CN112561320A - Training method of mechanism risk prediction model, mechanism risk prediction method and device - Google Patents
Training method of mechanism risk prediction model, mechanism risk prediction method and device Download PDFInfo
- Publication number
- CN112561320A CN112561320A CN202011475410.3A CN202011475410A CN112561320A CN 112561320 A CN112561320 A CN 112561320A CN 202011475410 A CN202011475410 A CN 202011475410A CN 112561320 A CN112561320 A CN 112561320A
- Authority
- CN
- China
- Prior art keywords
- risk
- prediction model
- sample
- target
- model
- 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
- 230000007246 mechanism Effects 0.000 title claims abstract description 281
- 238000012549 training Methods 0.000 title claims abstract description 81
- 238000013058 risk prediction model Methods 0.000 title claims abstract description 72
- 238000000034 method Methods 0.000 title claims abstract description 70
- 230000008520 organization Effects 0.000 claims description 41
- 238000003860 storage Methods 0.000 claims description 13
- 238000012545 processing Methods 0.000 claims description 12
- 238000012706 support-vector machine Methods 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 11
- 238000007637 random forest analysis Methods 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 10
- 238000012502 risk assessment Methods 0.000 claims description 7
- 238000004140 cleaning Methods 0.000 claims description 6
- 238000003062 neural network model Methods 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 5
- 230000017105 transposition Effects 0.000 claims description 3
- 238000004422 calculation algorithm Methods 0.000 description 25
- 238000010586 diagram Methods 0.000 description 17
- 239000010410 layer Substances 0.000 description 14
- 230000006870 function Effects 0.000 description 12
- 210000002569 neuron Anatomy 0.000 description 7
- 238000013528 artificial neural network Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 6
- 230000002159 abnormal effect Effects 0.000 description 4
- 238000012550 audit Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000010801 machine learning Methods 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000003066 decision tree Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000000835 fiber Substances 0.000 description 2
- 238000004900 laundering Methods 0.000 description 2
- 238000012417 linear regression Methods 0.000 description 2
- 238000007477 logistic regression Methods 0.000 description 2
- 239000013307 optical fiber Substances 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 238000012952 Resampling Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 239000002356 single layer Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a training method and a training device for a mechanism risk prediction model, and a mechanism risk prediction method and a mechanism risk prediction device, and relates to the technical field of computers. One embodiment of the method comprises: acquiring a mechanism sample set, wherein the mechanism sample set comprises a plurality of mechanism samples, and each mechanism sample comprises a plurality of mechanism indexes and risk labels; taking a plurality of mechanism indexes in the mechanism sample as input, taking a risk label in the mechanism sample as output, and training a plurality of initial prediction models to obtain a plurality of target prediction models; determining the weight corresponding to each target prediction model; and determining a mechanism risk prediction model according to the target prediction models and the weights corresponding to the target prediction models. The mechanism risk prediction result determined by the mechanism risk prediction model of the embodiment has higher accuracy.
Description
Technical Field
The invention relates to the technical field of computers, in particular to a training method of a mechanism risk prediction model, a mechanism risk prediction method and a mechanism risk prediction device.
Background
In the daily operation of an organization, the production operation of the organization may suffer a great impact due to various influences such as an imperfect internal control, a rapid change in the external economic environment, and the like. This may not only damage the institution itself, but may also damage the stakeholders of the company, and in addition, credit institutions such as banks may suffer bad account losses due to the financial risk of not identifying the institution in a timely manner.
Therefore, it is necessary to predict the risk of an organization, such as identifying whether the organization is at risk of fraud, money laundering, and breach of contract. The traditional institution risk assessment is usually obtained by manually scoring and calculating various indexes, and the institution risk prediction result lacks certain accuracy.
Disclosure of Invention
In view of this, embodiments of the present invention provide a training method for an organization risk prediction model, an organization risk prediction method, and an organization risk prediction device, which can more accurately predict a risk that may exist in an organization.
In a first aspect, an embodiment of the present invention provides a method for training a mechanism risk prediction model, including:
acquiring a mechanism sample set, wherein the mechanism sample set comprises a plurality of mechanism samples, and each mechanism sample comprises a plurality of mechanism indexes and risk labels;
taking a plurality of mechanism indexes in the mechanism sample as input, taking a risk label in the mechanism sample as output, and training a plurality of initial prediction models to obtain a plurality of target prediction models;
determining the weight corresponding to each target prediction model;
and determining a mechanism risk prediction model according to the target prediction models and the weights corresponding to the target prediction models.
Optionally, the determining a weight corresponding to each of the target prediction models includes:
determining a performance parameter of each of the target prediction models;
and determining the weight corresponding to each target prediction model according to the performance parameters.
Optionally, the performance parameter comprises at least one of: a receiver operating characteristic curve, a confusion matrix, a model score, a model accuracy or a model accuracy.
Optionally, the determining, according to the performance parameter, a weight corresponding to each of the target prediction models includes:
and if the current performance parameter is smaller than the performance threshold, the weight corresponding to the current target prediction model is 0, and the current target prediction model corresponds to the current performance parameter.
Optionally, the plurality of initial predictive models comprises: a first prediction model, wherein the first prediction model adopts a neural network model;
the training of the multiple initial prediction models by taking the multiple mechanism indexes in the mechanism sample as input and the risk labels in the mechanism sample as output to obtain multiple target prediction models comprises the following steps:
and training a first initial prediction model by taking a plurality of mechanism indexes in the mechanism sample as input and taking a risk label in the mechanism sample as output to obtain a first target prediction model.
Optionally, the plurality of initial predictive models comprises: a second prediction model, which adopts a support vector machine model;
the training of the multiple initial prediction models by taking the multiple mechanism indexes in the mechanism sample as input and the risk labels in the mechanism sample as output to obtain multiple target prediction models comprises the following steps:
and training a second initial prediction model by taking a plurality of mechanism indexes in the mechanism sample as input and taking a risk label in the mechanism sample as output to obtain a second target prediction model.
