any other variant are intended to non-row
His property includes, so that the process, method, article or the server-side that include a series of elements not only include those elements,
It but also including other elements that are not explicitly listed, or further include for this process, method, article or server-side institute
Intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that wrapping
Include in process, method, article or the server-side of the element that there is also other identical elements.
The serial number of the above embodiments of the invention is only for description, does not represent the advantages or disadvantages of the embodiments.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment side
Method can be realized by means of software and necessary general hardware platform, naturally it is also possible to by hardware, but in many cases
The former is more preferably embodiment.
The above is only a preferred embodiment of the present invention, is not intended to limit the scope of the invention, all to utilize this hair
Equivalent structure or equivalent flow shift made by bright specification and accompanying drawing content is applied directly or indirectly in other relevant skills
Art field, is included within the scope of the present invention.
Claims (10)
1. a kind of methods of risk assessment, which is characterized in that the methods of risk assessment the following steps are included:
After detecting risk assessment request, destination node that risk assessment request is directed toward and the destination node are obtained
Itself risk and the first anti-risk coefficient;
All first nodes being directly linked with the destination node are obtained from preset related network, and obtain all first segments
Point is transmitted to the first input risk of the destination node;
Itself risk of the destination node, the first input risk and the first anti-risk coefficient is pre- as ginseng input is entered
The association conduction model set obtains the output risk of the destination node.
2. methods of risk assessment as described in claim 1, which is characterized in that all first nodes of acquisition are transmitted to described
Destination node first input risk the step of include:
Obtain the output risk of each first node, and obtain respectively each first node to the destination node risk
The coefficient of conductivity;
According to the output risk of each first node and the risk coefficient of conductivity, calculates all first nodes of acquisition and be transmitted to
First input risk of the destination node.
3. methods of risk assessment as claimed in claim 2, which is characterized in that the output wind for obtaining each first node
Danger step include:
Itself risk and the second anti-risk coefficient of each first node are obtained, is obtained respectively direct with each first node
Associated second node, and the second input risk that the second node is transmitted to corresponding first node is obtained respectively;
It calculates according to itself risk of each first node, the second input risk and the second anti-risk coefficient and obtains each institute
State the output risk of first node.
4. methods of risk assessment as claimed in claim 2, which is characterized in that described to obtain each first node respectively to institute
The step of stating the risk coefficient of conductivity of destination node include:
Obtain node type, strength of association and the association type of the destination node and each first node;
The destination node and each first node are determined respectively according to the node type, strength of association and association type
Between the risk coefficient of conductivity.
5. methods of risk assessment as described in claim 1, which is characterized in that it is described after detecting risk assessment request, it obtains
Obtain itself risk of destination node and the destination node that the risk assessment request is directed toward and the step of the first anti-risk coefficient
Include: before rapid
When the venture influence factor for detecting arbitrary node in preset related network changes, triggering, which generates, is directed toward described appoint
It anticipates the risk assessment request of node, the venture influence factor includes the arbitrary node itself risk, in preset related network
Other nodes are transmitted to the input risk and at least one of anti-risk coefficient of the arbitrary node.
6. methods of risk assessment as described in claim 1, which is characterized in that it is described after detecting risk assessment request, it obtains
Obtain itself risk of destination node and the destination node that the risk assessment request is directed toward and the step of the first anti-risk coefficient
Include: before rapid
Training sample set is obtained, association conduction model is trained using the training sample set, obtains optimal weights matrix parameter,
Train the preset association conduction model of the matrix parameter containing optimal weights.
7. methods of risk assessment as claimed in claim 6, which is characterized in that it is described by itself risk of the destination node,
First input risk and the first anti-risk coefficient obtain the target as the preset association conduction model of ginseng input is entered
Include: after the step of output risk of node
The practical risk of the destination node is inputted the association conduction model by the practical risk for obtaining the destination node,
So that association conduction model is joined according to the discrepancy adjustment optimal weights matrix of the practical risk of the destination node and output risk
Number.
