CN118917551A - Financial data analysis system and method - Google Patents

Abstract

The invention discloses a financial data analysis system and a financial data analysis method, which relate to the technical field of financial data analysis, wherein the system comprises the following components: the system can continuously adjust parameters and structures of the prediction model according to market change by introducing an artificial intelligent self-optimization mechanism, deeply analyze investment, financing, income, cost, tax and profit distribution six basic financial data, provide comprehensive and accurate financial condition assessment for enterprises, improve the accuracy and timeliness of financial prediction, enable the model to adapt to market change in real time by the self-optimization mechanism, maintain prediction accuracy, and ensure the enterprises to respond to market dynamics rapidly and make timely adjustment by introducing an artificial intelligent self-optimization mechanism.

Description

A financial data analysis system and method

Technical Field

The present invention relates to the technical field of financial data analysis, and in particular to a financial data analysis system and method.

Background Art

In today's economic environment, the analysis of a company's financial status is crucial for decision makers. Accurate financial forecasts can help companies rationally plan resources and effectively avoid risks, thereby maintaining their advantages in the fierce market competition. However, financial data analysis is a complex process that involves the processing and analysis of large amounts of data, as well as the prediction of the company's future financial status.

However, in terms of financial forecasting models, existing technologies use data analysis and statistical methods, which have poor adaptability to data and cannot flexibly respond to complex patterns and dynamic changes in financial data. When the market environment, internal corporate structure or policies and regulations change significantly, it is difficult to quickly adjust and optimize. Secondly, existing self-optimization mechanisms are often not smart and automated enough, and do not make full use of new data. They cannot mine the potential characteristics of financial data from new data information in real time, and cannot provide accurate and real-time financial forecasts. This makes companies face the risk of outdated and erroneous information when making important decisions.

To sum up, the existing financial analysis technology shows obvious limitations and inadaptability when dealing with the rapidly changing market environment and the complex financial situation of enterprises. Therefore, developing an intelligent analysis system that can automatically learn, self-optimize and efficiently integrate multi-source financial data has become an important issue that needs to be urgently solved in the current financial management field.

Summary of the invention

The purpose of the present invention is to make up for the shortcomings of the existing technology and provide a financial data analysis system and method, which can automatically collect and process multi-source heterogeneous financial data, build a high-precision prediction model, and adjust the optimization strategy in real time according to market changes, thereby providing more accurate and timely support for the financial management of the enterprise.

In order to solve the above technical problems, the present invention provides the following technical solutions: On the one hand, a financial data analysis system, the components of which include: a basic data collection and preprocessing module, an intelligent prediction model construction and optimization module, a relationship map construction and analysis module, a project capital investment and output analysis module, and a decision support and suggestion generation module;

The basic data collection and preprocessing module is used to integrate multi-source heterogeneous data, including internal data sources and external data sources of the enterprise, and clean, standardize and convert the collected data;

The intelligent prediction model construction and optimization module uses machine learning algorithms to build an initial prediction model, perform feature engineering, train and verify the model, and self-optimize the model through automatic feedback and adjustment mechanisms;

The relationship map construction and analysis module is used to construct financial entities including investment, financing, income, cost, tax, profit distribution and their associated information, and to mine hidden data associations and potential risk points;

The project capital input and output analysis module is used to combine the output results of the intelligent prediction model with the relevant data in the relationship map, calculate the key indicators of the input-output ratio, and perform sensitivity analysis and risk warning;

The decision support and suggestion generation module automatically generates personalized financial decision suggestions based on the analysis results, including fund allocation strategies, cost control measures, and tax planning schemes, and provides an interactive interface to support users to adjust assumptions according to actual needs and conduct scenario simulation analysis.

Furthermore, the basic data collection and preprocessing module collects multi-source heterogeneous data, wherein the multi-source data respectively include:

External data sources include, but are not limited to, public databases, industry reports, and policy documents;

Internal data sources collect six types of basic financial data including investment, financing, income, cost, tax and profit distribution from the company's internal system.

Furthermore, the specific steps of the intelligent prediction model construction and optimization module are:

Model selection: Analyze the characteristics of financial data and select machine learning algorithms based on the forecasting objectives, i.e., predicting the trend of revenue and assessing the risk of profits. Perform a preliminary analysis of the financial data, i.e., determine the type of problem, and conduct time series forecasting, i.e., predict the future trend of revenue, i.e., determine the financial risk level of the enterprise, and predict the profit amount.

