Financial index prediction method and device, storage medium and computer equipment
By preprocessing and slicing financial data using machine learning and deep learning models, and combining this with distributed computing optimization, the efficiency and accuracy issues in financial indicator calculation are resolved, achieving efficient, stable, and scalable financial indicator prediction.
Patent Information
- Application Number
- CN202511064946.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-29
- Publication Date
- 2025-11-14
AI Technical Summary
Existing technologies for calculating financial indicators suffer from problems such as low data processing efficiency, slow calculation speed, insufficient accuracy, low resource utilization, lack of intelligent anomaly handling, and insufficient system scalability, making it difficult to meet real-time and complex business needs.
The raw financial data is preprocessed using a pre-set machine learning model, and the data is sliced based on the calculation rules and data volume of different product objects. Financial indicators are predicted using a pre-trained deep learning network model, and a comprehensive evaluation is performed through distributed computing and multi-threaded optimization.
It improves the accuracy of financial indicator forecasts, reduces calculation errors, enhances system stability and real-time performance, supports massive data processing and complex business scenarios, and significantly improves system scalability.
Smart Images
Figure CN120952234A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of data processing technology and can be applied to the fields of financial technology business and medical and health care and elderly care business. In particular, it relates to a method, device, storage medium and computer equipment for predicting financial indicators. Background Technology
[0002] Currently, methods for calculating financial indicators mainly rely on traditional numerical calculation methods (such as Newton's method) and manually assisted data processing workflows, which have the following drawbacks: Low data processing efficiency: Data sources are scattered, spanning long historical periods (e.g., 106 years of data), with massive data volumes and complex calculation formulas, leading to lengthy data preprocessing and cleaning times. Relying on manual operations for data synchronization, transformation, and logical processing (e.g., using the KETTLE task to synchronize data from SQL Server to Oracle) is inefficient and error-prone. Inadequate trade-off between computational efficiency and accuracy: Traditional methods cannot fully utilize multi-core processor resources, resulting in slow calculation speeds and difficulty meeting real-time requirements. While Newton's method is effective in certain scenarios, its convergence speed is slow or even non-convergent in cases of data anomalies or nonlinearity, affecting calculation accuracy. Lack of intelligent anomaly handling: Data anomalies (such as missing cash flow or numerical anomalies) typically require manual analysis and processing, which is time-consuming and difficult to respond to business needs in a timely manner. The lack of automated anomaly detection and repair mechanisms affects the stability and reliability of financial indicator calculations. Low resource utilization: Computational tasks are concentrated on a single node, failing to fully utilize distributed computing resources and resulting in wasted computing resources. Lack of scalability: The system struggles to adapt to the rapid growth of future data volumes and the diverse needs of complex business scenarios, exhibiting insufficient scalability. Fintech business platforms support functions such as shopping, social networking, interactive games, and resource transfer, and offer features like loan applications, credit card issuance, and the purchase of insurance and wealth management products. These platforms can be insurance systems, banking systems, transaction systems, order systems, etc. In healthcare and elderly care business systems, business data processing can support functions such as disease-assisted diagnosis, health management, and remote consultations. The platform for processing this business data can be a medical platform. Summary of the Invention
[0003] In view of this, the present invention provides a method, apparatus, storage medium and computer equipment for predicting financial indicators, the main purpose of which is to solve the problem of low data processing efficiency in the current process of determining financial indicators.
[0004] To address the above problems, this application provides a method for predicting financial indicators, including:
[0005] The raw financial data is preprocessed using a pre-defined machine learning model to obtain the target financial data.
[0006] Based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to the different product objects, the target financial data is sliced to obtain each financial dataset.
[0007] A pre-trained deep learning network model is used to predict financial indicators for each of the financial datasets, thereby obtaining initial financial indicator values corresponding to each of the financial datasets.
[0008] A comprehensive evaluation of the initial financial indicators is conducted to obtain the target financial indicator values.
[0009] Optionally, the step of using a preset machine learning model to preprocess the original financial data to obtain the target financial data specifically includes:
[0010] Anomaly processing is performed on the original financial data using a preset neural network model to obtain the first financial data.
[0011] The Lasso regression model is used to extract features from the first financial data to obtain the second financial data;
[0012] The first financial data is filtered for features using a random forest model to obtain the third financial data.
[0013] The second financial data and the third financial data are merged to obtain the target financial data.
