CN117217522A - Financial pre-billing risk management and control system based on artificial intelligence and operation method thereof - Google Patents
Financial pre-billing risk management and control system based on artificial intelligence and operation method thereof Download PDFInfo
- Publication number
- CN117217522A CN117217522A CN202311113547.8A CN202311113547A CN117217522A CN 117217522 A CN117217522 A CN 117217522A CN 202311113547 A CN202311113547 A CN 202311113547A CN 117217522 A CN117217522 A CN 117217522A
- Authority
- CN
- China
- Prior art keywords
- data
- enterprise
- credit
- information
- source heterogeneous
- 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
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 42
- 238000013480 data collection Methods 0.000 claims abstract description 28
- 238000004451 qualitative analysis Methods 0.000 claims abstract description 15
- 238000004445 quantitative analysis Methods 0.000 claims abstract description 15
- 239000011159 matrix material Substances 0.000 claims description 65
- 238000007726 management method Methods 0.000 claims description 46
- 238000013507 mapping Methods 0.000 claims description 35
- 239000013598 vector Substances 0.000 claims description 33
- 238000012549 training Methods 0.000 claims description 29
- 238000011176 pooling Methods 0.000 claims description 27
- 238000009826 distribution Methods 0.000 claims description 23
- 238000004590 computer program Methods 0.000 claims description 17
- 230000007246 mechanism Effects 0.000 claims description 15
- 238000012545 processing Methods 0.000 claims description 14
- 238000003860 storage Methods 0.000 claims description 14
- 208000000785 Invasive Pulmonary Aspergillosis Diseases 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 10
- 238000011161 development Methods 0.000 claims description 10
- 238000009411 base construction Methods 0.000 claims description 8
- 238000007781 pre-processing Methods 0.000 claims description 8
- 238000006243 chemical reaction Methods 0.000 claims description 7
- 238000004140 cleaning Methods 0.000 claims description 6
- 230000009467 reduction Effects 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 6
- 238000001212 derivatisation Methods 0.000 claims description 5
- 238000004801 process automation Methods 0.000 claims description 2
- 238000003062 neural network model Methods 0.000 description 12
- 230000008569 process Effects 0.000 description 12
- 230000006870 function Effects 0.000 description 11
- 238000013528 artificial neural network Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 238000010606 normalization Methods 0.000 description 6
- 238000012360 testing method Methods 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000011144 upstream manufacturing Methods 0.000 description 4
- 238000012550 audit Methods 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 238000011017 operating method Methods 0.000 description 2
- 238000013500 data storage Methods 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000013439 planning Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Landscapes
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The application relates to the technical field of financial risk management and control, and provides an artificial intelligence-based financial pre-billing risk management and control system and an operation method thereof. The method comprises the following steps: the method comprises the steps of collecting multi-source heterogeneous data based on an enterprise wind control data collection robot, and carrying out qualitative analysis, quantitative analysis and data statistics on the multi-source heterogeneous data to obtain structured data; carrying out credit grade classification on the structured data based on the enterprise credit prediction model to obtain a credit grade classification result; and under the condition that the bill center receives a pre-ticketing request sent by an enterprise, sending the structured data and the credit class classification result to the bill center for checking. The financial pre-billing risk management and control system based on artificial intelligence and the operation method thereof realize the automatic collection of multi-source heterogeneous data of enterprises, reduce the manual operation flow, improve the financial pre-billing checking efficiency and improve the comprehensiveness and accuracy of wind control rating.
Description
Technical Field
The application relates to the technical field of financial risk management and control, in particular to an artificial intelligence-based financial pre-billing risk management and control system and an operation method thereof.
Background
In the related technology, the financial risk of the pre-invoicing is controlled by adopting a manual checking mode, so that financial staff is required to widely collect external information of an enterprise, the capability and willingness of the enterprise to fulfill related contracts and economic commitments are generally evaluated according to the mastered external information, and finally, whether the pre-invoicing application of the enterprise is agreed is judged by combining own experience; when the number of the external information data searches is large, the data types are complicated, so that the manual data arrangement is more difficult, and when the financial staff performs the pre-invoicing audit according to working experience, the financial staff is influenced by subjective judgment, and different financial staff can generate different assessment results, so that the pre-invoicing audit decision is inaccurate.
Disclosure of Invention
The embodiment of the application provides a financial pre-invoicing risk management and control system based on artificial intelligence and an operation method thereof, which are used for solving the technical problems that data arrangement is difficult and pre-invoicing audit decision is inaccurate when a manual checking mode is adopted to control the financial risk of pre-invoicing.
In a first aspect, an embodiment of the present application provides an artificial intelligence based financial pre-ticketing risk management and control system, including:
The enterprise wind control data acquisition robot is built based on an intelligent process automation IPA development platform, and is used for acquiring multi-source heterogeneous data, and carrying out qualitative analysis, quantitative analysis and data statistics on the multi-source heterogeneous data to obtain structured data; wherein the multi-source heterogeneous data comprises enterprise internal data and enterprise external data;
the enterprise credit prediction model is trained based on a convolution attention mechanism and is used for classifying the credit level of the structured data to obtain a credit level classification result;
and the enterprise credit service module is used for sending the structured data and the credit class classification result to the bill center for checking under the condition that the bill center receives a pre-ticketing request sent by an enterprise.
In one embodiment, the enterprise wind control data collection robot comprises:
the internal information acquisition module is used for acquiring historical transaction data from a corporate staff stock holding plan ESOP system and a bill center database to obtain the corporate internal data;
The external information acquisition module is used for acquiring enterprise credit information from an enterprise information base through the Internet to obtain the enterprise external data;
the information standardization module is used for carrying out standardization processing on quantitative information in the multi-source heterogeneous data, carrying out digital processing on the qualitative information in the multi-source heterogeneous data, and counting the quantity of similar information in the multi-source heterogeneous data to obtain the structured data;
the enterprise information base construction module is used for constructing an enterprise information base according to the standardized data, and the enterprise information base is used for digitally managing multi-source heterogeneous data.
In one embodiment, the enterprise wind control data collection robot further comprises:
the data preprocessing module is used for carrying out data cleaning, format conversion, data derivatization and data dimension reduction on the multi-source heterogeneous data to obtain a data tag, wherein the data tag is a training tag associated with the enterprise credit prediction model.