Optionally, the plurality of initial predictive models comprises: a third prediction model, wherein the third prediction model adopts a random forest model;
the training of the multiple initial prediction models by taking the multiple mechanism indexes in the mechanism sample as input and the risk labels in the mechanism sample as output to obtain multiple target prediction models comprises the following steps:
and training a third initial prediction model by taking a plurality of mechanism indexes in the mechanism sample as input and taking a risk label in the mechanism sample as output to obtain a third target prediction model.
Optionally, the acquiring a set of mechanism samples comprises:
acquiring financial data and non-financial data for a plurality of institutions;
for each of the plurality of mechanisms: determining a plurality of metrics for the institution based on the financial data and non-financial data for the institution; determining a risk label for the organization; and constructing the mechanism sample according to the plurality of indexes of the mechanism and the risk label of the mechanism.
Optionally, before said determining a plurality of indicators for the institution based on the financial data and non-financial data for the institution, further comprising:
performing a cleaning process on the data according to the financial data and the non-financial data, the cleaning process including at least one of: missing value processing, outlier processing, data transposition, and data summation.
In a second aspect, an embodiment of the present invention provides an organization risk prediction method, including:
acquiring a plurality of index values of a target mechanism;
and inputting the index values into a mechanism risk prediction model to obtain a risk prediction result of the target mechanism, wherein the mechanism risk prediction model is generated by a plurality of target prediction models and weights corresponding to the target prediction models.
Optionally, the weight corresponding to the target prediction model is determined according to a performance parameter of the target prediction model.
In a third aspect, an embodiment of the present invention provides a training apparatus for a mechanism risk prediction model, including:
the system comprises a sample acquisition module, a risk analysis module and a risk analysis module, wherein the sample acquisition module is used for acquiring a mechanism sample set, the mechanism sample set comprises a plurality of mechanism samples, and each mechanism sample comprises a plurality of mechanism indexes and risk labels;
the model training module is used for training a plurality of initial prediction models to obtain a plurality of target prediction models by taking a plurality of mechanism indexes in the mechanism sample as input and taking a risk label in the mechanism sample as output;
the weight determining module is used for determining the weight corresponding to each target prediction model;
and the model determining module is used for determining the mechanism risk prediction model according to the target prediction models and the weights corresponding to the target prediction models.
In a fourth aspect, an embodiment of the present invention provides an organization risk prediction apparatus, including:
the index acquisition module is used for acquiring a plurality of index values of the target mechanism;
and the risk prediction module is used for inputting the index values into a mechanism risk prediction model to obtain a risk prediction result of the target mechanism, and the mechanism risk prediction model is generated by a plurality of target prediction models and weights corresponding to the target prediction models.
In a fifth aspect, an embodiment of the present invention provides an electronic device, including:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement the method of any of the embodiments described above.
In a sixth aspect, the present invention provides a computer-readable medium, on which a computer program is stored, where the computer program is executed by a processor to implement the method of any one of the above embodiments.
One embodiment of the above invention has the following advantages or benefits: and determining the mechanism risk prediction model according to the plurality of target prediction models and the weights corresponding to the target prediction models. The determined mechanism risk prediction model is obtained by training a mechanism sample set with more samples. Compared with the traditional mode of manually scoring indexes to carry out risk prediction on the mechanism, the accuracy of the mechanism risk prediction result obtained by the mechanism risk prediction model is higher.
Furthermore, single models tend to be sensitive to the choice of parameters and functions, and thus difficult to implement for large-scale training samples. The mechanism risk prediction model of the embodiment of the invention is obtained by combining a plurality of target prediction models, and can be better implemented for large-scale training samples.
Further effects of the above-mentioned non-conventional alternatives will be described below in connection with the embodiments.
Drawings
The drawings are included to provide a better understanding of the invention and are not to be construed as unduly limiting the invention. Wherein:
FIG. 1 is an exemplary system architecture diagram in which embodiments of the present invention may be employed;
FIG. 2 is a schematic diagram of a process of a training method of an organization risk prediction model according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a process of another method for training an organization risk prediction model according to an embodiment of the invention;
FIG. 4 is a schematic diagram of a process flow of an organization risk prediction method provided by an embodiment of the invention;
FIG. 5 is a schematic structural diagram of a training apparatus for an organization risk prediction model according to an embodiment of the present invention;
FIG. 6 is a schematic structural diagram of a risk prediction device for an organization according to an embodiment of the present invention;
fig. 7 is a schematic block diagram of a computer system suitable for use in implementing a terminal device or server of an embodiment of the invention.
Detailed Description
Exemplary embodiments of the present invention are described below with reference to the accompanying drawings; various details of embodiments of the invention are included to assist in understanding, and they are to be considered as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the invention. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
Fig. 1 shows an exemplary system architecture 100 to which a training method of a facility risk prediction model, a facility risk prediction method, a training apparatus of a facility risk prediction model, and a facility risk prediction apparatus according to embodiments of the present invention can be applied.
As shown in fig. 1, the system architecture 100 may include terminal devices 101, 102, 103, a network 104, and a server 105. The network 104 serves as a medium for providing communication links between the terminal devices 101, 102, 103 and the server 105. Network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, to name a few.
Financial data and non-financial data of a plurality of institutions collected through various channels are stored in the terminal devices 101, 102, 103. The terminal devices 101, 102, 103 may be cell phones, notebooks, servers, tablets, laptop portable computers, etc.
The terminal devices 101, 102, 103 interact with a server 105 via a network 104 to receive or send messages or the like. The terminal apparatuses 101, 102, 103 transmit the stored video to the server 105 via the network 104.
The server 105 receives financial data and non-financial data of a plurality of institutions transmitted by the terminal devices 101, 102 and 103; generating an institution sample set based on financial data and non-financial data for a plurality of institutions; taking a plurality of mechanism indexes in the mechanism sample as input, taking a risk label in the mechanism sample as output, and training a plurality of initial prediction models to obtain a plurality of target prediction models; determining the weight corresponding to each target prediction model; and determining a mechanism risk prediction model according to the target prediction models and the weights corresponding to the target prediction models. The determined mechanism risk prediction model is used for determining a risk prediction result of the mechanism according to a plurality of indexes of the mechanism.