Feature Engineering: For six types of basic financial data, PCA, a feature dimension reduction method, is used to screen and optimize the extracted features, highlighting the features that have an important impact on the prediction results;

Model training and validation: Using cross validation, the dataset containing six types of basic financial data is divided into a training set and a test set. The model is trained using the training set. The model parameters and model structure are adjusted to optimize the model's prediction accuracy and generalization ability in investment, financing, income, cost, taxation, and profit distribution. The model performance is validated using the test set.

Self-optimization mechanism: Continuously receive new basic financial data on investment, financing, income, cost, tax, and profit distribution, and input them into the trained model for prediction, calculate the model's performance indicators, and maintain accurate predictions for the six types of basic financial data.

Furthermore, the intelligent prediction model construction and optimization module uses the linear regression technology in the machine learning algorithm to define the model for data with obvious time series characteristics, namely, income and cost, that is, y _i =β ₀ +β ₁ X ₁ +β ₂ X ₂ +…+β _{n Xi} +∈, where y _i represents the predicted target variable, namely, income and cost, X ₁ to _Xi are independent variables, including but not limited to time, previous income and cost values, and related economic indicators, β ₀ is the intercept, β ₁ to β _n are regression coefficients, ∈ is the error term, and the mean square error is used as the loss function to evaluate the difference between the model prediction value and the true value, and the formula is: Where m is the number of samples, _yi is the true value, To predict the model value, use the test set data to evaluate the trained income and cost models, compare the performance indicators of different models, and apply the selected optimal model to the new time data to predict the income and cost.

Furthermore, the intelligent prediction model construction and optimization module uses the feature dimension reduction method PCA for six types of basic financial data to standardize the feature data so that each feature has zero mean and unit variance, that is, the original feature matrix is X, where each row represents a sample and each column represents a feature. The standardized feature matrix is recorded as X ^′ , and the standardization formula is: Where μ _j is the mean of the j-th column feature, σ _j is the standard deviation of the j-th column feature, and the covariance matrix C of the standardized feature matrix is calculated, that is, Where n is the number of samples. The covariance matrix C is decomposed by eigenvalue to obtain the eigenvalue λ _i and the corresponding eigenvector _vi . The eigenvalues are arranged in descending order, and the eigenvectors corresponding to the first k eigenvalues are selected to form the projection matrix P. The original derived feature matrix X ^′ is multiplied by the projection matrix P to obtain the reduced dimension feature matrix Y = X ^′ P, thereby achieving feature dimensionality reduction. Through feature importance evaluation, the key features that have a significant impact on the prediction results in the investment, financing, income, cost, tax, and profit distribution data are determined.

Furthermore, the specific steps of the relationship map construction and analysis module are:

Define entity types: clearly define investment, financing, income, cost, tax and profit distribution as types of financial entities, and define relevant attributes for each type, i.e., investment amount and term, financing method and interest rate, income source and time, cost type and amount, tax type and tax rate, profit distribution ratio and object;

Determine the relationship type: define the relationship between financial entities, i.e. the connection between investment and income, between financing and costs, and between income and profit distribution;

Collect and organize data: Obtain processed financial data from the basic data collection and preprocessing module and organize it according to the defined entity and relationship types;

Use graph database storage: Use a graph database to create nodes to represent financial entities, edges to represent the relationships between entities, and add corresponding attribute data to nodes and edges;

Application of graph analysis algorithm: Calculate the importance score of each financial entity. For investment projects, evaluate their importance in the entire financial system based on their connection relationship and strength with other entities.

Furthermore, the relationship graph construction and analysis module uses the PageRank algorithm in the graph analysis algorithm to assign an initial PageRank value to each financial entity node. For each node i, the update formula of its PageRank value PR(j) is: d is the damping coefficient, M(i) is the set of nodes pointing to node i, L(j) is the number of links pointed out by node j, for investment nodes, M(i) is the other financial nodes associated with it, namely income nodes and cost nodes, and L(j) is the number of links pointed out by these related nodes.

Furthermore, the project capital investment and output analysis module obtains the data of revenue forecast and cost forecast from the intelligent forecast model construction and optimization module, obtains the relationship data related to investment, financing, revenue, cost, tax and profit distribution from the relationship map construction and analysis module, associates and integrates the output results of the intelligent forecast model with the relevant data in the relationship map, establishes a unified analysis data set, calculates the input-output ratio ROI according to the predicted revenue data and the actual revenue situation, as well as the predicted cost data and the actual cost input, that is, clarifies the calculation scope of total revenue and total cost, the total revenue includes the sales revenue of products and services, investment income and all sources of revenue related to the project, the total cost covers direct cost, indirect cost, fixed cost and variable cost, and summarizes the various sources of revenue, that is, there are n sources of revenue, namely I ₁ , I ₂ ,…, I _n , then the total revenue TI is: Add up all kinds of costs, that is, there are m types of cost categories, namely C ₁ , C ₂ ,…, C _m , then the total cost TC is: The calculation formula for ROI is: Predict the net cash flow of the project in the future periods, including cash inflows and cash outflows. That is, the forecast period is T period, the net cash flow in period t is CF _t , and the net present value NPV is calculated, that is, r is the estimated value of internal rate of return. The value of internal rate of return r is adjusted through iteration so that the net present value NPV approaches zero, which is the required internal rate of return of the project.

