Financial statement intelligent generation method and device, computer equipment and storage medium

Abstract

The invention relates to the technical field of big data analysis, and discloses a financial statement intelligent generation method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring various data required by a financial statement to obtain initial data; preprocessing the initial data to obtain a preprocessing result; inputting the preprocessing result into an automatic report generation model to generate a report so as to obtain a generation result; and outputting the generation result. Wherein the automatic report generation model is obtained by training a deep learning model by taking a plurality of data required by a financial report as a sample set after preprocessing. By implementing the method, the generation speed of the financial statement can be increased, human intervention is reduced, and the data accuracy and reliability are improved.

Description

Technical Field

[0001] The present invention relates to big data analysis technology, and more specifically to a method, device, computer equipment and storage medium for intelligently generating financial statements. Background Art

[0002] In the daily financial operations of a company, it is crucial to prepare and present various financial statements to ensure the accuracy and timeliness of information. However, due to the wide range of business operations of a company, the required financial accounting data sources and systems are often scattered, which brings many challenges to the preparation of financial statements.

[0003] In addition, each business department usually uses independent systems for data management and storage. The lack of effective connectivity and standardization between these systems makes data integration extremely difficult. The current report generation method fails to effectively integrate data, resulting in poor information communication between business departments and the inability to share key data. In addition, similar data in different systems may be different, affecting the accuracy of subsequent analysis and decision-making; financial personnel need to manually download data from multiple systems and perform a large amount of manual operations and data integration. This process is not only cumbersome, but also prone to errors; manual operations consume a lot of time, resulting in low efficiency in report generation; further, due to the complexity of manual operations, errors are prone to occur during data entry and processing, further affecting the accuracy and reliability of data. Moreover, the manual processing of report generation is slow, especially when it is necessary to respond quickly to market changes, it is difficult to meet the needs of real-time decision-making. Due to the long report generation time, management cannot obtain important financial information in a timely manner, affecting the timeliness of business decisions. In a highly competitive market environment, delays in financial data may result in missed opportunities or increased operating risks.

[0004] Therefore, it is necessary to design a new method to increase the speed of financial statement generation, reduce human intervention, and improve data accuracy and reliability. Summary of the invention

[0005] The present invention provides a method, apparatus, computer equipment and storage medium for intelligently generating financial statements to solve the technical problem that due to the scattered sources of financial data and the lack of effective connectivity and standardization of the system, information integration is difficult, inefficient and prone to errors in the report generation process, thereby affecting the accuracy and timeliness of financial information and failing to meet the needs of rapid decision-making.

[0006] In the first aspect, a method for intelligently generating financial statements is provided, including:

[0007] Obtain various data required for financial statements to obtain initial data;

[0008] Preprocessing the initial data to obtain a preprocessing result;

[0009] Inputting the preprocessing result into the report automatic generation model to generate the report, so as to obtain the generation result;

[0010] Outputting the generated result;

[0011] Among them, the automatic report generation model is obtained by training the deep learning model with a sample set after preprocessing various data required for the financial report.

[0012] In a second aspect, a financial statement intelligent generation device is provided, comprising:

[0013] An initial data acquisition unit is used to acquire various data required for financial statements to obtain initial data;

[0014] A preprocessing unit, used for preprocessing the initial data to obtain a preprocessing result;

[0015] An automatic generation unit, used for inputting the preprocessing result into a report automatic generation model to generate a report, so as to obtain a generation result;

[0016] An output unit, used for outputting the generated result;

[0017] Among them, the automatic report generation model is obtained by training the deep learning model with a sample set after preprocessing various data required for the financial report.

[0018] In a third aspect, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned method for intelligently generating financial statements when executing the computer program.

[0019] In a fourth aspect, a storage medium is provided, which stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned method for intelligently generating financial statements are implemented.

[0020] The above-mentioned intelligent financial statement generation method obtains the required multiple data and preprocesses them, then inputs the processed data into the deep learning model to generate the report, and finally outputs the results. Through automated data acquisition and processing, the speed of report generation is significantly improved, and the need for manual intervention is reduced. This method improves the accuracy and reliability of data through precise preprocessing and deep learning model training, thereby enhancing the quality of the report.

[0021] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying any creative work.