It should be noted that the training method of the institution risk prediction model or the institution risk prediction method provided in the embodiment of the present invention is generally executed by the server 105, and accordingly, the training device of the institution risk prediction model or the institution risk prediction device is generally disposed in the server 105.
It should be understood that the number of terminal devices, networks, and servers in fig. 1 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation.
Fig. 2 is a schematic diagram of a flow of a training method of a mechanism risk prediction model according to an embodiment of the present invention. The embodiment of the invention provides a training method of a mechanism risk prediction model, as shown in fig. 2, the training method comprises the following steps:
step 201: the method comprises the steps of obtaining a mechanism sample set, wherein the mechanism sample set comprises a plurality of mechanism samples, and each mechanism sample comprises a plurality of mechanism indexes and risk labels.
The organization index may be a statistical value of the organization's business under various statistical dimensions. The statistical dimensions may include: risk alert dimensions, profitability dimensions, liquidity risk dimensions, operations, and the like. The institutional indexes may include: statistical net profits, end-of-term owner equity totals, operating revenues, operating cash inflows, reputation, audit opinions, guild penalty information, etc. over a certain statistical period. Wherein, the reputation, the audit opinion, the certificate punishment information and the like are expressed by specific scores.
The risk label may be that the organization has risk or does not have risk, or may be a risk level of the organization, such as high risk, medium risk, low risk, no risk, etc., where each level corresponds to a risk coefficient range. The risk label may also be represented by a specific score, with different scores indicating different risks that the target institution may present.
In one embodiment of the present invention, a set of institutional samples is acquired comprising:
acquiring financial data and non-financial data for a plurality of institutions;
for each of a plurality of institutions: determining a plurality of metrics for the organization based on financial data and non-financial data for the organization; determining a risk label for the organization; and constructing a mechanism sample according to the multiple indexes of the mechanism and the risk label of the mechanism.
The calculation mode of each index can be preset in the system, and each index of the mechanism in a certain statistical period can be determined according to financial data and non-financial data of each mechanism in the statistical period.
The actual operation conditions of the mechanisms in the same statistical period with the indexes can be obtained, such as the positive and negative conditions of good operation conditions, bankruptcy, default, fraud or money laundering and the like of the mechanisms in the statistical period, and the risk labels of the mechanisms in the statistical period are determined.
In one embodiment of the invention, before determining the plurality of metrics for the organization based on the financial data and the non-financial data for the organization, further comprising:
performing a cleaning process on the financial data and the non-financial data, the cleaning process including at least one of: missing value processing, outlier processing, data transposition, and data summation.
The financial data and the non-financial data may have values lacking or abnormal during processing and extraction. The financial data and the non-financial data are cleaned, so that the subsequent index determining step can be conveniently and smoothly executed, and the obtained index data has higher accuracy.
Step 202: and training the plurality of initial prediction models to obtain a plurality of target prediction models by taking a plurality of mechanism indexes in the mechanism sample as input and taking a risk label in the mechanism sample as output.
The training methods adopted by different initial models are different. The training method adopted by different initial models comprises the following steps: any machine learning algorithm, or any deep learning algorithm. Such as linear regression algorithms, support vector machine algorithms, nearest neighbor/k-nearest neighbor algorithms, logistic regression algorithms, decision tree algorithms, k-means algorithms, random forest algorithms, naive bayes algorithms, back propagation, random gradient descent, and the like.
The output result of each target prediction model can be 0 or 1,0 representing the mechanism with risk, and 1 representing the mechanism without risk. The output of each target prediction model may also be a risk level of the organization, such as high risk, medium risk, low risk, no risk, etc. The output may also be represented by a specific score, with different scores indicating the probability of different risks that an organization may be at.
It should be noted that the expression form of the output result of each target prediction model and the corresponding meaning of the output result are the same, which facilitates the execution of the step of combining each target prediction model into the mechanism risk prediction model in the following step 204.
Step 203: and determining the corresponding weight of each target prediction model.
The weight corresponding to each target prediction model can be preset in the system, and the weight corresponding to each target model can also be determined according to the performance parameters of each target prediction model, such as model accuracy and the like. Optionally, the sum of the weights corresponding to all target prediction models is 1.
Step 204: and determining the mechanism risk prediction model according to the plurality of target prediction models and the weights corresponding to the target prediction models.
The mechanism risk prediction model is obtained from a plurality of target prediction models and a combination of weights corresponding to the target prediction models. The results from the institutional risk prediction model may be a weighted sum of the results from the respective target prediction models.
For example, the system has three target prediction models, and the weights of the three target prediction models are 0.7, 0.2 and 0.1 respectively. For a certain target mechanism, the output results of the three target prediction models are respectively as follows: 0. 0 and 1. Wherein, 0 represents the mechanism with risk, and 1 represents the mechanism without risk. The probability of risk for the target mechanism is 0.7+ 0.2-0.9 and the probability of risk for the target mechanism not being present is 0.1.
As another example, the system has two target prediction models in common, and the weights of the two target prediction models are 0.6 and 0.4, respectively. For a certain target mechanism, the output fraud risk scores of the three target prediction models are respectively as follows: 80 and 50. The target institution fraud risk score is 80 x 0.6+50 x 0.4-68.
In the embodiment of the invention, the mechanism risk prediction model is determined according to the target prediction models and the weights corresponding to the target prediction models. The determined mechanism risk prediction model is obtained by training a mechanism sample set with more samples, and the accuracy of the obtained mechanism risk prediction result can be higher by a traditional mode of manually scoring indexes to carry out risk prediction on the mechanism.
Furthermore, single models tend to be sensitive to the choice of parameters and functions, and thus difficult to implement for large-scale training samples. The mechanism risk prediction model of the embodiment of the invention is obtained by combining a plurality of target prediction models, and can be better implemented for large-scale training samples.
The embodiment of the invention adopts a machine learning algorithm combining a neural network model, a support vector machine model and a random forest model, and verifies and improves the model by analyzing a large amount of data. Embodiments of the present invention focus on studies and algorithms that find patterns in data and use these patterns to make predictions.
