Financial statement intelligent generation method, device and equipment and readable storage medium
By using the frequency prediction model in the financial statement generation system, the supervision frequency is predicted based on the report supervision characteristic data and the generation time is adjusted, the problem of inflexible report generation frequency in the existing technology is solved, and more personalized and efficient report generation is achieved.
Patent Information
- Application Number
- CN202411538842.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-31
- Publication Date
- 2025-02-07
AI Technical Summary
The existing financial statement generation methods are less flexible and cannot dynamically adjust the generation frequency according to the specific needs of regulators.
By entering the report supervision feature data into the pre-trained frequency prediction model, predict the supervision frequency, and automatically adjust the report generation time based on the prediction results.
It improves the flexibility and personalization of financial statement generation, can more accurately meet the needs of regulators and reduce human intervention.
Smart Images
Figure CN119398891A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method, device, equipment and readable storage medium for intelligently generating financial statements. Background Art
[0002] The fund supervision business data generates corresponding reports that can reflect the business situation according to the manually set frequency and business rules, and provide them to the regulators for review and analysis.
[0003] Currently, reports are generated based on a manually set frequency. Generally, daily reports are generated every other day and monthly reports are generated every other month. The flexibility of report generation is low. Summary of the invention
[0004] The present application provides a method, apparatus, device and readable storage medium for intelligently generating financial statements, as follows:
[0005] A method for intelligently generating financial statements, comprising:
[0006] Inputting the report supervision characteristic data into the pre-trained frequency prediction model to obtain the predicted supervision frequency output by the frequency prediction model, wherein the report supervision characteristic data at least includes report supervision behavior data, and the report supervision behavior data includes the supervisor information and the historical report viewing frequency;
[0007] Based on the predicted supervision frequency, obtaining a report generation time;
[0008] In response to reaching the report generation time, a financial report is generated.
[0009] Optionally, the training process of the frequency prediction model includes:
[0010] Acquire first training data, wherein the first training data includes historical report supervision feature data with historical supervision frequency as a label, and the historical report supervision feature data includes historical report supervision behavior data;
[0011] Performing data preprocessing, feature engineering, and data segmentation on the first training data to obtain a first training set and a first validation set;
[0012] The first neural network model is trained based on the first training set, and the trained first neural network model is evaluated based on the first validation set until a first neural network model that meets the training completion condition is obtained as the frequency prediction model.
[0013] Optionally, after generating the financial statement, the method further comprises:
[0014] Extracting risk characteristic data from the financial statement, the risk characteristic data including values of a plurality of preset risk assessment characteristic items;
[0015] The risk characteristic data is input into a pre-trained risk prediction model to obtain a risk prediction result output by the risk prediction model.
[0016] Optionally, the training process of the risk prediction model includes:
[0017] Acquire second training data, wherein the second training data includes historical risk feature data with historical risk metrics as labels, and the historical risk feature data includes historical values of multiple preset risk assessment feature items;
[0018] Performing data preprocessing, feature engineering, and data segmentation on the second training data to obtain a second training set and a second validation set;
[0019] The second neural network model is trained based on the second training set, and the trained second neural network model is evaluated based on the second validation set until a second neural network model that meets the training completion condition is obtained as the risk prediction model.
[0020] Optionally, the report supervision characteristic data also includes: risk prediction characteristic data, the risk prediction characteristic data includes risk prediction results obtained within a preset time period; the historical report supervision characteristic data also includes historical risk prediction characteristic data.
[0021] Optionally, the method further comprises:
[0022] In response to the risk prediction result indicating that the risk metric value is higher than a preset risk threshold, a risk financial report is generated and a risk warning instruction is issued.
[0023] Optionally, based on the predicted supervision frequency, obtaining the report generation time includes:
[0024] Get the buffer duration based on at least the forecast duration of report generation;
[0025] The report generation time is obtained based on the predicted supervision frequency, the historical generation time of the financial report and the buffer time.