[0014] Optionally, the step of slicing the target financial data based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to the different product objects to obtain various financial datasets specifically includes:
[0015] The calculation process is performed based on the calculation rules corresponding to different product objects, the data volume of the target financial data corresponding to the different product objects, and the preset data volume slicing threshold to obtain the number of slices for the target financial data.
[0016] When the number of slices is less than or equal to the preset number of threads, the target financial data is processed by parallel data slicing to obtain each financial dataset.
[0017] Optionally, the method further includes:
[0018] When the number of slices is greater than the preset number of threads, the preset data volume slice threshold is dynamically updated according to the calculation rules and the preset number of threads with the goal of maximizing resource utilization.
[0019] Based on the updated preset data volume slicing threshold, parallel data slicing processing is performed on the target financial data to obtain each financial dataset.
[0020] Optionally, the step of using a pre-trained deep learning network model to predict financial indicators for each of the financial datasets, and obtaining initial financial indicator values corresponding to each of the financial datasets, specifically includes:
[0021] A target neural network using a pre-trained deep learning network model is used to predict financial indicators for each of the financial datasets, thereby obtaining the initial financial indicator values corresponding to each of the financial datasets.
[0022] Optionally, before the target neural network using the pre-trained deep learning network model predicts financial indicators for each of the financial datasets, the method further includes: constructing a deep learning network model;
[0023] The construction of the deep learning network model specifically includes:
[0024] Obtain multiple historical cash flow data sequences;
[0025] The Newton-Raphson iteration method is used to calculate and process each of the historical cash flow data sequences to obtain historical financial indicators corresponding to each of the historical cash flow data sequences.
[0026] Label data is obtained by annotating each of the historical cash flow data sequences based on the historical financial indicators.
[0027] The initial deep neural network is trained using the labeled data as training samples to obtain the target neural network that meets the model convergence condition.
[0028] The parameters of the target neural network are optimized using a reinforcement learning Q-network based on a preset reward function to obtain the deep learning network model.
[0029] Optionally, the step of comprehensively evaluating each of the initial financial indicator values to obtain the target financial indicator value specifically includes:
[0030] The target financial indicator value is obtained by performing a weighted average calculation based on the initial financial indicator value corresponding to each product object and the weight coefficient corresponding to each product object.
[0031] To address the aforementioned problems, this application provides a financial indicator forecasting device, comprising:
[0032] The preprocessing module is used to preprocess the raw financial data using a preset machine learning model to obtain the target financial data.
[0033] The data slicing module is used to slice the target financial data based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to the different product objects, so as to obtain each financial dataset.
[0034] The indicator prediction module is used to use a pre-trained deep learning network model to predict financial indicators for each of the financial datasets, and obtain the initial financial indicator values corresponding to each of the financial datasets.
[0035] The comprehensive evaluation module is used to comprehensively evaluate the initial financial indicator values to obtain the target financial indicator values.
[0036] To address the aforementioned problems, this application provides a storage medium storing a computer program that, when executed by a processor, implements the steps of the aforementioned financial indicator prediction method.
[0037] To address the aforementioned problems, this application provides a computer device, comprising at least a memory and a processor. The memory stores a computer program, and the processor, when executing the computer program in the memory, implements the steps of the aforementioned financial indicator prediction method.
[0038] The beneficial effects of this application are as follows: This application preprocesses the original financial data using a pre-set machine learning model to obtain target financial data; by replacing traditional numerical methods with a pre-set machine learning model, the prediction accuracy of financial indicators is improved, and calculation errors caused by data anomalies are reduced. Automated anomaly detection and repair reduce manual intervention, improve system stability, and ensure the timeliness and reliability of financial indicator calculations. Based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to the different product objects, the target financial data is sliced to obtain various financial datasets; a pre-trained deep learning network model is used to predict financial indicators for each of the financial datasets to obtain initial financial indicator values corresponding to each financial dataset; based on distributed computing methods and multi-threaded optimization, the data processing speed is significantly improved to meet real-time business needs; a comprehensive evaluation of each of the initial financial indicator values is performed to obtain the target financial indicator value. It supports massive data processing and complex business scenarios, significantly improves system scalability, and is suitable for future business growth needs.
[0039] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description
[0040] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:
[0041] Figure 1 This illustration shows an application environment diagram of a financial indicator forecasting method provided in an embodiment of this application;
[0042] Figure 2 A flowchart illustrating a financial indicator forecasting method provided in an embodiment of this application is shown.