In one embodiment of the invention, the plurality of initial predictive models includes: the first prediction model adopts a neural network model;
taking a plurality of mechanism indexes in a mechanism sample as input, taking a risk label in the mechanism sample as output, training a plurality of initial prediction models to obtain a plurality of target prediction models, and the method comprises the following steps:
and training the first initial prediction model by taking a plurality of mechanism indexes in the mechanism sample as input and taking the risk labels in the mechanism sample as output to obtain a first target prediction model.
The most basic structure of the neural network is a neuron which is often called, the neuron has a mapping function which combines with the neuron, linear or non-linear conversion is completed under the assistance of threshold and weight, and memory is performed (the learned things are memorized, and the memory is realized through the change of the weight, and the threshold is constant). The most basic mechanism of the BP neural network should be, in particular, the perceptron. One perceptron is in fact the simplest BP neural network. The neuron type multi-channel optical fiber comprises two layers of neurons, wherein an input layer receives an external input signal and then transmits the signal to an output layer, and the output layer is an M-P neuron (a neuron with a threshold function). Before the BP network training, network parameters need to be set, which mainly include setting of network structure parameters and training parameters. The network structure parameters mainly include the setting of the number of network layers and the number of nodes in each layer. The excessive number of layers of the neural network can cause the occurrence of an overfitting phenomenon, and the three-layer BP neural network is most commonly applied and can solve most of the nonlinear data processing problems in reality. The method also selects a BP neural network with a three-layer structure comprising an input layer, a single-layer hidden layer and an output layer as a prediction model. The number of nodes in the input layer is determined by the number of sample indexes. The number of nodes of the output layer is determined to be 2, the output vector is (1,0) and represents a company classified as 0, namely credit abnormal, and (0,1) represents a company classified as 1, namely credit normal.
In one embodiment of the invention, the plurality of initial predictive models includes: the second prediction model adopts a support vector machine model;
taking a plurality of mechanism indexes in a mechanism sample as input, taking a risk label in the mechanism sample as output, training a plurality of initial prediction models to obtain a plurality of target prediction models, and the method comprises the following steps:
and training the second initial prediction model by taking a plurality of mechanism indexes in the mechanism sample as input and taking the risk labels in the mechanism sample as output to obtain a second target prediction model.
The support vector machine has the characteristic of high classification accuracy on a small sample. After finding the optimal separation hyperplane, the support vector machine can classify the data population based on hyperplane (hyperplane). Support vector machines are good at performing binary classification operations between variable X and other variables, whether or not their relationship is linear. The method applies the support vector machine method to credit risk early warning, hopes that the SVM model can obtain good classification effect under the condition of less statistical samples, can rapidly judge companies with credit risk in crisis states in a plurality of local financial institutions better, and plays a role in credit risk early warning of the local financial institutions.
In one embodiment of the invention, the plurality of initial predictive models includes: a third prediction model, wherein the third prediction model adopts a random forest model;
taking a plurality of mechanism indexes in a mechanism sample as input, taking a risk label in the mechanism sample as output, training a plurality of initial prediction models to obtain a plurality of target prediction models, and the method comprises the following steps:
and training the third initial prediction model by taking a plurality of mechanism indexes in the mechanism sample as input and taking the risk labels in the mechanism sample as output to obtain a third target prediction model.
The random forest is actually obtained by randomly and repeatedly extracting k samples from an original training sample set N by using a self-help resampling technology, generating different training sample sets, and generating corresponding k classification trees according to the generated sample sets to form the random forest.
The advantages of random forests are many, the most prominent of which is that each tree in the algorithm grows to the greatest extent possible (and does not prune), which ensures that the learner learns deeper and more carefully. Meanwhile, two randomness properties, namely randomly selected samples and features, are added into the algorithm, so that the model is more difficult to fall into overfitting during deep learning.
The output result of the mechanism risk prediction model can divide the evaluated mechanism into a good mechanism and a bad mechanism according to the risk, if the actual risk occurrence frequency of the mechanism is matched with the probability of model estimation, the mechanism risk prediction model has good prediction capability, namely, the supervision model can discriminate the quality of the mechanism.
The method can also verify the prediction capability of the mechanism risk prediction model, and specifically comprises the steps of judging a bad mechanism by using the mechanism risk prediction model, namely, a mechanism which is expected to have negative conditions such as bankruptcy and default, comparing the bad mechanism with the actual mechanism which has negative conditions such as bankruptcy and default in a certain time range, and verifying the effectiveness of the mechanism risk prediction model.
According to the embodiment of the invention, through researching the relevant technology of a machine learning algorithm, a neural network model, a support vector machine model and a random forest model are established, and through learning the relation between financial data and non-financial data in a local financial institution and credit risk of the local financial institution, the structure, the mode and the law in the data are found out, so that under the condition that the financial data and the non-financial data of a certain local financial institution are known, the learned structure, the learned mode and the learned law are used for warning or normally predicting the credit risk of the local financial institution, whether the local financial institution is abnormal credit or normal credit at present is judged, and the credit risk is better prevented.
The method of the embodiment of the invention can assist in establishing the local financial institution risk model, improve the validity verification of the model, improve the fast and efficient judgment of whether the local financial institution is credit abnormal or credit normal at present by local financial supervision authorities of various regions, better prevent the credit risk and improve the local economic operation level.
Fig. 3 is a schematic diagram of a flow of another training method of an organization risk prediction model according to an embodiment of the present invention. The embodiment of the invention provides a training method of a mechanism risk prediction model, which comprises the following steps of:
step 301: the method comprises the steps of obtaining a mechanism sample set, wherein the mechanism sample set comprises a plurality of mechanism samples, and each mechanism sample comprises a plurality of mechanism indexes and risk labels.
Step 302: and training the plurality of initial prediction models to obtain a plurality of target prediction models by taking a plurality of mechanism indexes in the mechanism sample as input and taking a risk label in the mechanism sample as output.
Step 303: performance parameters of each target prediction model are determined.
The performance parameters are used to indicate the accuracy of the model. The performance parameters include at least one of: a receiver operating characteristic curve, a confusion matrix, a model score, a model accuracy or a model accuracy.