[0026] A financial report intelligent generation device, comprising:
[0027] A frequency prediction unit, used for inputting report supervision characteristic data into a pre-trained frequency prediction model to obtain a predicted supervision frequency output by the frequency prediction model, wherein the report supervision characteristic data at least includes report supervision behavior data, and the report supervision behavior data includes supervisor information and historical report viewing frequency;
[0028] A time prediction unit, used for obtaining a report generation time based on the predicted supervision frequency;
[0029] The report generation unit is used to generate a financial report in response to reaching the report generation time.
[0030] A financial report intelligent generation device, comprising: a memory and a processor;
[0031] The memory is used to store programs;
[0032] The processor is used to execute the program to implement each step of the method for intelligently generating financial statements.
[0033] A readable storage medium stores a computer program, which, when executed by a processor, implements each step of a method for intelligently generating financial statements.
[0034] It can be seen from the above technical solutions that the intelligent generation method, device, equipment and readable storage medium of financial statements provided by the embodiments of the present application input the report supervision characteristic data into the pre-trained frequency prediction model to obtain the predicted supervision frequency output by the frequency prediction model, and the report supervision characteristic data at least includes report supervision behavior data, and the report supervision behavior data includes the supervisory party information and the historical report viewing frequency. Based on the predicted supervision frequency, the report generation time is obtained. In response to reaching the report generation time, a financial statement is generated. It can be seen that this method predicts the frequency of the supervisory party viewing the financial statement based on the report supervision characteristic data through a pre-constructed frequency prediction model, thereby obtaining the time for the next automatic generation of the financial statement, and automatically generates the financial statement when the time is reached, thereby improving the flexibility and personalization of the generation of financial statements. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0036] Figure 1 A flowchart of a specific implementation method of a financial statement intelligent generation method provided in an embodiment of the present application;
[0037] Figure 2 A flowchart of a method for intelligently generating financial statements provided in an embodiment of the present application;
[0038] Figure 3A schematic diagram of the structure of a financial report intelligent generation device provided in an embodiment of the present application;
[0039] Figure 4 A schematic diagram of the structure of a financial report intelligent generation device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0040] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0041] The report data of fund supervision business contains a large amount of financial and business information, which can provide a lot of valuable information to data users. For example, by analyzing the supervision report data, abnormal transaction behaviors can be detected, such as abnormal account activities, unusual transaction amounts and frequencies, etc. These abnormal behaviors may be indicators of credit risk, market risk and operational risk. Banks can take corresponding regulatory measures in a timely manner based on these indicators; not only that, the fund supervision report data can also help banks assess the credit risk of customers. By analyzing indicators such as customer account balances, transaction frequencies and financial status, customers' credit status and risk preferences can be assessed, and corresponding measures can be taken to manage and reduce customer risks.
[0042] Statement analysis is generally done manually to identify abnormal fluctuations and trends in statements. On this basis, this paper proposes a statement risk prediction and early warning method based on artificial intelligence, which conducts intelligent analysis on the statement data of fund supervision business, establishes a risk early warning model, automatically identifies business risks, and issues risk warnings according to preset early warning rules.
[0043] This paper proposes an automatic report generation method based on risk prediction. Report data is generally provided to relevant data users for monitoring and evaluating the overall business. Therefore, it is necessary to intelligently identify data risks in reports, and automatically adjust the report generation frequency in combination with user supervision demand prediction, provide it to customers, and intervene in supervision in a timely manner. When there are abnormal fluctuations in report data, there is a certain lag in risk identification and supervision.
[0044] Figure 1 A flowchart of a method for intelligently generating financial statements provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, this method specifically includes:
[0045] S101. Inputting report supervision feature data into a pre-trained frequency prediction model to obtain a predicted supervision frequency output by the frequency prediction model.
[0046] In this embodiment, the report supervision characteristic data at least includes report supervision behavior data, and the report supervision behavior data includes the regulator information and the historical report viewing frequency, wherein the regulator information includes the regulator attributes (such as enterprise attributes, data scale, and fund security requirement level, etc.), the regulator event information (such as big promotion activities, enterprise activities at specific times, etc.), and the historical report viewing frequency indicates the frequency of the regulator viewing the financial statements, for example, triggering the financial report viewing operation once a month.