[0043] Figure 3 A flowchart illustrating a financial indicator forecasting method according to another embodiment of this application is shown;
[0044] Figure 4 A structural block diagram of a financial indicator forecasting device according to another embodiment of this application is shown;
[0045] Figure 5 A schematic diagram of the structure of a computer device according to an embodiment of this application is shown;
[0046] Figure 6 Another structural schematic diagram of a computer device according to one embodiment of this application is shown. Detailed Implementation
[0047] Various embodiments and features of this application are described herein with reference to the accompanying drawings.
[0048] It should be understood that various modifications can be made to the embodiments described herein. Therefore, the above description should not be considered as limiting, but merely as an example of embodiments. Other modifications within the scope and spirit of this application will be apparent to those skilled in the art.
[0049] The accompanying drawings, which are included in and form part of this specification, illustrate embodiments of the present application and, together with the general description of the present application given above and the detailed description of the embodiments given below, serve to explain the principles of the present application.
[0050] These and other features of this application will become apparent from the following description of preferred forms of embodiments given as non-limiting examples, with reference to the accompanying drawings.
[0051] It should also be understood that although this application has been described with reference to some specific examples, those skilled in the art can certainly implement many other equivalent forms of this application.
[0052] The above and other aspects, features and advantages of this application will become more apparent when taken in conjunction with the accompanying drawings and in view of the following detailed description.
[0053] Specific embodiments of this application are described thereafter with reference to the accompanying drawings; however, it should be understood that the claimed embodiments are merely examples of this application, which can be implemented in various ways. Well-known and / or repeated functions and structures are not described in detail to avoid unnecessary or redundant details that could obscure the application. Therefore, the specific structural and functional details claimed herein are not intended to be limiting, but merely serve as the basis and representative basis for the claims to teach those skilled in the art to use this application in a variety of substantially any suitable detailed structures.
[0054] This specification may use the phrases “in one embodiment,” “in another embodiment,” “in yet another embodiment,” or “in other embodiments,” all of which may refer to one or more of the same or different embodiments according to this application.
[0055] The financial indicator forecasting method provided in this application can be applied to, for example, Figure 1 In this application environment, the client communicates with the server via a network. The server uses a pre-set machine learning model to preprocess the original financial data to obtain target financial data; based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to different product objects, the target financial data is sliced to obtain various financial datasets; a pre-trained deep learning network model is used to predict financial indicators for each of the financial datasets to obtain initial financial indicator values corresponding to each financial dataset; a comprehensive evaluation of each initial financial indicator value is performed to obtain the target financial indicator value. The target financial indicator value is then fed back to the client. This application uses a pre-set machine learning model to preprocess the original financial data to obtain target financial data; by using a pre-set machine learning model to replace traditional numerical methods, the prediction accuracy of financial indicators is improved, and calculation errors caused by data anomalies are reduced. Automated anomaly detection and repair reduce manual intervention, improve system stability, and ensure the timeliness and reliability of financial indicator calculation. Based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to the different product objects, the target financial data is sliced to obtain each financial dataset; a pre-trained deep learning network model is used to predict financial indicators for each financial dataset to obtain the initial financial indicator values corresponding to each financial dataset; based on distributed computing methods and multi-threaded optimization, the data processing speed is significantly improved.
[0056] This application provides a method for predicting financial indicators, such as... Figure 2 As shown, it includes:
[0057] Step S101: Use a preset machine learning model to preprocess the raw financial data to obtain the target financial data;
[0058] In this step, a pre-set neural network model is used to process abnormal data in the original financial data to obtain first financial data; a Lasso regression model is used to extract features from the first financial data to obtain second financial data; a random forest model is used to filter features from the first financial data to obtain third financial data; and the second and third financial data are then fused to obtain the target financial data. In the fintech business field, financial data can be extracted from a business service platform. This business service platform can be an insurance system, a banking system, a transaction system, an order system, etc. In the healthcare and elderly care business system, business data can be medical sales data, such as medical service revenue data, elderly care business revenue data, health management revenue data, subsidy revenue, and other revenue data such as medical equipment leasing revenue.