[0023] Figure 1 A schematic diagram of an application scenario of the method for intelligently generating financial statements provided by an embodiment of the present invention;

[0024] Figure 2 A schematic diagram of a process for intelligently generating financial statements provided by an embodiment of the present invention;

[0025] Figure 3 A schematic diagram of a sub-process of a method for intelligently generating financial statements provided by an embodiment of the present invention;

[0026] Figure 4 A schematic block diagram of a device for intelligently generating financial statements provided by an embodiment of the present invention;

[0027] Figure 5 It is a schematic diagram of the structure of a computer device in an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0029] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0030] It should also be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0031] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0032] See also Figure 1 and Figure 2 , Figure 1 A schematic diagram of an application scenario of the method for intelligently generating financial statements provided in an embodiment of the present invention. Figure 2 A schematic flow chart of a method for intelligently generating financial statements provided in an embodiment of the present invention. The method for intelligently generating financial statements is applied in a server. The server interacts with a terminal, which is integrated with a data middle platform and can provide a variety of data required for financial statements. The server forms initial data by acquiring a variety of data (such as general ledgers, budgets, etc.), and performs preprocessing to obtain preprocessing results. The preprocessing steps include text deduplication, classification, clustering, standardization, noise removal, and data balancing to ensure data quality. The preprocessing results are input into a report automatic generation model, which performs parameter tuning and training of a deep learning model based on a preprocessed sample set, and the generated results are displayed in a visual manner to facilitate user understanding and analysis, thereby improving the generation speed of financial statements, reducing human intervention, and improving data accuracy and reliability.

[0033] Figure 2 Schematic diagram of the process of the intelligent generation method of financial statements provided by the embodiment of the present invention. Figure 2 As shown, the method includes the following steps S110 to S140.

[0034] S110. Obtain various data required for financial statements to obtain initial data.

[0035] In this embodiment, the initial data includes at least one of a general ledger, a budget, expense reimbursement, insurance business, actuarial assumptions, and investment data.

[0036] Specifically, through the data center, data from multiple sources can be effectively integrated, including:

[0037] OF general ledger data: records all financial transactions of the company and provides basic information for reporting.

[0038] Budget data: reflects the company's budget plan and helps evaluate the difference between actual performance and budget.

[0039] Expense reimbursement data: Detailed records of various expenses to provide a basis for cost analysis.

[0040] Insurance business data: related financial data of the company in the insurance field to help assess potential risks.

[0041] Actuarial assumption data: Assumption data used to predict and analyze future financial conditions.

[0042] Investment data: records the financial performance of the company's investment projects and supports financial decision-making.

[0043] These data cover key financial indicators such as revenue, cost, profit, assets and liabilities, providing a comprehensive basis for the generation of financial statements.

[0044] Using the existing data center for data collection and processing can significantly improve the efficiency of financial statement generation. Specifically, the automation capability of the data center can reduce manual intervention, increase data processing speed, and reduce human errors. The data center can achieve real-time data updates, ensure that financial statements reflect the latest financial status, and improve the timeliness of decision-making.

[0045] The data center usually has strong data security and compliance management capabilities:

[0046] Data security: It can ensure the secure storage and transmission of sensitive financial data and protect the interests of the company.

[0047] Compliance: Ensure that financial statements comply with relevant laws and regulations through compliant data collection and processing procedures.

[0048] By utilizing the existing data middle platform to collect and integrate various financial-related data, not only can the accuracy and reliability of financial statements be improved, but also the efficiency and automation of report generation can be improved, and the learning ability of the model can be enhanced, thereby providing the company with more powerful decision-making support and risk management capabilities.

[0049] S120: preprocess the initial data to obtain a preprocessing result.

[0050] In this embodiment, the preprocessing result refers to a cleaned and converted data set obtained after text deduplication, classification, clustering, standardization, noise removal, data balancing, etc.

[0051] Specifically, the initial data is subjected to text deduplication, classification, clustering, standardization, noise removal, and data balancing to obtain a preprocessing result.

[0052] Deduplication of text refers to identifying and removing duplicate records to ensure that each sample in the data set is unique. Reducing duplicate data can avoid bias in analysis. It can also reduce the size of the data set and reduce storage costs.

[0053] Classification is the process of grouping data based on certain features or labels, clustering similar samples together. Classified data can be more easily analyzed or visualized. Through classification, the performance of different categories of data can be quickly identified, providing a basis for decision making.