[0047] In this embodiment, the frequency prediction model is a neural network model trained with historical report supervision feature data as input data and historical supervision frequency as target output. An optional frequency prediction model method includes:
[0048] A1. Obtain first training data.
[0049] In this embodiment, the first training data includes historical report supervision feature data with historical supervision frequency as a label, and the historical report supervision feature data includes historical report supervision behavior data.
[0050] A2. Perform data preprocessing, feature engineering, and data segmentation on the first training data to obtain a first training set and a first validation set.
[0051] A3. Perform model training on the first neural network model based on the first training set, and perform model evaluation on the trained first neural network model based on the first validation set, until a first neural network model that meets the training completion condition is obtained as a frequency prediction model.
[0052] Optionally, the first neural network model may be a graph neural network, and the training completion condition includes that the number of iterations reaches a preset number threshold and / or the model accuracy reaches a preset accuracy threshold. It should be noted that the specific training method of the graph neural network refers to the prior art.
[0053] S102: Obtain a buffering time based at least on the predicted time of report generation.
[0054] In this embodiment, the predicted report generation time is used to indicate the average time required to generate a financial report. Optionally, the predicted report generation time is obtained based on statistics of the time required to generate historical reports.
[0055] S103. Obtaining report generation time based on the predicted supervision frequency, the historical generation time of financial reports and the buffer time.
[0056] In this embodiment, the historical generation time includes the time when the financial report was most recently generated. Optionally, a predicted viewing time is generated based on the predicted supervision frequency and the historical generation time, and the report generation time is obtained by subtracting the buffer time from the predicted viewing time.
[0057] S104: In response to reaching the report generation time, generate a financial report.
[0058] In this embodiment, the specific method of generating financial statements can refer to the prior art.
[0059] S105. Extract risk characteristic data from financial statements.
[0060] In this embodiment, the risk characteristic data includes the values of a plurality of preset risk assessment characteristic items. The risk assessment characteristic items include but are not limited to transaction amount, transaction frequency, and transaction object.
[0061] S106. Input the risk feature data into a pre-trained risk prediction model to obtain a risk prediction result output by the risk prediction model.
[0062] In this embodiment, the risk prediction model is trained with historical risk feature data as input and historical risk measurement values as target output. The historical risk feature data includes historical values of multiple preset risk assessment feature items, and the historical risk measurement value is the probability value (0 or 1) of whether the risk occurs when the historical risk feature data is available.
[0063] An optional risk prediction model training method includes:
[0064] B1. Obtain second training data.
[0065] In this embodiment, the second training data includes historical risk feature data with historical risk measurement values as labels, and the historical risk feature data includes historical values of multiple preset risk assessment feature items. Optionally, the historical risk measurement value is obtained based on the impact degree of the risk event. When the impact degree is greater than the degree threshold, the historical risk measurement value is recorded as 1, and when the impact degree is not greater than the degree threshold, the historical risk measurement value is recorded as 0.
[0066] A2. Perform data preprocessing, feature engineering, and data segmentation on the second training data to obtain a second training set and a second validation set.
[0067] An optional method for preprocessing, feature engineering, and data segmentation of training data includes:
[0068] Data preprocessing: First, the second training data needs to be cleaned, including preprocessing processes such as data cleaning, deduplication, and missing value processing, so that the data set can be analyzed and modeled.
[0069] Feature engineering: According to regulatory needs, select regulatory indicators that meet the requirements and determine important features that affect risks, such as year-on-year growth rate, month-on-month growth, abnormal fluctuation threshold, etc. Perform feature processing on the selected feature data, such as feature scaling and fast weighting.
[0070] Data segmentation: The second training data after data preprocessing and feature engineering is divided into training set and test set according to a certain ratio, usually 80:20 or 70:30.
[0071] A3. Perform model training on the second neural network model based on the second training set, and perform model evaluation on the trained second neural network model based on the second verification set, until a second neural network model that meets the training completion conditions is obtained as a risk prediction model.