[0059] Step S102: Based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to the different product objects, the target financial data is sliced to obtain each financial dataset;
[0060] In the specific implementation process, this step involves calculation based on the calculation rules corresponding to different product objects, the data volume of the target financial data corresponding to the different product objects, and the preset data volume slicing threshold to obtain the number of slices for the target financial data; when the number of slices is less than or equal to the preset number of threads, the target financial data is processed in parallel to obtain each financial dataset.
[0061] Step S103: Use a pre-trained deep learning network model to predict financial indicators for each of the financial datasets, and obtain the initial financial indicator values corresponding to each of the financial datasets.
[0062] In the specific implementation process of this step, a pre-trained deep learning network model is used to predict financial indicators for each of the financial datasets, thereby obtaining the initial financial indicator values corresponding to each of the financial datasets.
[0063] Step S104: Conduct a comprehensive evaluation of each of the initial financial indicator values to obtain the target financial indicator values.
[0064] In the specific implementation process of this step, a weighted average calculation is performed based on the initial financial indicator value corresponding to each product object and the weight coefficient corresponding to each product object to obtain the target financial indicator value.
[0065] This application preprocesses raw financial data using a pre-set machine learning model to obtain target financial data. By replacing traditional numerical methods with the pre-set machine learning model, the prediction accuracy of financial indicators is improved, reducing calculation errors caused by data anomalies. Automated anomaly detection and repair reduce manual intervention, improve system stability, and ensure the timeliness and reliability of financial indicator calculations. Based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to those product objects, the target financial data is sliced to obtain various financial datasets. A pre-trained deep learning network model is used to predict financial indicators for each of the financial datasets, obtaining initial financial indicator values corresponding to each dataset. Based on distributed computing methods and multi-threaded optimization, data processing speed is significantly improved to meet real-time business needs. A comprehensive evaluation of each initial financial indicator value yields the target financial indicator value. This application supports massive data processing and complex business scenarios, significantly improves system scalability, and is suitable for future business growth needs.
[0066] Another embodiment of this application provides a different method for predicting financial indicators, such as... Figure 3 As shown, it includes:
[0067] Step S201: Use a preset neural network model to process the original financial data for abnormal data to obtain the first financial data;
[0068] In this step, a preset neural network model is used to extract features and detect anomalies in the original financial data, automatically identifying missing and outlier values. The missing and outlier values are then processed to obtain the first set of financial data. This results in financial data suitable for model input and repairs the dynamically updated rule base of the knowledge graph, supporting flexible adjustments to business rules and providing intelligent repair suggestions. The original financial data includes transaction types, cash inflows, cash outflows, investment amounts, and cost amounts corresponding to different sales products. This source data can be obtained through cross-database queries from the financial system, sales system, and ERP system. The source data obtained from the cross-database queries is then processed to a unified format to obtain the original financial data.
[0069] Step S202: Use the Lasso regression model to extract features from the first financial data to obtain the second financial data;
[0070] In this step, the Lasso regression model is used to extract features from the first financial data to obtain the first feature data; the feature coefficients of each of the first feature data are obtained; and the first feature data is filtered based on the feature coefficients to remove unimportant features, thereby obtaining the second financial data.
[0071] Step S203: Use a random forest model to filter features of the first financial data to obtain the third financial data;
[0072] In this step, a random forest model is used to score the importance of the first financial data, obtaining importance scores corresponding to each feature data in the first financial data. Based on the importance scores, the first financial data is filtered to remove feature data with lower importance scores, thus obtaining the third financial data.
[0073] Step S204: Merge the second financial data and the third financial data to obtain the target financial data;
[0074] In this step, an improved LSTM model is used to fuse the second and third financial data to obtain the target financial data. Specifically, the LSTM layer of the improved LSTM model extracts features of the short-term dependencies between the second and third financial data, capturing local features of the cash flow data to obtain a first local feature corresponding to the second financial data and a second local feature corresponding to the third financial data. The Transformer layer of the improved LSTM model, which incorporates a self-attention mechanism, extracts global features of the second and third financial data to obtain a first global feature corresponding to the second financial data and a second global feature corresponding to the third financial data. The first local feature, the second local feature, the first global feature, and the second global feature are then fused to obtain the target financial data. Combining short-term and long-term features improves the model's prediction accuracy.