[0054] Clustering is the process of dividing data into groups based on similarity of features, with high similarity within groups and low similarity between groups. Clustering can help identify underlying patterns and trends in data and discover characteristics of different categories. Through clustering, complex data sets can be simplified into multiple groups that are easier to analyze.

[0055] Standardization is to convert data into the same dimension so that it has a similar range or distribution. Common methods include Z-score standardization and Min-Max standardization. Standardization can eliminate the dimensional effects between different features, making model training more effective. In many machine learning algorithms, standardization helps improve the convergence speed and accuracy of the model.

[0056] Noise removal refers to identifying and removing outliers or erroneous data that have no practical significance for analysis. Noise removal can reduce interference information in the data and improve the validity of the data. The cleaned data makes the model more reflective of the real situation and reduces prediction errors.

[0057] Data balancing is the processing of datasets with imbalanced categories. Common methods include undersampling, oversampling, or generating synthetic samples. Balancing datasets can reduce the bias of the model towards a certain category and improve the overall performance of the model. In an imbalanced dataset, balancing can improve the model's prediction ability for the minority class.

[0058] S130: Input the preprocessing result into a report automatic generation model to generate a report, so as to obtain a generation result.

[0059] In this embodiment, the generated result refers to an automatically generated financial statement.

[0060] Among them, the automatic report generation model is obtained by training the deep learning model with a sample set after preprocessing various data required for the financial report.

[0061] In one embodiment, see Figure 3 The above-mentioned automatic report generation model is obtained by training the deep learning model with a variety of data required for the financial report as a sample set after preprocessing, and may include steps S131 to S135.

[0062] S131. Obtain various data required for financial statements and perform preprocessing to obtain a sample set.

[0063] In this embodiment, the sample set refers to a variety of data required for financial statements, and is formed after preprocessing. The various data and preprocessing are consistent with steps S110 to S120 and are not described in detail here.

[0064] Collecting a large amount of corpus and data sets and ensuring the diversity and representativeness of the data sets has the following benefits:

[0065] Comprehensiveness: Diversified data sources can fully reflect the company's financial status, making the generated financial statements more accurate and reliable.

[0066] Pattern recognition: Through a wide range of data samples, the model is able to learn various underlying patterns and features, thereby improving the accuracy of predictions and analysis.

[0067] Improve the robustness of the model: A representative dataset can enhance the generalization ability of the model, reduce the risk of overfitting, and make the model more robust on new data.

[0068] Support decision-making: A rich data foundation can provide management with more reliable decision-making support and help identify potential opportunities and risks in business development.

[0069] S132. Build a deep learning model.

[0070] In this embodiment, the deep learning model includes a recurrent neural network.

[0071] Use the collected data set and optimized parameters to train the model. Here, a recurrent neural network is used for training. The basic formula of a recurrent neural network is as follows:

[0072] z=Uz previous +v, where z previous represents the vector of the input sequence, U is the weight matrix of the recurrent neural network, and v is the bias vector. Each element of the input sequence is input to the network through the input channel, and the recurrent variable is input through the recurrent channel. By sharing weights, the recurrent neural network is able to learn long-term dependencies in the input sequence, thereby improving its predictive ability.

[0073] Specifically, a recurrent neural network is a neural network that contains an input sequence and a recurrent variable. Each element of the input sequence is input to the recurrent neural network through an input channel, and each element of the recurrent variable is input to the recurrent neural network through a recurrent channel. Each weight matrix and bias vector of the recurrent neural network corresponds to a recurrent variable in the recurrent neural network and is connected to the recurrent variable through a shared weight channel. In this way, the recurrent neural network can learn long-term dependencies in the input sequence to improve its prediction ability.

[0074] Of course, in other embodiments, a long short-term memory network or a gated recurrent unit may be selected as a basic model for training, because they can more effectively capture long-term dependencies and reduce the gradient vanishing problem.

[0075] S133. Adjust parameters of the deep learning model.

[0076] In this embodiment, a back propagation algorithm is used to adjust the parameters of the deep learning model.

[0077] The process of model parameter adjustment depends on the collected data set. In this process, the back propagation algorithm is used to adjust the model parameters. The basic formula of the back propagation algorithm is: In this formula, W t represents the weight matrix at the tth time step, σ i is the weight of the ith neuron, f i (X t ,Y i ) is the error between input and output, and This formula means that at each time step, the error f of the current neuron is first calculated. i (X t ,Y i ), and then according to the neuron weight σ i The output error is calculated and finally added to the weight of the current neuron to update the weight. This method aims to improve the generalization ability and prediction performance of the model by minimizing the difference between the predicted results and the actual results.