[0072] In this embodiment, in the process of constructing the risk prediction model, the second neural network model is trained based on the second training set. In the process of training the model, supervised learning or unsupervised learning algorithms and cross-validation methods can be used to improve the accuracy of the model. Further, the trained second neural network model is evaluated based on the second validation set to determine the effect and error of the model. Commonly used methods include model evaluation indicators such as ROC curve analysis and F1 score. The training completion conditions include but are not limited to reaching the number of iterations or the model error is less than the preset error threshold.
[0073] It should be noted that the risk prediction results output by the risk prediction model include risk measurement values (0 to 1).
[0074] S107: In response to the risk prediction result indicating that the risk metric value is higher than a preset risk threshold, a risk financial report is generated and a risk warning instruction is issued.
[0075] In this embodiment, the risk prediction results are monitored. When the risk prediction results trigger the warning threshold, that is, the risk measurement value is higher than the preset risk threshold, a response warning behavior is performed according to the warning rules, where the warning behavior includes but is not limited to issuing a risk warning instruction.
[0076] It should be noted that a financial statement marked with a risk indicator is called a risk financial statement, which is used to indicate that the financial statement is generated when the risk prediction result triggers the warning threshold. The method for generating a risk financial statement is the same as that for generating a financial statement.
[0077] It can be seen from the above technical solutions that the embodiment of the present application provides a method for intelligently generating financial statements, which predicts the frequency of the regulator's review of financial statements based on the report supervision feature data through a pre-built frequency prediction model (graph neural network model), thereby obtaining the time for the next automatic generation of financial statements, and automatically generating financial statements when the time is reached, thereby improving the flexibility and personalization of financial statement generation. The risk prediction results are monitored in real time through a pre-built risk prediction model, and when the risk prediction results trigger a risk, financial statements are automatically generated and risk warnings are issued. The timeliness of automatic report generation is further achieved, and risk warnings are provided to improve regulatory security.
[0078] It should be noted that the above S101 to S107 are an optional specific implementation process of the present application, and the intelligent generation method of financial statements provided by the present application also includes other specific implementation processes.
[0079] For example, the report supervision characteristic data also includes: risk prediction characteristic data, the risk prediction characteristic data includes the risk prediction results obtained within a preset time period, and the historical report supervision characteristic data also includes historical risk prediction characteristic data. That is, the risk prediction results are added to the characteristic data used to train the frequency prediction model to improve the accuracy of the frequency prediction model in predicting the supervision frequency.
[0080] For another example, S103 to S104 are only an optional method for obtaining the report generation time. In other optional embodiments, the buffering time may be adjusted based on data such as network status.
[0081] In summary, the financial statement intelligent generation method provided by this application is summarized as follows: Figure 2 The process shown is as follows Figure 2 As shown, the method includes:
[0082] S201. Inputting the report supervision feature data into a pre-trained frequency prediction model to obtain a predicted supervision frequency output by the frequency prediction model.
[0083] In this embodiment, the report supervision characteristic data at least includes report supervision behavior data, and the report supervision behavior data includes supervisor information and historical report viewing frequency.
[0084] S202. Based on the predicted supervision frequency, obtain the report generation time.
[0085] S203: In response to reaching the report generation time, generate a financial report.
[0086] It can be seen from the above technical solution that an intelligent method for generating financial statements provided in an embodiment of the present application predicts the frequency of regulators reviewing financial statements based on report supervision characteristic data through a pre-constructed frequency prediction model, thereby obtaining the time for the next automatic generation of financial statements, and automatically generating financial statements when the time is reached, thereby improving the flexibility and personalization of the generation of financial statements.
[0087] Figure 3 A schematic diagram of the structure of a financial report intelligent generation device provided in an embodiment of the present application is shown. Figure 3 As shown, the device may include:
[0088] The frequency prediction unit 301 is used to input the report supervision characteristic data into the pre-trained frequency prediction model to obtain the predicted supervision frequency output by the frequency prediction model, wherein the report supervision characteristic data at least includes report supervision behavior data, and the report supervision behavior data includes the supervisor information and the historical report viewing frequency;
[0089] A time prediction unit 302, configured to obtain a report generation time based on the predicted supervision frequency;
[0090] The report generation unit 303 is used to generate a financial report in response to reaching the report generation time.