[0075] Step S205: Based on the calculation rules corresponding to different product objects, the data volume of the target financial data corresponding to the different product objects, and the preset data volume slicing threshold, perform calculation processing to obtain the number of data slices for the target financial data;
[0076] In the specific implementation process of this step, the calculation rules include IRR calculation functions, cost allocation functions, etc.; different product objects correspond to different calculation rules; a mapping relationship between different product objects and different calculation rules is established to obtain a rule mapping table; based on the calculation rules corresponding to different product objects, the data volume of the target financial data corresponding to the different product objects, and the preset data volume slicing threshold, calculation processing is performed to obtain the number of data slices for the target financial data; for example: the calculation rule for product A is to use the IRR calculation function for calculation, the data volume is 1.2 million time series transaction records, and the preset data volume slicing threshold corresponding to product A is 100,000 records / slice, then the number of slices = 1.2 million / 10 = 12 slices, that is, the number of slices obtained is 12 slices. For example, when the product object is a B-type product, the calculation rule corresponding to the B-type product is obtained as the cost allocation calculation rule based on the B-type product query rule mapping table; when the data volume of the B-type product is 500,000 records, involving 5 departments, and the preset data volume slicing threshold corresponding to the B-type product is 200,000 records / slice, the number of slices = 5 slices (grouped by department, each group is calculated independently), etc.
[0077] Step S206: When the number of slices is less than or equal to the preset number of threads, the target financial data is processed by parallel data slicing to obtain each financial dataset;
[0078] In the specific implementation of this step, when the number of slices is less than or equal to the preset number of threads, each thread processes an independent data slice and generates a financial dataset corresponding to the number of slices in parallel.
[0079] Step S207: When the number of slices is greater than the preset number of threads, the preset data volume slice threshold is dynamically updated according to the calculation rules and the preset number of threads with the goal of maximizing resource utilization.
[0080] In this step, the computational task complexity is determined based on the aforementioned calculation rules; a load factor is determined based on the task complexity; and calculations are performed based on the load factor, the preset number of threads, and the target financial data to obtain an updated data volume slice threshold. The load factor ranges from 1 to 2, and its value can be adjusted according to the actual task computational complexity. Specifically, the updated data volume slice threshold is obtained by dividing the total target financial data by the product of the number of threads and the load factor.
[0081] Step S208: Perform parallel data slicing processing on the target financial data based on the updated data volume slicing threshold to obtain each financial dataset;
[0082] In the specific implementation process, this step involves calculation based on the calculation rules corresponding to different product objects, the data volume of the target financial data corresponding to the different product objects, and the updated data volume slicing threshold, to obtain the first number of data slices for the target financial data; each thread processes an independent data slice and generates a financial dataset corresponding to the first number of slices in parallel.
[0083] Step S209: Construct a deep learning network model;
[0084] In this step, multiple historical cash flow data sequences are acquired; Newton's iteration method is used to calculate and process each historical cash flow data sequence to obtain historical financial indicators corresponding to each historical cash flow data sequence; based on the historical financial indicators, each historical cash flow data sequence is labeled to obtain label data; based on the label data as training samples, an initial deep neural network is trained to obtain a target neural network that meets the model convergence conditions; a reinforcement learning Q-network is used to optimize the parameters of the target neural network based on a preset reward function to obtain the deep learning network model.
[0085] Step S210: Use a target neural network of a deep learning network model to predict financial indicators for each of the financial datasets, and obtain the initial financial indicator values corresponding to each of the financial datasets;
[0086] In this step, each of the financial datasets is input into a deep learning network model, and the target neural network of the deep learning network model is used to predict the financial indicators of each financial dataset to obtain the initial financial indicator values corresponding to each financial dataset.
[0087] Step S211: Calculate the target financial indicator value based on the initial financial indicator value corresponding to each product object and the weight coefficient corresponding to each product object.
[0088] In the specific implementation process of this step, a weighted average calculation is performed based on the initial financial indicator value corresponding to each product object and the weight coefficient corresponding to each product object to obtain the target financial indicator value.