In one embodiment of the invention, a test sample set is generated, the test sample set comprising a plurality of institutional samples, each institutional sample comprising a plurality of institutional indicators and a risk label. Inputting sample data in a test sample set into the target prediction models respectively, processing the input sample data by the target prediction models respectively to obtain a plurality of prediction results, and obtaining the performance parameters of each target prediction model according to the difference between each prediction result and the sample risk label.
Step 304: and determining the weight corresponding to each target prediction model according to the performance parameters.
The smaller the difference between the prediction result and the sample risk label is, the higher the performance parameter of the target prediction model is, the higher the performance parameter of the model is, the more accurate the target prediction model is represented, and the weight corresponding to the target prediction model is larger.
Optionally, the ratio of the weight corresponding to each target prediction model is the same as the ratio of the performance parameter of the target prediction model. For example, the performance parameters of the target prediction model are 60%, 30% and 10%, respectively, and the weights of the target prediction model may be 0.6, 0.3 and 0.1, respectively.
In one embodiment of the present invention, determining the weight corresponding to each target prediction model according to the performance parameter includes:
and if the current performance parameter is smaller than the performance threshold, the weight corresponding to the current target prediction model is 0, and the current target prediction model corresponds to the current performance parameter.
The performance threshold may be set according to specific needs. If the current performance parameter of the current target prediction model is smaller than the performance threshold, it indicates that the current target prediction model cannot provide an accurate and effective mechanism risk prediction result, so that the weight corresponding to the current target prediction model is 0.
Step 305: and determining the mechanism risk prediction model according to the plurality of target prediction models and the weights corresponding to the target prediction models.
In the embodiment of the invention, the weight corresponding to each target prediction model is determined according to the performance parameters of each target prediction model, so that the weight of the target prediction model with higher performance parameters is smaller, and the weight of the target prediction model with lower performance parameters is smaller, and therefore, the mechanism risk prediction model has better mechanism risk identification effect.
Fig. 4 is a schematic diagram of a flow of a facility risk prediction method according to an embodiment of the present invention. The embodiment of the invention provides a mechanism risk prediction method, as shown in fig. 4, comprising the following steps:
step 401: a plurality of index values of the target mechanism are acquired.
The index value of the organization can be the statistical value of the organization operation condition under various statistical dimensions. The statistical dimensions may include: risk alert dimensions, profitability dimensions, liquidity risk dimensions, operations, and the like. The institutional indexes may include: statistical net profits, end-of-term owner equity totals, operating revenues, operating cash inflows, reputation, audit opinions, guild penalty information, etc. over a certain statistical period. Wherein, the reputation, the audit opinion, the certificate punishment information and the like are expressed by specific scores.
Step 402: and inputting the index values into a mechanism risk prediction model to obtain a risk prediction result of the target mechanism, wherein the mechanism risk prediction model is generated by the target prediction models and weights corresponding to the target prediction models.
Different target prediction models are obtained by training through different training methods. The training method is any machine learning algorithm or any deep learning algorithm. Such as linear regression algorithms, support vector machine algorithms, nearest neighbor/k-nearest neighbor algorithms, logistic regression algorithms, decision tree algorithms, k-means algorithms, random forest algorithms, naive bayes algorithms, back propagation, random gradient descent, and the like.
The output result of the mechanism risk prediction model can be 0 or 1,0 represents that the mechanism is at risk, and 1 represents that the mechanism is not at risk. The output of each target prediction model may also be a risk level of the organization, such as high risk, medium risk, low risk, no risk, etc. The output may also be represented by a specific score, with different scores indicating the probability of different risks that an organization may be at.
The mechanism risk prediction model is obtained from a plurality of target prediction models and a combination of weights corresponding to the target prediction models. The results from the institutional risk prediction model may be a weighted sum of the results from the respective target prediction models.
For example, the system has three target prediction models, and the weights of the three target prediction models are 0.7, 0.2 and 0.1 respectively. For a certain target mechanism, the output results of the three target prediction models are respectively as follows: 0. 0 and 1. Wherein, 0 represents the mechanism with risk, and 1 represents the mechanism without risk. The probability of risk for the target mechanism is 0.7+ 0.2-0.9 and the probability of risk for the target mechanism not being present is 0.1.
As another example, the system has two target prediction models in common, and the weights of the two target prediction models are 0.6 and 0.4, respectively. For a certain target mechanism, the output fraud risk scores of the three target prediction models are respectively as follows: 80 and 50. The target institution fraud risk score is 80 x 0.6+50 x 0.4-68.
In the embodiment of the invention, the mechanism risk prediction model is determined according to the target prediction models and the weights corresponding to the target prediction models. The determined mechanism risk prediction model is obtained by training a mechanism sample set with more samples, and the accuracy of the obtained mechanism risk prediction result can be higher by a traditional mode of manually scoring indexes to carry out risk prediction on the mechanism.
Furthermore, single models tend to be sensitive to the choice of parameters and functions, and thus difficult to implement for large-scale training samples. The mechanism risk prediction model of the embodiment of the invention is obtained by combining a plurality of target prediction models, and can be better implemented for large-scale training samples.
Optionally, the weight corresponding to the target prediction model is determined according to a performance parameter of the target prediction model.
The performance parameters are used to indicate the accuracy of the model. The performance parameters include at least one of: a receiver operating characteristic curve, a confusion matrix, a model score, a model accuracy or a model accuracy.
In the embodiment of the invention, the weight corresponding to each target prediction model is determined according to the performance parameters of each target prediction model, so that the weight of the target prediction model with higher performance parameters is smaller, and the weight of the target prediction model with lower performance parameters is smaller, and therefore, the mechanism risk prediction model has better mechanism risk identification effect.
Fig. 5 is a schematic structural diagram of a training apparatus for an organization risk prediction model according to an embodiment of the present invention, including:
the system comprises a sample acquisition module 501, a risk analysis module and a risk analysis module, wherein the sample acquisition module is used for acquiring a mechanism sample set, the mechanism sample set comprises a plurality of mechanism samples, and each mechanism sample comprises a plurality of mechanism indexes and risk labels;
a model training module 502, configured to train multiple initial prediction models to obtain multiple target prediction models, where multiple mechanism indexes in a mechanism sample are used as inputs and risk labels in the mechanism sample are used as outputs;
a weight determining module 503, configured to determine a weight corresponding to each target prediction model;
a model determining module 504, configured to determine the mechanism risk prediction model according to the multiple target prediction models and weights corresponding to the target prediction models.