In one embodiment, the enterprise credit prediction model includes:
The full-connection layers are used for expanding the feature matrix corresponding to the structured data to obtain an expanded feature matrix;
the convolution layer is used for carrying out feature mapping on the expanded feature matrix through convolution check to obtain a mapping matrix;
the pooling layer is used for carrying out sampling treatment on the mapping matrix to obtain a pooling matrix;
the attention layer is used for mapping the feature vectors corresponding to the enterprise internal data and the feature vectors corresponding to the enterprise external data in the pooling matrix to a unified feature space to obtain a feature hiding matrix, and attention weight distribution information is obtained based on the feature hiding matrix and mapping parameters;
and the Softmax layer is used for carrying out classification probability calculation on the attention weight distribution information to obtain the credit class classification result.
In a second aspect, an embodiment of the present application provides an operation method of an artificial intelligence based financial pre-ticketing risk management and control system, including:
the method comprises the steps of collecting multi-source heterogeneous data based on an enterprise wind control data collection robot, and carrying out qualitative analysis, quantitative analysis and data statistics on the multi-source heterogeneous data to obtain structured data; the multi-source heterogeneous data comprises enterprise internal data and enterprise external data, and the enterprise wind control data acquisition robot is built based on an intelligent flow automatic IPA development platform;
Carrying out credit grade classification on the structured data based on an enterprise credit prediction model to obtain a credit grade classification result, wherein the enterprise credit prediction model is trained based on a convolution attention mechanism;
and under the condition that the bill center receives a pre-ticketing request sent by an enterprise, sending the structured data and the credit class classification result to the bill center for checking.
In one embodiment, the collecting multi-source heterogeneous data based on the enterprise wind control data collecting robot, and performing qualitative analysis, quantitative analysis and data statistics on the multi-source heterogeneous data to obtain structured data, including:
acquiring historical transaction data from a corporate staff stock holding plan ESOP system and a bill center database based on an internal information acquisition module to obtain the internal data of the corporation; the external information acquisition module is used for acquiring enterprise credit information from an enterprise information base through the Internet to obtain the enterprise external data;
based on an information standardization module, carrying out standardization processing on quantitative information in the multi-source heterogeneous data, carrying out digital processing on qualitative information in the multi-source heterogeneous data, and counting the quantity of similar information in the multi-source heterogeneous data to obtain the structured data;
And constructing an enterprise information base according to the standardized data based on an enterprise information base construction module, wherein the enterprise information base is used for digitally managing the multi-source heterogeneous data.
In one embodiment, the enterprise credit prediction model includes a plurality of fully connected, convolved, pooled, attention, and Softmax layers;
the credit rating classification of the structured data based on the enterprise credit prediction model is performed to obtain a credit rating classification result, which comprises the following steps:
the full connection layer is used for expanding the feature matrix corresponding to the structured data to obtain an expanded feature matrix;
performing feature mapping on the extended feature matrix through convolution check based on a convolution layer to obtain a mapping matrix;
sampling the mapping matrix based on a pooling layer to obtain a pooling matrix;
mapping the feature vectors corresponding to the enterprise internal data and the feature vectors corresponding to the enterprise external data in the pooling matrix to a unified feature space based on the attention layer to obtain a feature hiding matrix, and obtaining attention weight distribution information based on the feature hiding matrix and mapping parameters;
and carrying out classification probability calculation on the attention weight distribution information based on a Softmax layer to obtain the credit class classification result.
In a third aspect, an embodiment of the present application provides an electronic device, including a processor and a memory storing a computer program, where the processor implements the steps of the operating method of the artificial intelligence based financial pre-ticketing risk management and control system according to the first aspect when executing the program.
In a fourth aspect, embodiments of the present application provide a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method of operating an artificial intelligence based financial pre-ticketing risk management and control system of the first aspect.
In a fifth aspect, an embodiment of the present application provides a computer program product comprising a computer program which, when executed by a processor, implements the steps of the method for operating an artificial intelligence based financial pre-ticketing risk management and control system according to the first aspect.
According to the financial pre-invoicing risk management and control system based on artificial intelligence and the operation method thereof, provided by the embodiment of the application, multi-source heterogeneous data are collected through an enterprise wind control data collection robot, and information standardization analysis is carried out on the multi-source heterogeneous data to obtain structured data; and finally, the structured data and the credit class classification result are sent to the bill center for checking under the condition that the bill center receives a pre-opening ticket request sent by an enterprise through an enterprise credit service module, so that the automatic collection of multi-source heterogeneous data of the enterprise is realized, the manual operation flow is reduced, the financial pre-opening ticket checking efficiency is improved, and meanwhile, the comprehensiveness and the accuracy of the wind control rating are improved.
Drawings
In order to more clearly illustrate the application or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the application, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic diagram of an artificial intelligence based financial pre-ticketing risk management and control system according to an embodiment of the present application;
FIG. 2 is a second schematic diagram of an artificial intelligence based financial pre-ticketing risk management and control system according to an embodiment of the present application;
FIG. 3 is a schematic flow chart of a working mechanism of an enterprise wind control data acquisition robot according to an embodiment of the present application;
FIG. 4 is a flow chart of a method of operation of an artificial intelligence based financial pre-ticketing risk management and control system provided by an embodiment of the application;
fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present application;
reference numerals:
110: an enterprise wind control data acquisition robot; 120: an enterprise credit prediction model; 130: and the enterprise credit service module.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
Fig. 1 is a schematic structural diagram of an artificial intelligence-based financial pre-ticketing risk management and control system according to an embodiment of the present application. Referring to fig. 1, an embodiment of the present application provides an operation method of an artificial intelligence based financial pre-ticketing risk management and control system, which includes an enterprise wind control data acquisition robot 110, an enterprise credit prediction model 120, and an enterprise credit service module 130.
The enterprise wind control data acquisition robot 110 is built based on an intelligent flow automatic IPA development platform, and the enterprise wind control data acquisition robot 110 is used for acquiring multi-source heterogeneous data, and carrying out qualitative analysis, quantitative analysis and data statistics on the multi-source heterogeneous data to obtain structured data; wherein the multi-source heterogeneous data includes intra-enterprise data and extra-enterprise data.