[0089] This application employs a preset neural network model to process abnormal data from the original financial data, obtaining first financial data; uses a Lasso regression model to extract features from the first financial data, obtaining second financial data; uses a random forest model to filter features from the first financial data, obtaining third financial data; fuses the second and third financial data to obtain the target financial data; calculates the number of slices for the target financial data based on calculation rules corresponding to different product objects, the data volume of the target financial data corresponding to different product objects, and a preset data volume slicing threshold; when the number of slices is less than or equal to a preset number of threads, parallel data slicing is performed on the target financial data to obtain each financial dataset; when the number of slices is greater than the preset number of threads, the preset data volume slicing threshold is dynamically updated according to the calculation rules and the preset number of threads with the goal of maximizing resource utilization; distributed computing and dynamic task allocation improve resource utilization and reduce computing costs. Based on the updated data volume slicing threshold, parallel data slicing processing is performed on the target financial data to obtain each financial dataset; a deep learning network model is constructed; the target neural network of the deep learning network model is used to predict financial indicators for each financial dataset to obtain the initial financial indicator value corresponding to each financial dataset; the target financial indicator value is obtained by calculation based on the initial financial indicator value corresponding to each product object and the weight coefficient corresponding to each product object. This application uses a preset machine learning model to replace traditional numerical methods, which improves the prediction accuracy of financial indicators and reduces calculation errors caused by data anomalies. Automated anomaly detection and repair reduce manual intervention, improve system stability, and ensure the timeliness and reliability of financial indicator calculation. Based on distributed computing methods and multi-threaded optimization, the data processing speed is significantly improved to meet real-time business needs; it supports massive data processing and complex business scenarios, significantly improves system scalability, and is suitable for future business growth needs.
[0090] Another embodiment of this application provides a financial indicator prediction device, such as... Figure 4 As shown, it includes:
[0091] Preprocessing module 1 is used to preprocess the raw financial data using a preset machine learning model to obtain the target financial data;
[0092] Data slicing module 2 is used to slice the target financial data based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to the different product objects, so as to obtain each financial dataset;
[0093] The indicator prediction module 3 is used to use a pre-trained deep learning network model to predict financial indicators for each of the financial datasets, and obtain the initial financial indicator values corresponding to each of the financial datasets.
[0094] The comprehensive evaluation module 4 is used to conduct a comprehensive evaluation of each of the initial financial indicator values to obtain the target financial indicator value.
[0095] In the specific implementation process, the preprocessing module 1 is specifically used to: perform abnormal data processing on the original financial data using a preset neural network model to obtain first financial data; perform feature extraction on the first financial data using a Lasso regression model to obtain second financial data; perform feature filtering on the first financial data using a random forest model to obtain third financial data; and perform fusion processing on the second financial data and the third financial data to obtain the target financial data.
[0096] In the specific implementation process, the data slicing module 2 is specifically used to: perform calculation processing based on the calculation rules corresponding to different product objects, the data volume of the target financial data corresponding to the different product objects, and the preset data volume slicing threshold to obtain the number of slices for data slicing of the target financial data; when the number of slices is less than or equal to the preset number of threads, perform parallel data slicing processing on the target financial data to obtain each financial dataset.
[0097] In the specific implementation process, the data slicing module 2 is also used to: when the number of slices is greater than the preset number of threads, dynamically update the preset data volume slicing threshold according to the calculation rules and the preset number of threads with the goal of maximizing resource utilization; and perform parallel data slicing processing on the target financial data based on the updated preset data volume slicing threshold to obtain each of the financial datasets.
[0098] In the specific implementation process, the indicator prediction module 3 specifically includes: using a pre-trained deep learning network model to predict financial indicators for each of the financial datasets, and obtaining the initial financial indicator values corresponding to each of the financial datasets.
[0099] In specific implementation, the device further includes a model building module, which is specifically used to: acquire multiple historical cash flow data sequences; perform calculations on each of the historical cash flow data sequences using the Newton-Raphson iteration method to obtain historical financial indicators corresponding to each of the historical cash flow data sequences; label each of the historical cash flow data sequences based on the historical financial indicators to obtain label data; train an initial deep neural network based on the label data as training samples to obtain a target neural network that meets the model convergence conditions; and optimize the parameters of the target neural network using a reinforcement learning Q-network based on a preset reward function to obtain the deep learning network model.
[0100] In the specific implementation process, the comprehensive evaluation module 4 is specifically used to: perform weighted average calculation based on the initial financial indicator value corresponding to each product object and the weight coefficient corresponding to each product object to obtain the target financial indicator value.