[0091] Optionally, the device also includes a frequency prediction model construction unit, which is used to obtain first training data, wherein the first training data includes historical report supervision feature data labeled with historical supervision frequency, and the historical report supervision feature data includes historical report supervision behavior data; the first training data is subjected to data preprocessing, feature engineering, and data segmentation processing to obtain a first training set and a first verification set; a first neural network model is trained based on the first training set, and a model evaluation is performed on the trained first neural network model based on the first verification set, until a first neural network model that meets the training completion conditions is obtained as the frequency prediction model.
[0092] Optionally, the device also includes a risk prediction unit, which is used to extract risk characteristic data from the financial statement after the financial statement is generated, and the risk characteristic data includes values of multiple preset risk assessment characteristic items; input the risk characteristic data into a pre-trained risk prediction model to obtain a risk prediction result output by the risk prediction model.
[0093] Optionally, the device also includes a risk prediction model construction unit, which is used to obtain second training data, wherein the second training data includes historical risk feature data with historical risk measurement values as labels, and the historical risk feature data includes historical values of multiple preset risk assessment feature items; the second training data is subjected to data preprocessing, feature engineering, and data segmentation processing to obtain a second training set and a second verification set; a second neural network model is trained based on the second training set, and a model evaluation is performed on the trained second neural network model based on the second verification set until a second neural network model that meets the training completion conditions is obtained as the risk prediction model.
[0094] Optionally, the report supervision characteristic data also includes: risk prediction characteristic data, the risk prediction characteristic data includes risk prediction results obtained within a preset time period; the historical report supervision characteristic data also includes historical risk prediction characteristic data.
[0095] Optionally, the device further comprises a risk warning unit for generating a risk financial report and issuing a risk warning instruction in response to the risk prediction result indicating that the risk metric value is higher than a preset risk threshold.
[0096] Optionally, the time prediction unit is used to obtain the report generation time based on the predicted regulatory frequency, including: the time prediction unit is specifically used to: obtain the buffer time based at least on the predicted report generation duration; obtain the report generation time based on the predicted regulatory frequency, the historical generation time of the financial statement and the buffer time.
[0097] Figure 4 A schematic diagram of the structure of the financial statement intelligent generation device is shown, and the device may include: at least one processor 401, at least one communication interface 402, at least one memory 403 and at least one communication bus 404;
[0098] In the embodiment of the present application, the number of the processor 401, the communication interface 402, the memory 403, and the communication bus 404 is at least one, and the processor 401, the communication interface 402, and the memory 403 communicate with each other through the communication bus 404;
[0099] The processor 401 may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention;
[0100] The memory 403 may include a high-speed RAM memory, and may also include a non-volatile memory, such as at least one disk memory;
[0101] The memory stores a program, and the processor can execute the program stored in the memory to implement the steps of a financial statement intelligent generation method provided in an embodiment of the present application, as follows:
[0102] A method for intelligently generating financial statements, comprising:
[0103] Inputting the report supervision characteristic data into the pre-trained frequency prediction model to obtain the predicted supervision frequency output by the frequency prediction model, wherein the report supervision characteristic data at least includes report supervision behavior data, and the report supervision behavior data includes the supervisor information and the historical report viewing frequency;
[0104] Based on the predicted supervision frequency, obtaining a report generation time;
[0105] In response to reaching the report generation time, a financial report is generated.
[0106] Optionally, the training process of the frequency prediction model includes:
[0107] Acquire first training data, the first training data including historical report supervision feature data with historical supervision frequency as a label, the historical report supervision feature data including historical report supervision behavior data; perform data preprocessing, feature engineering, and data segmentation on the first training data to obtain a first training set and a first validation set; perform model training on a first neural network model based on the first training set, and perform model evaluation on the trained first neural network model based on the first validation set, until a first neural network model that meets the training completion conditions is obtained as the frequency prediction model.