8. a kind of risk evaluating system, which is characterized in that the risk evaluating system includes:
First obtains module, the target section being directed toward for after detecting risk assessment request, obtaining the risk assessment request
Itself risk and the first anti-risk coefficient of point and the destination node;
Second obtains module, for obtaining all first segments being directly linked with the destination node from preset related network
Point, and obtain the first input risk that all first nodes are transmitted to the destination node;
Computing module, for itself risk of the destination node, first to be inputted risk and the first anti-risk coefficient
Preset association conduction model is inputted as ginseng is entered, obtains the output risk of the destination node.
9. a kind of risk assessment equipment, which is characterized in that the risk assessment equipment includes processor, memory and storage
On the memory and the risk assessment procedures that can be executed by the processor, wherein the risk assessment procedures are by the place
When managing device and executing, the step of realizing methods of risk assessment as described in any one of claims 1 to 7.
10. a kind of storage medium, which is characterized in that risk assessment procedures are stored on the storage medium, wherein the risk
When appraisal procedure is executed by processor, the step of realizing methods of risk assessment as described in any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811529145.5A CN109740865A (en) | 2018-12-13 | 2018-12-13 | Methods of risk assessment, system, equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811529145.5A CN109740865A (en) | 2018-12-13 | 2018-12-13 | Methods of risk assessment, system, equipment and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109740865A true CN109740865A (en) | 2019-05-10 |
Family
ID=66359378
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811529145.5A Pending CN109740865A (en) | 2018-12-13 | 2018-12-13 | Methods of risk assessment, system, equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109740865A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110738388A (en) * | 2019-09-02 | 2020-01-31 | 深圳壹账通智能科技有限公司 | Method, device, equipment and storage medium for risk conduction of associated map evaluation |
CN110782115A (en) * | 2019-08-19 | 2020-02-11 | 腾讯科技(深圳)有限公司 | Method and device for processing risks of Internet of vehicles system, electronic equipment and storage medium |
CN111353728A (en) * | 2020-05-06 | 2020-06-30 | 支付宝(杭州)信息技术有限公司 | Risk analysis method and system |
CN111915206A (en) * | 2020-08-11 | 2020-11-10 | 成都市食品药品检验研究院 | Method for recognizing food risk conduction |
CN112511567A (en) * | 2021-02-05 | 2021-03-16 | 浙江地芯引力科技有限公司 | Method and device for managing secret communication priority of intelligent security chip |
CN112734270A (en) * | 2021-01-19 | 2021-04-30 | 中国科学院地理科学与资源研究所 | Measuring method, system and data platform for energy risk conduction |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107292509A (en) * | 2017-06-16 | 2017-10-24 | 兴业数字金融服务(上海)股份有限公司 | A kind of enterprise's credit risk early-warning monitoring method |
CN107862205A (en) * | 2017-11-01 | 2018-03-30 | 龚土婷 | One kind assesses accurate information security risk evaluation system |
CN107909274A (en) * | 2017-11-17 | 2018-04-13 | 平安科技(深圳)有限公司 | Enterprise investment methods of risk assessment, device and storage medium |
CN108090709A (en) * | 2018-02-09 | 2018-05-29 | 重庆誉存大数据科技有限公司 | A kind of enterprise evaluation method and system based on risk conduction model |
-
2018
- 2018-12-13 CN CN201811529145.5A patent/CN109740865A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107292509A (en) * | 2017-06-16 | 2017-10-24 | 兴业数字金融服务(上海)股份有限公司 | A kind of enterprise's credit risk early-warning monitoring method |
CN107862205A (en) * | 2017-11-01 | 2018-03-30 | 龚土婷 | One kind assesses accurate information security risk evaluation system |
CN107909274A (en) * | 2017-11-17 | 2018-04-13 | 平安科技(深圳)有限公司 | Enterprise investment methods of risk assessment, device and storage medium |
CN108090709A (en) * | 2018-02-09 | 2018-05-29 | 重庆誉存大数据科技有限公司 | A kind of enterprise evaluation method and system based on risk conduction model |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110782115A (en) * | 2019-08-19 | 2020-02-11 | 腾讯科技(深圳)有限公司 | Method and device for processing risks of Internet of vehicles system, electronic equipment and storage medium |