[0101] This application preprocesses raw financial data using a pre-set machine learning model to obtain target financial data. By replacing traditional numerical methods with the pre-set machine learning model, the prediction accuracy of financial indicators is improved, reducing calculation errors caused by data anomalies. Automated anomaly detection and repair reduce manual intervention, improve system stability, and ensure the timeliness and reliability of financial indicator calculations. Based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to those product objects, the target financial data is sliced to obtain various financial datasets. A pre-trained deep learning network model is used to predict financial indicators for each of the financial datasets, obtaining initial financial indicator values corresponding to each dataset. Based on distributed computing methods and multi-threaded optimization, data processing speed is significantly improved to meet real-time business needs. A comprehensive evaluation of each initial financial indicator value yields the target financial indicator value. This application supports massive data processing and complex business scenarios, significantly improves system scalability, and is suitable for future business growth needs.
[0102] Another embodiment of this application provides a storage medium storing a computer program, which, when executed by a processor, implements the following method steps:
[0103] Step 1: Use a pre-set machine learning model to preprocess the raw financial data to obtain the target financial data;
[0104] Step 2: Based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to the different product objects, perform data slicing on the target financial data to obtain each financial dataset;
[0105] Step 3: Use a pre-trained deep learning network model to predict financial indicators for each of the financial datasets, and obtain the initial financial indicator values corresponding to each of the financial datasets.
[0106] Step 4: Conduct a comprehensive evaluation of the initial financial indicator values to obtain the target financial indicator values.
[0107] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, 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. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
[0108] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to 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.
[0109] The specific implementation process of the above method steps can be found in the embodiments of the above-mentioned arbitrary financial indicator prediction methods, which will not be repeated here.
[0110] This application preprocesses raw financial data using a pre-set machine learning model to obtain target financial data. By replacing traditional numerical methods with the pre-set machine learning model, the prediction accuracy of financial indicators is improved, reducing calculation errors caused by data anomalies. Automated anomaly detection and repair reduce manual intervention, improve system stability, and ensure the timeliness and reliability of financial indicator calculations. Based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to those product objects, the target financial data is sliced to obtain various financial datasets. A pre-trained deep learning network model is used to predict financial indicators for each of the financial datasets, obtaining initial financial indicator values corresponding to each dataset. Based on distributed computing methods and multi-threaded optimization, data processing speed is significantly improved to meet real-time business needs. A comprehensive evaluation of each initial financial indicator value yields the target financial indicator value. This application supports massive data processing and complex business scenarios, significantly improves system scalability, and is suitable for future business growth needs.
[0111] Another embodiment of this application provides a computer device, which may be a server, and its internal structure is shown in the figure below. Figure 5 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with external clients via a network connection. When the computer device program is executed by the processor, it implements the functions or steps of a financial indicator forecasting method on the server side.
[0112] In one embodiment, a computer device is provided, which may be a client. Its internal structure is illustrated below. Figure 6 As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it implements the functions or steps of a financial indicator forecasting method on the client side.
[0113] Another embodiment of this application provides a computer device, including at least a memory and a processor. The memory stores a computer program, and the processor, when executing the computer program in the memory, performs the following method steps:
[0114] Step 1: Use a pre-set machine learning model to preprocess the raw financial data to obtain the target financial data;
[0115] Step 2: Based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to the different product objects, perform data slicing on the target financial data to obtain each financial dataset;
[0116] Step 3: Use a pre-trained deep learning network model to predict financial indicators for each of the financial datasets, and obtain the initial financial indicator values corresponding to each of the financial datasets.
[0117] Step 4: Conduct a comprehensive evaluation of the initial financial indicator values to obtain the target financial indicator values.
[0118] The specific implementation process of the above method steps can be found in the embodiments of the above-mentioned financial indicator prediction methods, which will not be repeated here.
[0119] This application preprocesses raw financial data using a pre-set machine learning model to obtain target financial data. By replacing traditional numerical methods with the pre-set machine learning model, the prediction accuracy of financial indicators is improved, reducing calculation errors caused by data anomalies. Automated anomaly detection and repair reduce manual intervention, improve system stability, and ensure the timeliness and reliability of financial indicator calculations. Based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to those product objects, the target financial data is sliced to obtain various financial datasets. A pre-trained deep learning network model is used to predict financial indicators for each of the financial datasets, obtaining initial financial indicator values corresponding to each dataset. Based on distributed computing methods and multi-threaded optimization, data processing speed is significantly improved to meet real-time business needs. A comprehensive evaluation of each initial financial indicator value yields the target financial indicator value. This application supports massive data processing and complex business scenarios, significantly improves system scalability, and is suitable for future business growth needs.