In this embodiment, the enterprise wind control collection robot is built using an IPA (Intelligent Process Automatic, IPA) capability platform.
In this embodiment, the enterprise wind control data collection robot 110 may collect enterprise internal data, for example, the enterprise wind control data collection robot 110 collects historical transaction data such as arrears, cash flows, enterprise-related revenue sizes, asset liabilities, and product ordering sizes from ticketing parties in (Employee Stock Ownership Plans, corporate staff holding stock plans) systems and ticket center databases.
In this embodiment, the enterprise wind control data collection robot 110 may collect enterprise external data, for example, by capturing enterprise information on an enterprise credit website, including data on revenue size, business efficiency, financing, equity change, industry ranking, upstream and downstream enterprise operations, etc., as enterprise external data.
In this embodiment, the enterprise wind control data collection robot 110 may perform quantitative analysis, qualitative analysis and statistics of the same type of information on the collected multi-source heterogeneous data, and convert the enterprise information in different formats into a unified format to obtain corresponding structured data.
The enterprise credit prediction model 120 is trained based on a convolution attention mechanism, and the enterprise credit prediction model 120 is used for classifying the credit level of the structured data to obtain a credit level classification result.
In this embodiment, the enterprise credit prediction model 120 classifies the credit class of the ticket buyer enterprise into a plurality of classes, for example, the credit class is classified into 5 classes a-E, including a (almost risk-free), B (risk-less), C (risk-less), D (risk-greater), and E (high risk), and the enterprise credit prediction model 120 predicts the credit class of the structured data to obtain a corresponding credit class classification result.
In this embodiment, the enterprise credit prediction model 120 may be trained based on a convolutional attention neural network algorithm.
For example, a self-attention layer is set in the convolutional neural network, and the input characteristic vectors corresponding to the enterprise internal data and the enterprise external data are mapped on a unified characteristic control, so that a more accurate credit class classification result is obtained.
The enterprise credit service module 130 is configured to send the structured data and the credit class classification result to the ticket center for checking if the ticket center receives a pre-ticketing request sent by the enterprise.
In this embodiment, the enterprise credit prediction model 120 is automatically invoked to rank predict the collected multi-source heterogeneous data when the ticket center accepts the enterprise's pre-ticketing application, and based on this determination, whether to approve the enterprise's pre-ticketing application.
Fig. 2 is a second schematic structural diagram of an artificial intelligence-based financial pre-ticketing risk management and control system, where in the embodiment shown in fig. 2, an enterprise wind control data collection robot 110 captures (corresponds to enterprise internal and external credit data), selects enterprise credit related data (structured data after information standardization and data preprocessing) from the captured data, and inputs the selected data to an enterprise credit prediction model 120 for classification, so as to obtain a credit class classification result; the enterprise credit forecast evaluation service capability of the enterprise credit service module 130 is automatically invoked when the ticket center receives an enterprise pre-ticketing application to obtain enterprise credit related data and credit class classification results.
Enterprise wind control data collection robot 110 and enterprise credit prediction model 120 provide enterprise credit related data and credit rating classification results, respectively, to enterprise credit service module 130 (corresponding to the enterprise credit prediction rating service).
According to the financial pre-billing risk management and control system based on the artificial intelligence, which is provided by the embodiment of the application, multisource heterogeneous data is collected through an enterprise wind control data collection robot, and information standardization analysis is carried out on the multisource heterogeneous data to obtain structured data; and finally, the structured data and the credit class classification result are sent to the bill center for checking under the condition that the bill center receives a pre-opening ticket request sent by an enterprise through an enterprise credit service module, so that the automatic collection of multi-source heterogeneous data of the enterprise is realized, the manual operation flow is reduced, the financial pre-opening ticket checking efficiency is improved, and meanwhile, the comprehensiveness and the accuracy of the wind control rating are improved.
In some embodiments, enterprise wind control data collection robot 110 includes: the internal information acquisition module is used for acquiring historical transaction data from the enterprise staff stock plan ESOP system and the bill center database to obtain enterprise internal data; the external information acquisition module is used for acquiring enterprise credit information from an enterprise information base through the Internet to obtain enterprise external data; the information standardization module is used for carrying out standardization processing on quantitative information in the multi-source heterogeneous data, carrying out digital processing on qualitative information in the multi-source heterogeneous data, and counting the quantity of similar information in the multi-source heterogeneous data to obtain structured data; the enterprise information base construction module is used for constructing an enterprise information base according to the standardized data, and the enterprise information base is used for digitally managing the multi-source heterogeneous data.
In this embodiment, the information normalization module is capable of normalizing quantitative information in multi-source heterogeneous data, e.g., normalizing data by unit, order of magnitude, and presentation.
In this embodiment, the information normalization module can also digitize qualitative information related to the multi-source heterogeneous data, for example, replace characters with numbers 100, 80, 60, or 40 for characters with good, medium, bad, or very good, general, bad, etc. evaluation with respect to information such as industry information, basic quality management, etc.
In this embodiment, the information standardization module may further obtain, in a digital form, one piece of enterprise rating data corresponding to each enterprise through statistics on information such as intellectual property rights, lawsuits, upstream links and downstream links, and finally form 106-dimensional enterprise credit rating data information.
FIG. 3 is a schematic flow chart of a working mechanism of an enterprise wind control data collection robot 110 according to an embodiment of the present application, in the embodiment shown in FIG. 3, enterprise internal data is collected from databases of ESOP and bill center through an internal information collection module, enterprise external data is collected from an enterprise credit website A and an application program B through an external information collection module, and quantitative information standardization, qualitative information standardization and similar information quantity statistics and data preprocessing are respectively performed on the collected multi-source heterogeneous data through an information standardization module; after information standardization analysis is carried out on the multi-source heterogeneous data, structured data can be obtained, the structured data is multi-dimensional data, and an enterprise credit information base expressed in a standardized number can be constructed through the structured data.
In this embodiment, sample data may be provided to the training enterprise credit prediction model 120 via an enterprise credit information base.
For example, the structured data is 106-dimensional enterprise credit rating data information, and an enterprise credit information base is constructed through the 106-dimensional enterprise credit rating data information.