[0078] S134. Use the sample set to train the deep learning model after parameter adjustment to obtain a preliminary trained model.

[0079] In this embodiment, the initially trained model refers to a model obtained by training the parameter-adjusted model using a sample set.

[0080] In this embodiment, the training set is input into the deep learning model with adjusted parameters; the output is calculated through forward propagation, and the weights and biases of the model are updated through the back propagation algorithm based on the error between the true label and the predicted result to reduce the prediction error; multiple rounds of training are iterated to continuously optimize the model parameters to better fit the training data.

[0081] After several rounds of training, a preliminary training model is generated. At this point, the model can perform well on the training set, but its generalization ability still needs to be further evaluated.

[0082] S135. Perform performance evaluation and optimization on the preliminary training model to obtain an automatic report generation model.

[0083] In this embodiment, the generated preliminary training model is subjected to in-depth performance evaluation and necessary optimization to ensure the accuracy and reliability of the model. The specific steps include:

[0084] Use the validation set to test the initially trained model and evaluate its performance on unseen data.

[0085] Calculate various performance indicators, such as accuracy, recall, F1-score, ROC curve, etc., to comprehensively evaluate model performance.

[0086] Analyze the model’s predictions to identify weaknesses in specific categories or data types.

[0087] Observe the performance of the model on different categories to identify potential overfitting or underfitting problems.

[0088] Based on the evaluation results, adjust the model structure or further optimize the hyperparameters to improve the model's predictive ability. You can consider increasing regularization, adjusting the learning rate, using different optimization algorithms, and other strategies.

[0089] After completing the performance optimization, the model was verified and adjusted multiple times to form the final report automatic generation model.

[0090] This model can automatically generate accurate reports in practical applications to meet business needs.

[0091] Through the above steps, the accuracy and applicability of the model are ensured, which can effectively support the subsequent data analysis and decision-making process.

[0092] Deploy the report automatic generation model to actual applications, such as the intelligent analysis system of financial statements. According to actual needs, the model can be fine-tuned and optimized to improve the accuracy and efficiency of the model.

[0093] In this embodiment, the above-mentioned model parameter adjustment and training processes use an artificial intelligence agent tool Auto-GPT, which can reasonably decompose complex tasks into small tasks that can be executed independently, and can decompose complex model parameter adjustment and model training into small tasks for separate execution, thereby improving efficiency.

[0094] Specifically, Auto-GPT can decompose the complex model parameter adjustment and training process into multiple independent small tasks. This process usually includes the following aspects:

[0095] Identify the various components of the entire model tuning and training process, such as data preprocessing, feature selection, hyperparameter tuning, model training, model evaluation, etc.

[0096] Further break down each component into smaller tasks. For example, hyperparameter tuning can be broken down into selecting different hyperparameter combinations, running cross validation, evaluating model performance, etc.

[0097] Each small task after disassembly can be executed independently, thus achieving higher efficiency. The specific steps include:

[0098] By utilizing a multithreaded or distributed computing framework, multiple small tasks can be processed simultaneously, which greatly shortens the overall execution time.

[0099] Auto-GPT is responsible for managing the execution order and status of each task, ensuring that all tasks are called and completed at the appropriate time.

[0100] Through reasonable decomposition and independent execution, Auto-GPT improves the efficiency of the entire model parameter adjustment and training process. This is reflected in the following aspects:

[0101] Parallel execution of small tasks reduces the overall time required, allowing researchers and developers to obtain results faster.

[0102] Adjusting the parameters or logic of a small task will not affect the execution of other tasks, which improves the flexibility and adjustability of the overall process.

[0103] Auto-GPT’s intelligent capabilities make it possible to automatically select and optimize the order in which tasks are executed, further improving efficiency. For example, after model training is complete, performance evaluation and result recording are automatically performed.

[0104] During the task execution process, Auto-GPT can collect the execution results of each small task and make dynamic adjustments based on these feedbacks:

[0105] When the result of a small task is not ideal, it can be fed back to the scheduling system in real time, prompting the system to automatically select different strategies for retry or adjustment.