[0108] Optionally, after generating the financial statement, the method further comprises:
[0109] Extract risk characteristic data from the financial statements, the risk characteristic data including values of a plurality of preset risk assessment characteristic items; input the risk characteristic data into a pre-trained risk prediction model to obtain a risk prediction result output by the risk prediction model.
[0110] Optionally, the training process of the risk prediction model includes:
[0111] Acquire second training data, wherein the second training data includes historical risk feature data labeled with historical risk measurement values, wherein the historical risk feature data includes historical values of multiple preset risk assessment feature items; perform data preprocessing, feature engineering, and data segmentation on the second training data to obtain a second training set and a second validation set; perform model training on a second neural network model based on the second training set, and perform model evaluation on the trained second neural network model based on the second validation set, until a second neural network model that meets the training completion conditions is obtained as the risk prediction model.
[0112] Optionally, the report supervision characteristic data also includes: risk prediction characteristic data, the risk prediction characteristic data includes risk prediction results obtained within a preset time period; the historical report supervision characteristic data also includes historical risk prediction characteristic data.
[0113] Optionally, the method further comprises:
[0114] In response to the risk prediction result indicating that the risk metric value is higher than a preset risk threshold, a risk financial report is generated and a risk warning instruction is issued.
[0115] Optionally, based on the predicted supervision frequency, obtaining the report generation time includes:
[0116] The buffer time is obtained based at least on the predicted report generation time; the report generation time is obtained based on the predicted supervision frequency, the historical generation time of the financial report and the buffer time.
[0117] It should be noted that the specific implementation methods of the various steps of the method for intelligently generating financial statements provided in the embodiment of the present application refer to the above embodiment and will not be repeated here.
[0118] The embodiment of the present application further provides a readable storage medium, which can store a computer program suitable for execution by a processor. When the computer program is executed by the processor, the steps of a financial statement intelligent generation method provided in the embodiment of the present application are implemented as follows:
[0119] A method for intelligently generating financial statements, comprising:
[0120] Inputting the report supervision characteristic data into the pre-trained frequency prediction model to obtain the predicted supervision frequency output by the frequency prediction model, wherein the report supervision characteristic data at least includes report supervision behavior data, and the report supervision behavior data includes the supervisor information and the historical report viewing frequency;
[0121] Based on the predicted supervision frequency, obtaining a report generation time;
[0122] In response to reaching the report generation time, a financial report is generated.
[0123] Optionally, the training process of the frequency prediction model includes:
[0124] Acquire first training data, the first training data including historical report supervision feature data with historical supervision frequency as a label, the historical report supervision feature data including historical report supervision behavior data; perform data preprocessing, feature engineering, and data segmentation on the first training data to obtain a first training set and a first validation set; perform model training on a first neural network model based on the first training set, and perform model evaluation on the trained first neural network model based on the first validation set, until a first neural network model that meets the training completion conditions is obtained as the frequency prediction model.
[0125] Optionally, after generating the financial statement, the method further comprises:
[0126] Extract risk characteristic data from the financial statements, the risk characteristic data including values of a plurality of preset risk assessment characteristic items; input the risk characteristic data into a pre-trained risk prediction model to obtain a risk prediction result output by the risk prediction model.
[0127] Optionally, the training process of the risk prediction model includes:
[0128] Acquire second training data, wherein the second training data includes historical risk feature data labeled with historical risk measurement values, wherein the historical risk feature data includes historical values of multiple preset risk assessment feature items; perform data preprocessing, feature engineering, and data segmentation on the second training data to obtain a second training set and a second validation set; perform model training on a second neural network model based on the second training set, and perform model evaluation on the trained second neural network model based on the second validation set, until a second neural network model that meets the training completion conditions is obtained as the risk prediction model.
[0129] Optionally, the report supervision characteristic data also includes: risk prediction characteristic data, the risk prediction characteristic data includes risk prediction results obtained within a preset time period; the historical report supervision characteristic data also includes historical risk prediction characteristic data.