According to the financial pre-billing risk management and control system based on artificial intelligence, which is provided by the embodiment of the application, historical transaction data is collected from a corporate staff stock holding plan ESOP system and a bill center database through an internal information collection module, so that enterprise internal data is obtained; acquiring enterprise credit information from an enterprise information base through an external information acquisition module to obtain enterprise external data; the quantitative information in the multi-source heterogeneous data is normalized through the information normalization module, the qualitative information in the multi-source heterogeneous data is digitized, and the quantity of the same kind of information in the multi-source heterogeneous data is counted to obtain structured data, so that the efficiency of acquiring the multi-source heterogeneous data of enterprises is improved, and meanwhile, the quality and normalization of financial risk management and control are improved; the enterprise information base is constructed according to the standardized data through the enterprise information base construction module, so that financial staff can directly acquire data from the information base, manual operation flow is greatly reduced, and digital management of credit data is realized.
In some embodiments, the enterprise wind control data collection robot 110 further includes: the data preprocessing module is used for carrying out data cleaning, format conversion, data derivatization and data dimension reduction on the multi-source heterogeneous data to obtain a data tag.
In this embodiment, the enterprise wind control data collection robot 110 pre-processes the collected multi-source heterogeneous data, including data cleansing and format conversion, and selects tag data and values from the tag data as input to the enterprise credit prediction model 120.
In this embodiment, the specific content of the data preprocessing includes: and performing data cleaning, format conversion, data derivatization and data dimension reduction on the multi-source heterogeneous data to obtain a training label, namely a data label.
According to the financial pre-ticketing risk management and control system based on artificial intelligence, which is provided by the embodiment of the application, the data preprocessing module is used for carrying out data cleaning, format conversion, data derivatization and data dimension reduction on the multi-source heterogeneous data to obtain the data label, and the training label is provided for the pre-training process of the model so as to realize automatic judgment of the risk level of an enterprise, thereby improving the pre-ticketing decision efficiency of the enterprise.
In some embodiments, the enterprise credit prediction model 120 includes: the full-connection layers are used for expanding the feature matrix corresponding to the structured data to obtain an expanded feature matrix; the convolution layer is used for performing feature mapping on the feature matrix after the expansion through convolution check to obtain a mapping matrix; the pooling layer is used for carrying out sampling treatment on the mapping matrix to obtain a pooling matrix; the attention layer is used for mapping the feature vectors corresponding to the enterprise internal data and the feature vectors corresponding to the enterprise external data in the pooling matrix to a unified feature space to obtain a feature hiding matrix, and attention weight distribution information is obtained based on the feature hiding matrix and mapping parameters; the Softmax layer is used for carrying out classification probability calculation on the attention weight distribution information to obtain a credit class classification result.
In this embodiment, the enterprise credit prediction model 120 is trained based on a convolutional attention neural network algorithm, which is trained as follows:
(1) Obtaining sample data, randomly dividing a multi-source heterogeneous database of a constructed enterprise, dividing the target database into a training set and a testing set, processing each sample in the training set in a training process, and calculating a loss value corresponding to a training output result and a label classification result for a model training process; the test set is used for obtaining the output result and the label classification result to calculate the accuracy rate, and the test set does not participate in training.
In this embodiment, the convolutional attention neural network model is trained and tested by using the credit class classification result obtained in the training process, and the enterprise credit prediction model 120 is obtained after the model converges.
(2) Constructing a convolutional attention neural network model, wherein the model comprises six layers in total, and the input of the model is 106 indexes contained in an enterprise credit information base and can be expressed as vectors with the size of 1 multiplied by 106; the output is a credit grade classification result, which is a vector with the size of 1 multiplied by 5, the data in each dimension respectively represents the probability of the enterprise to predict the corresponding grade, namely the probability corresponding to the corresponding (A, B, C, D, E) risk grade, and the maximum probability can be taken as the final grade prediction result of the enterprise.
In this embodiment, the specific operational steps of the convolved attention neural network model are as follows:
the first layer of the convolution attention neural network model is a fully connected layer, expands the vector of 1×106 to 1×900, and converts the output result into a matrix of 30×30; the second layer is a convolution layer, comprises 32 convolution kernels, the size of each convolution kernel is 3 multiplied by 3, each convolution kernel performs feature mapping by sliding on an input feature matrix, and the output result is a matrix with 30 multiplied by 32, and different from the traditional feature operator, the parameters of the convolution kernels are learned through network training; the third layer is a pooling layer, and the output result is subjected to max-pooling (max-pooling) through the pooling layer, namely, the characteristic with the largest value is selected from each characteristic mapping, the result is a matrix of 15×15×32, and the input data can be subjected to 'downsampling' by using pooling, so that the number and the calculated amount of network parameters are reduced to a certain extent, and the network parameters are converted into vectors of 1×7200 (namely, a flattening operation is directly performed on the matrix of 15×15×32); the fourth layer is a full connection layer, each node is connected with all nodes of the upper layer, all the extracted characteristics are integrated, and the result is converted into a vector of 1 multiplied by 1024; the fifth layer is an attention layer, for the internal and external information of the enterprise, the probability distribution of each information is different, the generated vector features are in different feature spaces, the layer utilizes an attention mechanism to map the two information into the same feature space, and simultaneously, the most-influencing decision feature is automatically obtained
In this embodiment, the vector matrix H of each of the internal and external information is obtained through the fourth layer i And H o Splicing the two to form a characteristic hiding matrix H w ＝[H i ；H o ]The attention weight distribution is calculated by the following formula:
M w ＝tanh(w 1 H w )
wherein M is w Hiding mapping function of matrix for feature, w 1 And w 2 For mapping parameters (Projection Parameters), α w For the weight of the attention model, v et The feature vectors after the weight is re-assigned to the attention model.
In this embodiment, the sixth layer of the convolutional attention neural network model is a fully connected layer, the calculation result of the attention weight distribution is input to the Softmax layer to perform classification probability calculation, a 1×5 vector is obtained, and finally the model is converted into a probability vector with the sum of elements being 1, and the larger the element value corresponding to a certain class in the probability vector is, the higher the probability of belonging to the class is.
In this embodiment, the convolutional attention neural network model uses a Cross-Entropy cost function (Cross-Entropy Loss) as the Loss function Loss of the rating discrimination model, defines y as the target rating distribution,for predictive rating distribution, the goal of training is to minimize y and +.>Cross entropy deviation between.