CN110782115B (en) * | 2019-08-19 | 2024-04-26 | 腾讯科技(深圳)有限公司 | Method and device for processing risk of Internet of vehicles system, electronic equipment and storage medium |
CN110738388A (en) * | 2019-09-02 | 2020-01-31 | 深圳壹账通智能科技有限公司 | Method, device, equipment and storage medium for risk conduction of associated map evaluation |
CN110738388B (en) * | 2019-09-02 | 2023-09-12 | 深圳壹账通智能科技有限公司 | Method, device, equipment and storage medium for evaluating risk conduction through association map |
CN111353728A (en) * | 2020-05-06 | 2020-06-30 | 支付宝(杭州)信息技术有限公司 | Risk analysis method and system |
CN111915206A (en) * | 2020-08-11 | 2020-11-10 | 成都市食品药品检验研究院 | Method for recognizing food risk conduction |
CN111915206B (en) * | 2020-08-11 | 2024-02-27 | 成都市食品药品检验研究院 | Method for identifying food risk conduction |
CN112734270A (en) * | 2021-01-19 | 2021-04-30 | 中国科学院地理科学与资源研究所 | Measuring method, system and data platform for energy risk conduction |
CN112734270B (en) * | 2021-01-19 | 2024-01-23 | 中国科学院地理科学与资源研究所 | Energy risk conduction measurement method, system and data platform |
CN112511567A (en) * | 2021-02-05 | 2021-03-16 | 浙江地芯引力科技有限公司 | Method and device for managing secret communication priority of intelligent security chip |
CN112511567B (en) * | 2021-02-05 | 2021-05-11 | 浙江地芯引力科技有限公司 | Method and device for managing secret communication priority of intelligent security chip |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109740865A (en) | Methods of risk assessment, system, equipment and storage medium | |
CN109492945A (en) | Business risk identifies monitoring method, device, equipment and storage medium | |
CN106529729A (en) | Method and system for forecasting default of credit card user based on BP_Adaboost model | |
Bracking | How do investors value environmental harm/care? Private equity funds, development finance institutions and the partial financialization of nature‐based industries | |
CN109410036A (en) | A kind of fraud detection model training method and device and fraud detection method and device | |
CN109977151A (en) | A kind of data analysing method and system | |
CN108399509A (en) | Determine the method and device of the risk probability of service request event | |
Malito | Measuring corruption indicators and indices | |
CN110417721A (en) | Safety risk estimating method, device, equipment and computer readable storage medium | |
CN108416663A (en) | The method and device of the financial default risk of assessment | |
CN107203939A (en) | Determine method and device, the computer equipment of consumer's risk grade | |
CN106296391A (en) | A kind of assessment exceeds the time limit the method and apparatus of probability | |
CN107038511A (en) | A kind of method and device for determining risk assessment parameter | |
CN106651573A (en) | Business data processing method and apparatus | |
CN108389120A (en) | Method, system and device for automatically rating internet credit assets | |
CN112767136A (en) | Credit anti-fraud identification method, credit anti-fraud identification device, credit anti-fraud identification equipment and credit anti-fraud identification medium based on big data | |
CN107481135A (en) | A kind of personal credit evaluation method and system based on BP neural network | |
Huang et al. | Improve reputation evaluation of crowdsourcing participants using multidimensional index and machine learning techniques | |
CN109636565A (en) | Processing method, device, equipment and the computer readable storage medium of risk data | |
Arora et al. | PREDICTION & WARNING: a method to improve student's performance | |
CN112132441A (en) | Risk propagation information evaluation method, risk propagation information evaluation system, storage medium and computer equipment | |
CN110473039A (en) | Assess the method and device of the credit risk of user | |
Li et al. | Application of XGBoost in P2P default prediction | |
CN109146667A (en) | A kind of construction method of the external interface integrated application model based on quantitative statistics | |
Kim et al. | Inductive modeling of expert decision making in loan evaluation: a decision strategy perspective |
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 |