Optionally, the weight determining module 503 is specifically configured to: determining performance parameters of each target prediction model;
and determining the weight corresponding to each target prediction model according to the performance parameters.
Optionally, the performance parameter comprises at least one of: a receiver operating characteristic curve, a confusion matrix, a model score, a model accuracy or a model accuracy.
Optionally, the weight determining module 503 is specifically configured to: and if the current performance parameter is smaller than the performance threshold, the weight corresponding to the current target prediction model is 0, and the current target prediction model corresponds to the current performance parameter.
Optionally, the plurality of initial predictive models comprises: the first prediction model adopts a neural network model;
the model training module 502 is specifically configured to: and training the first initial prediction model by taking a plurality of mechanism indexes in the mechanism sample as input and taking the risk labels in the mechanism sample as output to obtain a first target prediction model.
Optionally, the plurality of initial predictive models comprises: the second prediction model adopts a support vector machine model;
the model training module 502 is specifically configured to: and training the second initial prediction model by taking a plurality of mechanism indexes in the mechanism sample as input and taking the risk labels in the mechanism sample as output to obtain a second target prediction model.
Optionally, the plurality of initial predictive models comprises: a third prediction model, wherein the third prediction model adopts a random forest model;
the model training module 502 is specifically configured to: and training the third initial prediction model by taking a plurality of mechanism indexes in the mechanism sample as input and taking the risk labels in the mechanism sample as output to obtain a third target prediction model.
Optionally, the sample acquiring module 501 is specifically configured to: acquiring financial data and non-financial data for a plurality of institutions;
for each of a plurality of institutions: determining a plurality of metrics for the organization based on financial data and non-financial data for the organization; determining a risk label for the organization; and constructing a mechanism sample according to the multiple indexes of the mechanism and the risk label of the mechanism.
Fig. 6 is a schematic structural diagram of a mechanism risk prediction apparatus according to an embodiment of the present invention, including:
an index obtaining module 601, configured to obtain a plurality of index values of a target mechanism;
the risk prediction module 602 is configured to input the plurality of index values into a mechanism risk prediction model to obtain a risk prediction result of the target mechanism, where the mechanism risk prediction model is generated by the plurality of target prediction models and weights corresponding to the target prediction models.
Optionally, the weight corresponding to the target prediction model is determined according to a performance parameter of the target prediction model.
An embodiment of the present invention provides an electronic device, including:
one or more processors;
a storage device for storing one or more programs,
when the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the method of any of the embodiments described above.
Referring now to FIG. 7, shown is a block diagram of a computer system 700 suitable for use with a terminal device implementing an embodiment of the present invention. The terminal device shown in fig. 7 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present invention.
As shown in fig. 7, the computer system 700 includes a Central Processing Unit (CPU)701, which can perform various appropriate actions and processes in accordance with a program stored in a Read Only Memory (ROM)702 or a program loaded from a storage section 708 into a Random Access Memory (RAM) 703. In the RAM 703, various programs and data necessary for the operation of the system 700 are also stored. The CPU 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input/output (I/O) interface 705 is also connected to bus 704.
The following components are connected to the I/O interface 705: an input portion 706 including a keyboard, a mouse, and the like; an output section 707 including a display such as a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, and a speaker; a storage section 708 including a hard disk and the like; and a communication section 709 including a network interface card such as a LAN card, a modem, or the like. The communication section 709 performs communication processing via a network such as the internet. A drive 710 is also connected to the I/O interface 705 as needed. A removable medium 711 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 710 as necessary, so that a computer program read out therefrom is mounted into the storage section 708 as necessary.
In particular, according to the embodiments of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication section 709, and/or installed from the removable medium 711. The computer program performs the above-described functions defined in the system of the present invention when executed by the Central Processing Unit (CPU) 701.
It should be noted that the computer readable medium shown in the present invention can be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present invention, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present invention, however, a computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, fiber optic cable, RF, etc., or any suitable combination of the foregoing.
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The modules described in the embodiments of the present invention may be implemented by software or hardware. The described modules may also be provided in a processor, which may be described as: a processor includes a sample acquisition module, a model training module, a weight determination module, and a model determination module. Where the names of these modules do not in some cases constitute a limitation on the module itself, for example, a sample acquisition module may also be described as "acquiring a facility sample set comprising a plurality of facility samples, each of which comprises a plurality of facility indices and risk labels.
As another aspect, the present invention also provides a computer-readable medium that may be contained in the apparatus described in the above embodiments; or may be separate and not incorporated into the device. The computer readable medium carries one or more programs which, when executed by a device, cause the device to comprise:
acquiring a mechanism sample set, wherein the mechanism sample set comprises a plurality of mechanism samples, and each mechanism sample comprises a plurality of mechanism indexes and risk labels;
taking a plurality of mechanism indexes in the mechanism sample as input, taking a risk label in the mechanism sample as output, and training a plurality of initial prediction models to obtain a plurality of target prediction models;
determining the weight corresponding to each target prediction model;
and determining a mechanism risk prediction model according to the target prediction models and the weights corresponding to the target prediction models.
According to the technical scheme of the embodiment of the invention, the mechanism risk prediction model is determined by utilizing a plurality of target prediction models and weights corresponding to the target prediction models. The determined mechanism risk prediction model is obtained by training a mechanism sample set with more samples. Compared with the traditional mode of manually scoring indexes to carry out risk prediction on the mechanism, the accuracy of the mechanism risk prediction result determined by the mechanism risk prediction model is higher.
Furthermore, single models tend to be sensitive to the choice of parameters and functions, and thus difficult to implement for large-scale training samples. The mechanism risk prediction model of the embodiment of the invention is obtained by combining a plurality of target prediction models, and can be better implemented for large-scale training samples.