[0106] Based on the results of each round of execution, Auto-GPT can iteratively optimize task strategies and continuously improve model performance.

[0107] Through the efficient task management and execution described above, a fully tuned and validated model is finally generated:

[0108] The results of all small tasks will be integrated to form the final model and evaluation report for subsequent analysis and decision-making.

[0109] The final model can not only adapt to the current data and task requirements, but also has certain scalability and can continue to be used in subsequent model iterations.

[0110] In summary, using Auto-GPT for model parameter adjustment and training can significantly improve work efficiency, promote rapid iteration and optimization of models, and meet the needs of complex tasks. This method is particularly suitable for processing large-scale data and complex models, and provides strong support for the research and application of machine learning and deep learning.

[0111] S140: Output the generated result.

[0112] In this embodiment, the generation result is displayed in a visual manner.

[0113] The corresponding financial statements are generated according to the trained model, and a visual display is provided so that users can easily view and compare. The report generation will use a custom template, and the display format can include a variety of visual expressions such as tables and charts.

[0114] The method of this embodiment can automatically obtain relevant data from multiple sources (such as databases, financial software, APIs, etc.). Through the built-in data cleaning algorithm, duplicate data and erroneous data are automatically excluded to ensure the accuracy of input data. Advanced analysis algorithms (such as machine learning models) are used to automatically generate financial indicators, trend analysis, etc., reducing the time and potential errors of manual analysis. In other words, the method of this embodiment can automatically process data, from data collection, cleaning to analysis, and the entire process reduces manual intervention.

[0115] In addition, the model and initial data can also be set to be automatically updated regularly to ensure that the information in the report always reflects the latest financial status. It covers financial statements (such as balance sheets, income statements, cash flow statements) and related financial indicators to provide a comprehensive perspective for decision-making. This enables the method of this embodiment to generate comprehensive and accurate financial information based on real-time updates.

[0116] Users can quickly query and filter information through an intuitive interface, such as generating customized reports by time period, department, project, etc. In addition, it can intelligently recommend commonly used report formats and analysis dimensions based on historical data and user behavior analysis to improve user experience. Not only does it generate static reports, it also provides interactive functions, allowing users to deeply analyze data.

[0117] Through seamless integration of the data center with the existing enterprise resource planning (ERP) system and other financial software, the financial process is made more efficient. Digital reports not only improve information processing efficiency, but also enhance the company's professional image in the industry and enhance the trust of customers and investors, thereby promoting the digital transformation of corporate finance.

[0118] The above-mentioned intelligent financial statement generation method obtains the required multiple data and preprocesses them, then inputs the processed data into the deep learning model to generate the report, and finally outputs the results. Through automated data acquisition and processing, the speed of report generation is significantly improved, and the need for manual intervention is reduced. This method improves the accuracy and reliability of data through precise preprocessing and deep learning model training, thereby enhancing the quality of the report.

[0119] Figure 4is a schematic block diagram of a financial statement intelligent generation device 300 provided by an embodiment of the present invention. Figure 4 As shown, corresponding to the above financial statement intelligent generation method, the present invention also provides a financial statement intelligent generation device 300. The financial statement intelligent generation device 300 includes a unit for executing the above financial statement intelligent generation method, and the device can be configured in a server. Figure 4 The financial statement intelligent generation device 300 includes an initial data acquisition unit 301, a preprocessing unit 302, an automatic generation unit 303 and an output unit 304.

[0120] The initial data acquisition unit 301 is used to acquire various data required for the financial report to obtain the initial data; the preprocessing unit 302 is used to preprocess the initial data to obtain the preprocessing result; the automatic generation unit 303 is used to input the preprocessing result into the report automatic generation model to generate the report to obtain the generation result; the output unit 304 is used to output the generation result;

[0121] The automatic report generation model is obtained by training the deep learning model with various data required for the financial report as sample sets after preprocessing. The output unit 304 is used to output the generation result.

[0122] In one embodiment, the preprocessing unit 302 is used to perform text deduplication, classification, clustering, standardization, noise removal, and data balancing on the initial data to obtain a preprocessing result.

[0123] In one embodiment, the financial statement intelligent generation device 300 also includes a model generation unit, which is used to train a deep learning model by using a variety of data required for financial statements as a sample set after preprocessing, so as to form an automatic report generation model.