[0120] The above embodiments are merely exemplary embodiments of this application and are not intended to limit this application. The scope of protection of this application is defined by the claims. Those skilled in the art can make various modifications or equivalent substitutions to this application within its substance and scope of protection, and such modifications or equivalent substitutions should also be considered to fall within the scope of protection of this application.
Claims
1. A method for predicting financial indicators, characterized in that, include: The raw financial data is preprocessed using a pre-defined machine learning model to obtain the target financial data. Based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to the different product objects, the target financial data is sliced to obtain each financial dataset. A pre-trained deep learning network model is used to predict financial indicators for each of the financial datasets, thereby obtaining initial financial indicator values corresponding to each of the financial datasets. A comprehensive evaluation of the initial financial indicator values is conducted to obtain the target financial indicator values.
2. The method as described in claim 1, characterized in that, The process of preprocessing the raw financial data using a preset machine learning model to obtain the target financial data specifically includes: Anomaly processing is performed on the original financial data using a preset neural network model to obtain the first financial data. The Lasso regression model is used to extract features from the first financial data to obtain the second financial data; The first financial data is filtered for features using a random forest model to obtain the third financial data. The second financial data and the third financial data are merged to obtain the target financial data.
3. The method as described in claim 1, characterized in that, The method involves slicing the target financial data based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to different product objects to obtain various financial datasets, specifically including: The calculation process is performed based on the calculation rules corresponding to different product objects, the data volume of the target financial data corresponding to the different product objects, and the preset data volume slicing threshold to obtain the number of slices for the target financial data. When the number of slices is less than or equal to the preset number of threads, the target financial data is processed by parallel data slicing to obtain each financial dataset.
4. The method as described in claim 3, characterized in that, The method further includes: When the number of slices is greater than the preset number of threads, the preset data volume slice threshold is dynamically updated according to the calculation rules and the preset number of threads with the goal of maximizing resource utilization. Based on the updated preset data volume slicing threshold, parallel data slicing processing is performed on the target financial data to obtain each financial dataset.
5. The method as described in claim 1, characterized in that, The step of using a pre-trained deep learning network model to predict financial indicators for each of the financial datasets, and obtaining initial financial indicator values corresponding to each of the financial datasets, specifically includes: A target neural network using a pre-trained deep learning network model is used to predict financial indicators for each of the financial datasets, thereby obtaining the initial financial indicator values corresponding to each of the financial datasets.
6. The method as described in claim 1, characterized in that, Before using a pre-trained deep learning network model to predict financial indicators for each of the financial datasets, the method further includes: constructing a deep learning network model; The construction of the deep learning network model specifically includes: Obtain multiple historical cash flow data sequences; The Newton-Raphson iteration method is used to calculate and process each of the historical cash flow data sequences to obtain historical financial indicators corresponding to each of the historical cash flow data sequences. Label data is obtained by annotating each of the historical cash flow data sequences based on the historical financial indicators. The initial deep neural network is trained using the labeled data as training samples to obtain the target neural network that meets the model convergence condition. The parameters of the target neural network are optimized using a reinforcement learning Q-network based on a preset reward function to obtain the deep learning network model.
7. The method as described in claim 1, characterized in that, The process of comprehensively evaluating each of the initial financial indicator values to obtain the target financial indicator value specifically includes: The target financial indicator value is obtained by performing a weighted average calculation based on the initial financial indicator value corresponding to each product object and the weight coefficient corresponding to each product object.
8. A financial indicator prediction device, characterized in that, include: The preprocessing module is used to preprocess the raw financial data using a preset machine learning model to obtain the target financial data. The data slicing module is used to slice the target financial data based on the calculation rules corresponding to different product objects and the data volume of the target financial data corresponding to the different product objects, so as to obtain each financial dataset. The indicator prediction module is used to use a pre-trained deep learning network model to predict financial indicators for each of the financial datasets, and obtain the initial financial indicator values corresponding to each of the financial datasets. The comprehensive evaluation module is used to comprehensively evaluate the initial financial indicator values to obtain the target financial indicator values.
9. A storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the steps of the financial indicator prediction method according to any one of claims 1-7.
10. A computer device, characterized in that, It includes at least a memory and a processor, wherein the memory stores a computer program, and the processor, when executing the computer program in the memory, implements the steps of the financial indicator prediction method according to any one of claims 1-7.