The loss function loss is expressed as:
Wherein i is an entity pair subscript, j is a relationship class subscript; θ is a training parameter set, λ is a regular term, and error back propagation of the overall loss function is performed, so that the purposes of influencing the regular term and guiding network training are achieved, the generalization capability of the model can be improved to a certain extent, and overfitting is avoided.
In the convolutional attention neural network model training process, because the data size is small, in order to avoid over fitting, each neuron weight of each layer is randomly set to 0 by using probability p after each layer of the model, and the probability p is 0.55. (p is the probability of data discard), the activation function used by the first four layers is ReLU, and the activation function used by the last two layers is softmax; since the optimizer used by the overall depth scoring model is an SGD, its parameters are set as: the learning rate is 0.05, the momentum is 0.8, the learning rate taper rate is 1e-6, and the loss function is a cross entropy function; 1600 cycles of learning with a batch training size of 64 are performed to obtain a trained enterprise credit prediction model 120.
According to the financial pre-billing risk management and control system based on artificial intelligence, which is provided by the embodiment of the application, the enterprise credit prediction model is established through the convolution attention neural network, the characteristic of better interpretability of an attention mechanism is utilized, the mining of risk factors is completed through calculating the attention distribution, and the characteristic of higher contribution degree of enterprises with low grades is used as the risk factors for recommendation and display, so that the accuracy of a prediction result is improved.
The operation method of the financial pre-invoicing risk management and control system based on artificial intelligence provided by the embodiment of the application is described below, and the operation method of the financial pre-invoicing risk management and control system based on artificial intelligence described below and the financial pre-invoicing risk management and control system based on artificial intelligence described above can be correspondingly referred to each other.
Fig. 4 is a flow chart of an operation method of an artificial intelligence based financial pre-ticketing risk management and control system according to an embodiment of the present application, where the operation method of the artificial intelligence based financial pre-ticketing risk management and control system includes the following steps:
step 410, collecting multi-source heterogeneous data based on an enterprise wind control data collection robot, and carrying out qualitative analysis, quantitative analysis and data statistics on the multi-source heterogeneous data to obtain structured data; the multi-source heterogeneous data comprise enterprise internal data and enterprise external data, and the enterprise wind control data acquisition robot is built based on an intelligent flow automatic IPA development platform.
In the step, an enterprise wind control data acquisition robot is built based on an intelligent flow automatic IPA development platform.
In this embodiment, the enterprise wind control data collection robot may collect enterprise internal data, for example, the enterprise wind control data collection robot collects historical transaction data of arrears, cash flows, enterprise-related revenue sizes, asset liabilities rates, and product ordering sizes from ticketing parties in (Employee Stock Ownership Plans, corporate staff holding stock planning) systems and ticket center databases.
In this embodiment, the enterprise wind control data collection robot may collect enterprise external data, for example, by capturing enterprise information on an enterprise credit website, including data on revenue scale, business efficiency, financing status, equity change, industry ranking, upstream and downstream enterprise operations, etc., as enterprise external data.
In the embodiment, the enterprise wind control data acquisition robot can respectively perform quantitative analysis, qualitative analysis and statistics of the same type of information on the acquired multi-source heterogeneous data, and convert enterprise information in different formats into a unified format to obtain corresponding structured data.
And step 420, classifying the credit level of the structured data based on an enterprise credit prediction model to obtain a credit level classification result, wherein the enterprise credit prediction model is trained based on a convolution attention mechanism.
In this step, the enterprise credit prediction model classifies the credit class of the ticket buyer enterprise into a plurality of classes, for example, the credit class is classified into 5 classes a-E, including a (almost risk-free), B (risk-less), C (certain risk), D (risk-greater), and E (high risk), and the enterprise credit prediction model predicts the credit class of the structured data to obtain a corresponding credit class classification result.
In this embodiment, the enterprise credit prediction model may be trained based on a convolutional attention neural network algorithm.
For example, a self-attention layer is set in the convolutional neural network, and the input characteristic vectors corresponding to the enterprise internal data and the enterprise external data are mapped on a unified characteristic control, so that a more accurate credit class classification result is obtained.
Step 430, in the case that the ticket center receives the pre-opening request sent by the enterprise, the structured data and the credit class classification result are sent to the ticket center for checking.
In the embodiment, when the bill center receives the pre-ticketing application of the enterprise, the enterprise credit prediction model is automatically called to conduct level prediction on the collected multi-source heterogeneous data, and whether the pre-ticketing application of the enterprise is agreed is determined according to the level prediction.
According to the operation method of the financial pre-billing risk management and control system based on the artificial intelligence, which is provided by the embodiment of the application, multi-source heterogeneous data is collected through an enterprise wind control data collection robot, and information standardization analysis is carried out on the multi-source heterogeneous data to obtain structured data; and finally, the structured data and the credit class classification result are sent to the bill center for checking under the condition that the bill center receives a pre-opening ticket request sent by an enterprise through an enterprise credit service module, so that the automatic collection of multi-source heterogeneous data of the enterprise is realized, the manual operation flow is reduced, the financial pre-opening ticket checking efficiency is improved, and meanwhile, the comprehensiveness and the accuracy of the wind control rating are improved.
In some embodiments, the multi-source heterogeneous data is collected based on the enterprise wind control data collection robot, and qualitative analysis, quantitative analysis and data statistics are performed on the multi-source heterogeneous data to obtain structured data, including: acquiring historical transaction data from a corporate staff stock holding plan ESOP system and a bill center database based on an internal information acquisition module to obtain enterprise internal data; the external information acquisition module is used for acquiring enterprise credit information from an enterprise information base through the Internet to obtain enterprise external data; based on an information standardization module, carrying out standardization processing on quantitative information in the multi-source heterogeneous data, carrying out digital processing on qualitative information in the multi-source heterogeneous data, and counting the quantity of similar information in the multi-source heterogeneous data to obtain structured data; and constructing an enterprise information base according to the standardized data based on the enterprise information base construction module, wherein the enterprise information base is used for digitally managing the multi-source heterogeneous data.