[0124] In one embodiment, the model generation unit includes

[0125] The sample set generation subunit is used to obtain various data required for financial statements and perform preprocessing to obtain a sample set; the model construction subunit is used to build a deep learning model; the parameter adjustment subunit is used to adjust the parameters of the deep learning model; the training subunit is used to use the sample set to train the deep learning model after parameter adjustment to obtain a preliminary trained model; the optimization subunit is used to perform performance evaluation and optimization on the preliminary training model to obtain an automatic report generation model.

[0126] In one embodiment, the parameter adjustment unit is used to adjust the parameters of the deep learning model using a back propagation algorithm.

[0127] In one embodiment, the output unit 304 is used to display the generated result in a visual manner.

[0128] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the above-mentioned financial statement intelligent generation device 300 and each unit can refer to the corresponding description in the aforementioned method embodiment, and for the convenience and brevity of description, it will not be repeated here.

[0129] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 5 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, it implements the functions or steps on the server side of a method for intelligently generating financial statements.

[0130] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0131] Acquire various data required for financial statements to obtain initial data; preprocess the initial data to obtain preprocessing results; input the preprocessing results into a report automatic generation model to generate reports to obtain generation results; output the generation results;

[0132] Among them, the automatic report generation model is obtained by training the deep learning model with a sample set after preprocessing various data required for the financial report.

[0133] In one embodiment, a storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0134] Acquire various data required for financial statements to obtain initial data; preprocess the initial data to obtain preprocessing results; input the preprocessing results into a report automatic generation model to generate reports to obtain generation results; output the generation results;

[0135] Among them, the automatic report generation model is obtained by training the deep learning model with a sample set after preprocessing various data required for the financial report.

[0136] It should be noted that the functions or steps that can be implemented by the above storage medium or computer device can be correspondingly referred to the relevant descriptions on the server side and the client side in the aforementioned method embodiment. In order to avoid repetition, they will not be described one by one here.

[0137] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0138] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0139] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for intelligently generating financial statements, characterized in that: include: Obtain various data required for financial statements to obtain initial data; Preprocessing the initial data to obtain a preprocessing result; Inputting the preprocessing result into the report automatic generation model to generate the report, so as to obtain the generation result; Outputting the generated result; Among them, the automatic report generation model is obtained by training the deep learning model with a sample set after preprocessing various data required for the financial report.

2. The method for intelligently generating financial statements according to claim 1, characterized in that: The initial data includes at least one of a general ledger, a budget, expense reimbursement, insurance business, actuarial assumptions, and investment data.

3. The method for intelligently generating financial statements according to claim 1, characterized in that: The preprocessing of the initial data to obtain a preprocessing result includes: The initial data is subjected to text deduplication, classification, clustering, standardization, noise removal, and data balancing to obtain a preprocessing result.

4. The method for intelligently generating financial statements according to claim 3, characterized in that: The automatic report generation model is obtained by training the deep learning model with a variety of data required for financial reports after preprocessing as a sample set, including: Obtain various data required for financial statements and pre-process them to obtain a sample set; Build deep learning models; Adjusting parameters of the deep learning model; Using the sample set to train the adjusted deep learning model to obtain a preliminarily trained model; The preliminary training model is subjected to performance evaluation and optimization to obtain an automatic report generation model.

5. The method for intelligently generating financial statements according to claim 4, characterized in that: The deep learning model includes a recurrent neural network.

6. The method for intelligently generating financial statements according to claim 4, characterized in that: The step of adjusting parameters of the deep learning model includes: The parameters of the deep learning model are adjusted using a back-propagation algorithm.

7. The method for intelligently generating financial statements according to claim 1, characterized in that: The outputting the generated result includes: The generated results are presented in a visual manner.

8. An intelligent financial statement generation device, characterized in that: include: An initial data acquisition unit is used to acquire various data required for financial statements to obtain initial data; A preprocessing unit, used for preprocessing the initial data to obtain a preprocessing result; An automatic generation unit, used for inputting the preprocessing result into a report automatic generation model to generate a report, so as to obtain a generation result; An output unit, used for outputting the generated result; Among them, the automatic report generation model is obtained by training the deep learning model with a sample set after preprocessing various data required for the financial report.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method for intelligently generating financial statements as described in any one of claims 1 to 7 are implemented.

10. A storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for intelligently generating financial statements as claimed in any one of claims 1 to 7 are implemented.

Images