The above-described embodiments should not be construed as limiting the scope of the invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions can occur, depending on design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (15)
1. A method for training a mechanism risk prediction model is characterized by comprising the following steps:
acquiring a mechanism sample set, wherein the mechanism sample set comprises a plurality of mechanism samples, and each mechanism sample comprises a plurality of mechanism indexes and risk labels;
taking a plurality of mechanism indexes in the mechanism sample as input, taking a risk label in the mechanism sample as output, and training a plurality of initial prediction models to obtain a plurality of target prediction models;
determining the weight corresponding to each target prediction model;
and determining a mechanism risk prediction model according to the target prediction models and the weights corresponding to the target prediction models.
2. The method of claim 1, wherein determining the weight corresponding to each of the target prediction models comprises:
determining a performance parameter of each of the target prediction models;
and determining the weight corresponding to each target prediction model according to the performance parameters.
3. The method of claim 2, wherein the performance parameter comprises at least one of: a receiver operating characteristic curve, a confusion matrix, a model score, a model accuracy or a model accuracy.
4. The method of claim 2, wherein determining the weight for each of the target predictive models based on the performance parameters comprises:
and if the current performance parameter is smaller than the performance threshold, the weight corresponding to the current target prediction model is 0, and the current target prediction model corresponds to the current performance parameter.
5. The method of claim 1, wherein the plurality of initial predictive models comprises: a first prediction model, wherein the first prediction model adopts a neural network model;
the training of the multiple initial prediction models by taking the multiple mechanism indexes in the mechanism sample as input and the risk labels in the mechanism sample as output to obtain multiple target prediction models comprises the following steps:
and training a first initial prediction model by taking a plurality of mechanism indexes in the mechanism sample as input and taking a risk label in the mechanism sample as output to obtain a first target prediction model.
6. The method of claim 1, wherein the plurality of initial predictive models comprises: a second prediction model, which adopts a support vector machine model;
the training of the multiple initial prediction models by taking the multiple mechanism indexes in the mechanism sample as input and the risk labels in the mechanism sample as output to obtain multiple target prediction models comprises the following steps:
and training a second initial prediction model by taking a plurality of mechanism indexes in the mechanism sample as input and taking a risk label in the mechanism sample as output to obtain a second target prediction model.
7. The method of claim 1, wherein the plurality of initial predictive models comprises: a third prediction model, wherein the third prediction model adopts a random forest model;
the training of the multiple initial prediction models by taking the multiple mechanism indexes in the mechanism sample as input and the risk labels in the mechanism sample as output to obtain multiple target prediction models comprises the following steps:
and training a third initial prediction model by taking a plurality of mechanism indexes in the mechanism sample as input and taking a risk label in the mechanism sample as output to obtain a third target prediction model.
8. The method of claim 1, wherein said obtaining a set of institution samples comprises:
acquiring financial data and non-financial data for a plurality of institutions;
for each of the plurality of mechanisms: determining a plurality of metrics for the institution based on the financial data and non-financial data for the institution; determining a risk label for the organization; and constructing the mechanism sample according to the plurality of indexes of the mechanism and the risk label of the mechanism.
9. The method of claim 8, prior to said determining a plurality of metrics for the organization based on financial data and non-financial data for the organization, further comprising:
performing a cleaning process on the data according to the financial data and the non-financial data, the cleaning process including at least one of: missing value processing, outlier processing, data transposition, and data summation.
10. An organization risk prediction method, comprising:
acquiring a plurality of index values of a target mechanism;
and inputting the index values into a mechanism risk prediction model to obtain a risk prediction result of the target mechanism, wherein the mechanism risk prediction model is generated by a plurality of target prediction models and weights corresponding to the target prediction models.
11. The method of claim 10, wherein the weights corresponding to the target prediction model are determined according to performance parameters of the target prediction model.
12. An apparatus for training a risk prediction model of a facility, comprising:
the system comprises a sample acquisition module, a risk analysis module and a risk analysis module, wherein the sample acquisition module is used for acquiring a mechanism sample set, the mechanism sample set comprises a plurality of mechanism samples, and each mechanism sample comprises a plurality of mechanism indexes and risk labels;
the model training module is used for training a plurality of initial prediction models to obtain a plurality of target prediction models by taking a plurality of mechanism indexes in the mechanism sample as input and taking a risk label in the mechanism sample as output;
the weight determining module is used for determining the weight corresponding to each target prediction model;
and the model determining module is used for determining the mechanism risk prediction model according to the target prediction models and the weights corresponding to the target prediction models.
13. An organization risk prediction apparatus, comprising:
the index acquisition module is used for acquiring a plurality of index values of the target mechanism;
and the risk prediction module is used for inputting the index values into a mechanism risk prediction model to obtain a risk prediction result of the target mechanism, and the mechanism risk prediction model is generated by a plurality of target prediction models and weights corresponding to the target prediction models.
14. An electronic device, comprising:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement the method of any one of claims 1-11.