[0130] Optionally, the method further comprises:
[0131] In response to the risk prediction result indicating that the risk metric value is higher than a preset risk threshold, a risk financial report is generated and a risk warning instruction is issued.
[0132] Optionally, based on the predicted supervision frequency, obtaining the report generation time includes:
[0133] The buffer time is obtained based at least on the predicted report generation time; the report generation time is obtained based on the predicted supervision frequency, the historical generation time of the financial report and the buffer time.
[0134] It should be noted that the method, device, equipment and readable storage medium for intelligent generation of financial statements provided by the present invention can be used in the financial field or other fields. It can be used in the field of artificial intelligence or the financial field. The above is only an example and does not limit the application field of the method, device, equipment and readable storage medium for intelligent generation of financial statements provided by the present invention.
[0135] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.
[0136] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.
[0137] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for intelligently generating financial statements, characterized in that: include: Inputting the report supervision characteristic data into the pre-trained frequency prediction model to obtain the predicted supervision frequency output by the frequency prediction model, wherein the report supervision characteristic data at least includes report supervision behavior data, and the report supervision behavior data includes the supervisor information and the historical report viewing frequency; Based on the predicted supervision frequency, obtaining a report generation time; In response to reaching the report generation time, a financial report is generated.
2. The method according to claim 1, characterized in that The training process of the frequency prediction model includes: Acquire first training data, wherein the first training data includes historical report supervision feature data with historical supervision frequency as a label, and the historical report supervision feature data includes historical report supervision behavior data; Performing data preprocessing, feature engineering, and data segmentation on the first training data to obtain a first training set and a first validation set; The first neural network model is trained based on the first training set, and the trained first neural network model is evaluated based on the first validation set until a first neural network model that meets the training completion condition is obtained as the frequency prediction model.
3. The method according to claim 2, characterized in that The method further comprises, after generating the financial statement: Extracting risk characteristic data from the financial statement, the risk characteristic data including values of a plurality of preset risk assessment characteristic items; The risk characteristic data is input into a pre-trained risk prediction model to obtain a risk prediction result output by the risk prediction model.
4. The method according to claim 3, characterized in that The training process of the risk prediction model includes: Acquire second training data, wherein the second training data includes historical risk feature data with historical risk metrics as labels, and the historical risk feature data includes historical values of multiple preset risk assessment feature items; Performing data preprocessing, feature engineering, and data segmentation on the second training data to obtain a second training set and a second validation set; The second neural network model is trained based on the second training set, and the trained second neural network model is evaluated based on the second validation set until a second neural network model that meets the training completion condition is obtained as the risk prediction model.
5. The method according to claim 3, characterized in that: The report supervision characteristic data also includes: risk prediction characteristic data, and the risk prediction characteristic data includes risk prediction results obtained within a preset time period; the historical report supervision characteristic data also includes historical risk prediction characteristic data.
6. The method according to claim 3, characterized in that The method further comprises: In response to the risk prediction result indicating that the risk metric value is higher than a preset risk threshold, a risk financial report is generated and a risk warning instruction is issued.
7. The method according to any one of claims 1 to 6, characterized in that: The obtaining of report generation time based on the predicted supervision frequency includes: Get the buffer duration based on at least the forecast duration of report generation; The report generation time is obtained based on the predicted supervision frequency, the historical generation time of the financial report and the buffer time.
8. A financial report intelligent generation device, characterized in that: include: A frequency prediction unit, used for inputting report supervision characteristic data into a pre-trained frequency prediction model to obtain a predicted supervision frequency output by the frequency prediction model, wherein the report supervision characteristic data at least includes report supervision behavior data, and the report supervision behavior data includes supervisor information and historical report viewing frequency; A time prediction unit, used for obtaining a report generation time based on the predicted supervision frequency; The report generation unit is used to generate a financial report in response to reaching the report generation time.
9. A financial report intelligent generation device, characterized in that: include: Memory and processor; The memory is used to store programs; The processor is used to execute the program to implement each step of the method for intelligently generating financial statements according to any one of claims 1 to 7.
10. A readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, each step of the method for intelligently generating financial statements according to any one of claims 1 to 7 is implemented.