In this embodiment, the information normalization module is capable of normalizing quantitative information in multi-source heterogeneous data, e.g., normalizing data by unit, order of magnitude, and presentation.
In this embodiment, the information normalization module can also digitize qualitative information related to the multi-source heterogeneous data, for example, replace characters with numbers 100, 80, 60, or 40 for characters with good, medium, bad, or very good, general, bad, etc. evaluation with respect to information such as industry information, basic quality management, etc.
In this embodiment, the information standardization module may further obtain, in a digital form, one piece of enterprise rating data corresponding to each enterprise through statistics on information such as intellectual property rights, lawsuits, upstream links and downstream links, and finally form 106-dimensional enterprise credit rating data information.
In this embodiment, sample data may be provided to the training enterprise credit prediction model 120 via an enterprise credit information base.
For example, the structured data is 106-dimensional enterprise credit rating data information, and an enterprise credit information base is constructed through the 106-dimensional enterprise credit rating data information.
In this embodiment, the enterprise wind control data acquisition robot performs preprocessing on the multi-source heterogeneous data, specifically includes performing processes of data cleaning, format conversion, data derivation, data dimension reduction and the like on the multi-source heterogeneous data, and selecting tag data and a value as model input from the preprocessed data.
According to the operation method of the financial pre-billing risk management and control system based on the artificial intelligence, which is provided by the embodiment of the application, the internal data of an enterprise is obtained through an internal information acquisition module; obtaining enterprise external data through an external information acquisition module; the structured data is obtained through the information standardization module, so that the efficiency of obtaining multi-source heterogeneous data of enterprises is improved, and meanwhile, the quality and standardization of financial risk management and control are improved; the enterprise information base is constructed according to the standardized data through the enterprise information base construction module, so that financial staff can directly acquire data from the information base, manual operation flow is greatly reduced, and digital management of credit data is realized.
In one embodiment, the enterprise credit prediction model includes a plurality of fully connected, convolved, pooled, attention, and Softmax layers; carrying out credit grade classification on the structured data based on the enterprise credit prediction model to obtain a credit grade classification result, wherein the method comprises the following steps: the full connection layer is used for expanding the feature matrix corresponding to the structured data to obtain an expanded feature matrix; performing feature mapping on the feature matrix after the expansion through convolution check based on the convolution layer to obtain a mapping matrix; sampling is carried out on the mapping matrix based on the pooling layer, so as to obtain a pooling matrix; mapping the feature vectors corresponding to the enterprise internal data and the feature vectors corresponding to the enterprise external data in the pooling matrix to a unified feature space based on the attention layer to obtain a feature hiding matrix, and obtaining attention weight distribution information based on the feature hiding matrix and mapping parameters; and carrying out classification probability calculation on the attention weight distribution information based on the Softmax layer to obtain a credit class classification result.
In this embodiment, the enterprise credit prediction model is trained based on a convolutional attention neural network algorithm, and the training process is as follows:
(1) Obtaining sample data, randomly dividing a multi-source heterogeneous database of a constructed enterprise, dividing the target database into a training set and a testing set, processing each sample in the training set in a training process, and calculating a loss value corresponding to a training output result and a label classification result for a model training process; the test set is used for obtaining the output result and the label classification result to calculate the accuracy rate, and the test set does not participate in training.
In the embodiment, the convolutional attention neural network model is trained and tested by using the credit class classification result obtained in the training process, and the enterprise credit prediction model is obtained after the model converges.
(2) Constructing a convolutional attention neural network model, wherein the model comprises six layers in total, and the input of the model is 106 indexes contained in an enterprise credit information base and can be expressed as vectors with the size of 1 multiplied by 106; the output is a credit grade classification result, which is a vector with the size of 1 multiplied by 5, the data in each dimension respectively represents the probability of the enterprise to predict the corresponding grade, namely the probability corresponding to the corresponding (A, B, C, D, E) risk grade, and the maximum probability can be taken as the final grade prediction result of the enterprise.
In this embodiment, the specific operational steps of the convolved attention neural network model are as follows:
the first layer of the convolution attention neural network model is a fully connected layer, expands the vector of 1×106 to 1×900, and converts the output result into a matrix of 30×30; the second layer is a convolution layer, comprises 32 convolution kernels, and has a size of 3×3, each convolution kernel performs feature mapping by sliding on an input feature matrix, and outputs a matrix with a result of 30×30×32; the third layer is a pooling layer, and the output result is subjected to maximum pooling (max-pooling) through the pooling layer; the fourth layer is a full connection layer, each node is connected with all nodes of the upper layer, all the extracted characteristics are integrated, and the result is converted into a vector of 1 multiplied by 1024; the fifth layer is an attention layer, and for the internal and external data of the enterprise, the generated vector features are in different feature spaces due to different probability distributions, and the layer utilizes an attention mechanism to map the internal and external data to the same feature space, and simultaneously automatically acquires the features which have the most influence on decision making; the sixth layer is a full-connection layer, a calculation result of attention weight distribution is input to the Softmax layer for classification probability calculation to obtain a 1×5 vector, and finally the model is converted into a probability vector with the sum of elements being 1 and output, wherein the larger the element value corresponding to a certain class in the probability vector is, the higher the probability of belonging to the class is.
In this embodiment, the convolutional attention neural network model uses a Cross-Entropy cost function (Cross-Entropy Loss) as the Loss function of the rank discrimination modelloss, defining y as the target rating distribution,for predictive rating distribution, the goal of training is to minimize y and +.>Cross entropy deviation between.
According to the operation method of the financial pre-billing risk management and control system based on artificial intelligence, which is provided by the embodiment of the application, the enterprise credit prediction model is established through the convolution attention neural network, the characteristic of better interpretability of an attention mechanism is utilized, the mining of risk factors is completed through calculating the attention distribution, the characteristic of higher contribution degree of enterprises with low grades is used as the risk factors for recommendation and display, and the accuracy of a prediction result is improved.
Fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present application, as shown in fig. 5, the electronic device may include: processor 510, communication interface (Communication Interface) 520, memory 530, and communication bus 540, wherein processor 510, communication interface 520, memory 530 complete communication with each other through communication bus 540. Processor 510 may invoke a computer program in memory 530 to perform the steps of an artificial intelligence based financial pre-billing risk management system operating method comprising: the method comprises the steps of collecting multi-source heterogeneous data based on an enterprise wind control data collection robot, and carrying out qualitative analysis, quantitative analysis and data statistics on the multi-source heterogeneous data to obtain structured data; the multi-source heterogeneous data comprises enterprise internal data and enterprise external data, and the enterprise wind control data acquisition robot is built based on an intelligent flow automatic IPA development platform; carrying out credit grade classification on the structured data based on an enterprise credit prediction model to obtain a credit grade classification result, wherein the enterprise credit prediction model is trained based on a convolution attention mechanism; and under the condition that the bill center receives a pre-ticketing request sent by an enterprise, sending the structured data and the credit class classification result to the bill center for checking.
Further, the logic instructions in the memory 530 described above may be implemented in the form of software functional units and may be stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
In another aspect, embodiments of the present application further provide a computer program product, where the computer program product includes a computer program, where the computer program may be stored on a non-transitory computer readable storage medium, where the computer program when executed by a processor is capable of executing the method for operating the financial pre-billing risk management system based on artificial intelligence provided in the foregoing embodiments, where the method includes: the method comprises the steps of collecting multi-source heterogeneous data based on an enterprise wind control data collection robot, and carrying out qualitative analysis, quantitative analysis and data statistics on the multi-source heterogeneous data to obtain structured data; the multi-source heterogeneous data comprises enterprise internal data and enterprise external data, and the enterprise wind control data acquisition robot is built based on an intelligent flow automatic IPA development platform; carrying out credit grade classification on the structured data based on an enterprise credit prediction model to obtain a credit grade classification result, wherein the enterprise credit prediction model is trained based on a convolution attention mechanism; and under the condition that the bill center receives a pre-ticketing request sent by an enterprise, sending the structured data and the credit class classification result to the bill center for checking.
In another aspect, an embodiment of the present application further provides a processor readable storage medium, where a computer program is stored in the processor readable storage medium, where the computer program is configured to cause a processor to execute the method for operating the financial pre-billing risk management system based on artificial intelligence provided in the foregoing embodiments, where the method includes: the method comprises the steps of collecting multi-source heterogeneous data based on an enterprise wind control data collection robot, and carrying out qualitative analysis, quantitative analysis and data statistics on the multi-source heterogeneous data to obtain structured data; the multi-source heterogeneous data comprises enterprise internal data and enterprise external data, and the enterprise wind control data acquisition robot is built based on an intelligent flow automatic IPA development platform; carrying out credit grade classification on the structured data based on an enterprise credit prediction model to obtain a credit grade classification result, wherein the enterprise credit prediction model is trained based on a convolution attention mechanism; and under the condition that the bill center receives a pre-ticketing request sent by an enterprise, sending the structured data and the credit class classification result to the bill center for checking.
The processor-readable storage medium may be any available medium or data storage device that can be accessed by a processor, including, but not limited to, magnetic storage (e.g., floppy disks, hard disks, magnetic tape, magneto-optical disks (MOs), etc.), optical storage (e.g., CD, DVD, BD, HVD, etc.), semiconductor storage (e.g., ROM, EPROM, EEPROM, nonvolatile storage (NAND FLASH), solid State Disk (SSD)), and the like.
The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present application, and are not limiting; although the application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims (10)
1. Financial pre-billing risk management and control system based on artificial intelligence, characterized by comprising:
the enterprise wind control data acquisition robot is built based on an intelligent process automation IPA development platform, and is used for acquiring multi-source heterogeneous data of an enterprise, and carrying out qualitative analysis, quantitative analysis and data statistics on the multi-source heterogeneous data to obtain structured data; wherein the multi-source heterogeneous data comprises enterprise internal data and enterprise external data;
the enterprise credit prediction model is trained based on a convolution attention mechanism and is used for classifying the credit level of the structured data to obtain a credit level classification result;
And the enterprise credit service module is used for sending the structured data and the credit class classification result to the bill center for checking under the condition that the bill center receives a pre-ticketing request sent by an enterprise.
2. The artificial intelligence based financial pre-ticketing risk management and control system of claim 1, wherein the enterprise wind control data collection robot comprises:
the internal information acquisition module is used for acquiring historical transaction data from a corporate staff stock holding plan ESOP system and a bill center database to obtain the corporate internal data;
the external information acquisition module is used for acquiring enterprise credit information from an enterprise information base through the Internet to obtain the enterprise external data;
the information standardization module is used for carrying out standardization processing on quantitative information in the multi-source heterogeneous data, carrying out digital processing on the qualitative information in the multi-source heterogeneous data, and counting the quantity of similar information in the multi-source heterogeneous data to obtain the structured data;
the enterprise information base construction module is used for constructing an enterprise information base according to the standardized data, and the enterprise information base is used for digitally managing multi-source heterogeneous data.
3. The artificial intelligence based financial pre-ticketing risk management and control system of claim 1, wherein the enterprise wind control data collection robot further comprises:
the data preprocessing module is used for carrying out data cleaning, format conversion, data derivatization and data dimension reduction on the multi-source heterogeneous data to obtain a data tag, wherein the data tag is a training tag associated with the enterprise credit prediction model.
4. The artificial intelligence based financial pre-ticketing risk management and control system of claim 1, wherein the enterprise credit prediction model comprises:
the full-connection layers are used for expanding the feature matrix corresponding to the structured data to obtain an expanded feature matrix;
the convolution layer is used for carrying out feature mapping on the expanded feature matrix through convolution check to obtain a mapping matrix;
the pooling layer is used for carrying out sampling treatment on the mapping matrix to obtain a pooling matrix;
the attention layer is used for mapping the feature vectors corresponding to the enterprise internal data and the feature vectors corresponding to the enterprise external data in the pooling matrix to a unified feature space to obtain a feature hiding matrix, and obtaining attention weight distribution information based on the feature hiding matrix and mapping parameters;
And the Softmax layer is used for carrying out classification probability calculation on the attention weight distribution information to obtain the credit class classification result.