15. A computer-readable medium, on which a computer program is stored, which, when being executed by a processor, carries out the method according to any one of claims 1-11.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011475410.3A CN112561320A (en) | 2020-12-14 | 2020-12-14 | Training method of mechanism risk prediction model, mechanism risk prediction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011475410.3A CN112561320A (en) | 2020-12-14 | 2020-12-14 | Training method of mechanism risk prediction model, mechanism risk prediction method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112561320A true CN112561320A (en) | 2021-03-26 |
Family
ID=75063770
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011475410.3A Pending CN112561320A (en) | 2020-12-14 | 2020-12-14 | Training method of mechanism risk prediction model, mechanism risk prediction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112561320A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113379530A (en) * | 2021-06-09 | 2021-09-10 | 中国工商银行股份有限公司 | User risk determination method and device and server |
| CN113823411A (en) * | 2021-09-26 | 2021-12-21 | 萱闱(北京)生物科技有限公司 | Risk prediction model training method, device, medium and computing equipment |
| CN114091902A (en) * | 2021-11-22 | 2022-02-25 | 支付宝(杭州)信息技术有限公司 | Risk prediction model training method and device, and risk prediction method and device |
| CN114154891A (en) * | 2021-12-08 | 2022-03-08 | 中国建设银行股份有限公司 | Retraining method and retraining device for risk control model |
| CN114629797A (en) * | 2022-03-11 | 2022-06-14 | 阿里巴巴（中国）有限公司 | Bandwidth prediction method, model generation method and equipment |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106408423A (en) * | 2016-11-25 | 2017-02-15 | 泰康保险集团股份有限公司 | Method and system for risk assessment and method for constructing system for risk assessment |
| CN109359818A (en) * | 2018-09-20 | 2019-02-19 | 重庆誉存大数据科技有限公司 | A kind of pair of enterprise carries out the method and six rib star models of risk quantification assessment |
| CN109523117A (en) * | 2018-10-11 | 2019-03-26 | 平安科技（深圳）有限公司 | Risk Forecast Method, device, computer equipment and storage medium |
| CN109993412A (en) * | 2019-03-01 | 2019-07-09 | 百融金融信息服务股份有限公司 | The construction method and device of risk evaluation model, storage medium, computer equipment |
| CN110378545A (en) * | 2018-04-13 | 2019-10-25 | 北京三快在线科技有限公司 | Trade company's credit estimation method and device, storage medium, electronic equipment |
| CN110599335A (en) * | 2019-09-27 | 2019-12-20 | 北京淇瑀信息科技有限公司 | User financial risk assessment method and device based on multiple models |
| CN110866696A (en) * | 2019-11-15 | 2020-03-06 | 成都数联铭品科技有限公司 | Method and device for training shop falling risk assessment model |
| CN110956273A (en) * | 2019-11-07 | 2020-04-03 | 中信银行股份有限公司 | Credit scoring method and system integrating multiple machine learning models |
-
2020
- 2020-12-14 CN CN202011475410.3A patent/CN112561320A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106408423A (en) * | 2016-11-25 | 2017-02-15 | 泰康保险集团股份有限公司 | Method and system for risk assessment and method for constructing system for risk assessment |
| CN110378545A (en) * | 2018-04-13 | 2019-10-25 | 北京三快在线科技有限公司 | Trade company's credit estimation method and device, storage medium, electronic equipment |
| CN109359818A (en) * | 2018-09-20 | 2019-02-19 | 重庆誉存大数据科技有限公司 | A kind of pair of enterprise carries out the method and six rib star models of risk quantification assessment |
| CN109523117A (en) * | 2018-10-11 | 2019-03-26 | 平安科技（深圳）有限公司 | Risk Forecast Method, device, computer equipment and storage medium |
| CN109993412A (en) * | 2019-03-01 | 2019-07-09 | 百融金融信息服务股份有限公司 | The construction method and device of risk evaluation model, storage medium, computer equipment |
| CN110599335A (en) * | 2019-09-27 | 2019-12-20 | 北京淇瑀信息科技有限公司 | User financial risk assessment method and device based on multiple models |
| CN110956273A (en) * | 2019-11-07 | 2020-04-03 | 中信银行股份有限公司 | Credit scoring method and system integrating multiple machine learning models |
| CN110866696A (en) * | 2019-11-15 | 2020-03-06 | 成都数联铭品科技有限公司 | Method and device for training shop falling risk assessment model |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113379530A (en) * | 2021-06-09 | 2021-09-10 | 中国工商银行股份有限公司 | User risk determination method and device and server |
| CN113823411A (en) * | 2021-09-26 | 2021-12-21 | 萱闱(北京)生物科技有限公司 | Risk prediction model training method, device, medium and computing equipment |
| CN114091902A (en) * | 2021-11-22 | 2022-02-25 | 支付宝(杭州)信息技术有限公司 | Risk prediction model training method and device, and risk prediction method and device |
| CN114154891A (en) * | 2021-12-08 | 2022-03-08 | 中国建设银行股份有限公司 | Retraining method and retraining device for risk control model |
| CN114629797A (en) * | 2022-03-11 | 2022-06-14 | 阿里巴巴（中国）有限公司 | Bandwidth prediction method, model generation method and equipment |
| CN114629797B (en) * | 2022-03-11 | 2024-03-08 | 阿里巴巴（中国）有限公司 | Bandwidth prediction method, model generation method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20210049700A1 (en) | System and method for machine learning architecture for enterprise capitalization | |
| Shah et al. | A comprehensive review on multiple hybrid deep learning approaches for stock prediction | |
| CN112561320A (en) | Training method of mechanism risk prediction model, mechanism risk prediction method and device | |
| CN110378786B (en) | Model training method, default transmission risk identification method, device and storage medium | |
| Keramati et al. | A proposed classification of data mining techniques in credit scoring | |
| CN110852881B (en) | Risk account identification method and device, electronic equipment and medium | |
| US20090043715A1 (en) | Method to Continuously Diagnose and Model Changes of Real-Valued Streaming Variables | |
| CN113095927B (en) | Method and equipment for identifying suspected transactions of backwashing money | |
| Kompella et al. | Stock market prediction using machine learning methods | |
| CN113627566A (en) | Early warning method and device for phishing and computer equipment | |
| CN110930038A (en) | Loan demand identification method, loan demand identification device, loan demand identification terminal and loan demand identification storage medium | |
| Agarwal et al. | Sentiment analysis in stock price prediction: a comparative study of algorithms | |
| CN115760332A (en) | Risk prediction method, system, medium and device based on enterprise data analysis | |
| Boz et al. | Reassessment and monitoring of loan applications with machine learning | |
| CN117252688A (en) | Financial risk assessment method, system, terminal equipment and storage medium | |
| Wimmer et al. | Leveraging vision-language models for granular market change prediction | |
| CN114612239A (en) | Stock public opinion monitoring and wind control system based on algorithm, big data and artificial intelligence | |
| CN110704803A (en) | Target object evaluation value calculation method and device, storage medium and electronic device | |
| CN116911994B (en) | External trade risk early warning system | |
| Hajek et al. | Interval-valued fuzzy cognitive maps with genetic learning for predicting corporate financial distress | |
| CN109741172B (en) | Credit early warning method, device, system and storage medium | |
| Yang et al. | An evidential reasoning rule-based ensemble learning approach for evaluating credit risks with customer heterogeneity | |
| CN116452225A (en) | Object classification method, device, computer equipment and storage medium | |
| Lee et al. | Application of machine learning in credit risk scorecard | |
| Nam | Modelling Stock Market Manipulation in Online Forums |
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 |