5. An operation method of an artificial intelligence-based financial pre-ticketing risk management and control system is characterized by comprising the following steps:
the method comprises the steps of collecting multi-source heterogeneous data based on an enterprise wind control data collection robot, and carrying out qualitative analysis, quantitative analysis and data statistics on the multi-source heterogeneous data to obtain structured data; the multi-source heterogeneous data comprises enterprise internal data and enterprise external data, and the enterprise wind control data acquisition robot is built based on an intelligent flow automatic IPA development platform;
carrying out credit grade classification on the structured data based on an enterprise credit prediction model to obtain a credit grade classification result, wherein the enterprise credit prediction model is trained based on a convolution attention mechanism;
and under the condition that the bill center receives a pre-ticketing request sent by an enterprise, sending the structured data and the credit class classification result to the bill center for checking.
6. The method for operating an artificial intelligence based financial pre-billing risk management and control system according to claim 5, wherein the enterprise wind control data acquisition robot acquires multi-source heterogeneous data, performs qualitative analysis, quantitative analysis and data statistics on the multi-source heterogeneous data to obtain structured data, and comprises the following steps:
Acquiring historical transaction data from a corporate staff stock holding plan ESOP system and a bill center database based on an internal information acquisition module to obtain the internal data of the corporation; the external information acquisition module is used for acquiring enterprise credit information from an enterprise information base through the Internet to obtain the enterprise external data;
based on an information standardization module, carrying out standardization processing on quantitative information in the multi-source heterogeneous data, carrying out digital processing on qualitative information in the multi-source heterogeneous data, and counting the quantity of similar information in the multi-source heterogeneous data to obtain the structured data;
and constructing an enterprise information base according to the standardized data based on an enterprise information base construction module, wherein the enterprise information base is used for digitally managing the multi-source heterogeneous data.
7. The method of claim 5, wherein the enterprise credit prediction model comprises a plurality of fully connected, convolved, pooled, attention and Softmax layers;
the credit rating classification of the structured data based on the enterprise credit prediction model is performed to obtain a credit rating classification result, which comprises the following steps:
The full connection layer is used for expanding the feature matrix corresponding to the structured data to obtain an expanded feature matrix;
performing feature mapping on the extended feature matrix through convolution check based on a convolution layer to obtain a mapping matrix;
sampling the mapping matrix based on a pooling layer to obtain a pooling matrix;
mapping the feature vectors corresponding to the enterprise internal data and the feature vectors corresponding to the enterprise external data in the pooling matrix to a unified feature space based on the attention layer to obtain a feature hiding matrix, and obtaining attention weight distribution information based on the feature hiding matrix and mapping parameters;
and carrying out classification probability calculation on the attention weight distribution information based on a Softmax layer to obtain the credit class classification result.
8. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor, when executing the program, implements a method of operating the financial pre-invoicing risk management system of any of claims 5 to 7 for artificial intelligence.
9. A non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements a method of operating the financial pre-invoicing risk management system of any of claims 5 to 7 for artificial intelligence.
10. A computer program product comprising a computer program which when executed by a processor implements a method of operating an artificial intelligence financial pre-invoicing risk management system according to any of claims 5 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311113547.8A CN117217522A (en) | 2023-08-30 | 2023-08-30 | Financial pre-billing risk management and control system based on artificial intelligence and operation method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311113547.8A CN117217522A (en) | 2023-08-30 | 2023-08-30 | Financial pre-billing risk management and control system based on artificial intelligence and operation method thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117217522A true CN117217522A (en) | 2023-12-12 |
Family
ID=89043448
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311113547.8A Pending CN117217522A (en) | 2023-08-30 | 2023-08-30 | Financial pre-billing risk management and control system based on artificial intelligence and operation method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117217522A (en) |
-
2023
- 2023-08-30 CN CN202311113547.8A patent/CN117217522A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | Risk prediction in financial management of listed companies based on optimized BP neural network under digital economy | |
| Chung et al. | Insolvency prediction model using multivariate discriminant analysis and artificial neural network for the finance industry in New Zealand | |
| CN110956273A (en) | Credit scoring method and system integrating multiple machine learning models | |
| CN111861698B (en) | Pre-loan approval early warning method and system based on loan multi-head data | |
| KR102330423B1 (en) | Online default forecasting system using image recognition deep learning algorithm | |
| CN112700324A (en) | User loan default prediction method based on combination of Catboost and restricted Boltzmann machine | |
| CN109255029A (en) | A method of automatic Bug report distribution is enhanced using weighted optimization training set | |
| CN113283582A (en) | Textile industry financial loss prediction method and device, storage medium and processor | |
| CN117217522A (en) | Financial pre-billing risk management and control system based on artificial intelligence and operation method thereof | |
| Yang et al. | An evidential reasoning rule-based ensemble learning approach for evaluating credit risks with customer heterogeneity | |
| CN114581209A (en) | Method, device and equipment for training financial analysis model and storage medium | |
| CN107977804B (en) | Guarantee warehouse business risk assessment method | |
| Gaikwad et al. | Unveiling Market Dynamics through Machine Learning: Strategic Insights and Analysis | |
| Zeng | A comparison study on the era of internet finance China construction of credit scoring system model | |
| CN113327162B (en) | Credit wind control rule making system | |
| KR102596740B1 (en) | Method for predicting macroeconomic factors and stock returns in the context of economic uncertainty news sentiment using machine learning | |
| Kocaoğlu et al. | Sector-Based Stock Price Prediction with Machine Learning Models | |
| Wang | Data-driven Investment Decisions in P2P Lending: Strategies of Integrating Credit Scoring and Profit Scoring | |
| SANTOSH et al. | ANALYZING LARGE SCALE BUSINESS DATA TO PREDICT COMPANY'S GROWTH USING AN INTEGRATED HYBRID APPROACH OF DATA REDUCTION UNIT AND CONVOLUTIONAL NEURAL NETWORKS | |
| Ashok Bhavikatti | Credit Risk Modelling | |
| Khiem Tran et al. | Towards Improved Bankruptcy Prediction: Utilizing Variational Autoencoder Latent Representations in a Norwegian Context | |
| Fernández | Co-teaching strategies for credit risk assessment in the presence of label noise | |
| CN115237970A (en) | Data prediction method, device, equipment, storage medium and program product | |
| CN117764692A (en) | Method for predicting credit risk default probability | |
| Pang et al. | Credit risk prediction based on an interpretable three-way decision method: Evidence from Chinese SMEs |
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 |