On the other hand, a financial data analysis method, the specific steps of the method are:

S100, basic data collection and preprocessing: used to integrate multi-source heterogeneous data, including internal and external data sources, and clean, standardize and convert the collected data;

S200, intelligent prediction model construction and optimization: use machine learning algorithms to build initial prediction models, perform feature engineering, train and validate models, and self-optimize models through automatic feedback and adjustment mechanisms;

S300, relationship map construction and analysis: used to construct financial entities and their associated information including investment, financing, income, cost, tax, profit distribution, and to mine hidden data associations and potential risk points;

S400, project capital input and output analysis: used to combine the output results of the intelligent prediction model with the relevant data in the relationship map, calculate the key indicators of the input-output ratio, and conduct sensitivity analysis and risk warning;

S500, decision support and suggestion generation: Based on the analysis results, it automatically generates personalized financial decision suggestions, including fund allocation strategies, cost control measures, and tax planning schemes. It provides an interactive interface to support users to adjust assumptions according to actual needs and conduct scenario simulation analysis.

Furthermore, the scenario simulation analysis in S500 designs an input interface for scenario simulation, inputs different market condition parameters, namely, market growth rate, degree of competition, price fluctuation range, and calls the prediction model according to the input conditions to simulate the financial results under different decision paths, including changes in key indicators of revenue, cost, profit, and cash flow, and presents the simulation results in the form of a data table to clearly and intuitively present the potential risks and benefits of different decision strategies.

Compared with the prior art, this financial data analysis system and method has the following beneficial effects:

1. By introducing an artificial intelligence self-optimization mechanism, this system can continuously adjust the parameters and structure of the forecasting model according to market changes, and deeply analyze six types of basic financial data: investment, financing, income, cost, tax and profit distribution, providing enterprises with a comprehensive and accurate financial status assessment, thereby improving the accuracy and timeliness of financial forecasts. Through the self-optimization mechanism, the model can adapt to market changes in real time and maintain forecast accuracy. This real-time performance ensures that enterprises can respond quickly to market dynamics and make timely adjustments.

2. The present invention provides scientific and reasonable financial decision-making basis for enterprises through comprehensive data analysis and prediction, and supports more refined decision-making through precise prediction data provided by intelligent prediction model construction and optimization module, combined with correlation analysis in relationship map construction and analysis module. Enterprises can carry out resource allocation and strategic planning based on more comprehensive and in-depth information, thereby improving overall competitiveness.

Other advantages, objectives and features of the present invention will be set forth in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be taught from the practice of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

FIG1 is an operation flow chart of a financial data analysis system.

FIG. 2 is a flow chart of a financial data analysis method.

DETAILED DESCRIPTION

The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. 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.

Embodiment 1

This embodiment provides a financial data analysis system, which, through a self-optimizing financial forecasting model based on artificial intelligence, can deeply mine the potential patterns and associations in the data, and continuously adjust the model parameters according to new data and market changes, thereby significantly improving the accuracy of financial forecasts and providing enterprises with more reliable decision-making basis.

In the specific implementation, first, various data from inside and outside the enterprise are collected and pre-processed to facilitate subsequent analysis. External data sources include but are not limited to public databases, industry reports, and policy documents. These data help to understand industry trends, competitor conditions, and the macroeconomic environment. Internal data sources are data collected from the internal system of the enterprise, mainly including six types of basic financial data: investment, financing, income, cost, tax, and profit distribution. Check and correct or delete errors, missing or outliers in the data, identify logical errors in the data or records that do not comply with the rules, fill in missing data points or mark them as unavailable, identify and process data points that are out of the normal range, ensure that all data are on the same scale for comparison and analysis, convert non-numeric data into numerical form, such as encoding classified data into numerical values, convert data in different units into the same unit, such as unifying the monetary unit into RMB, output the processed data to the next processing stage or store it in a data warehouse/database, store the data in a suitable data warehouse or database, and provide a data interface for subsequent processing stages.

Then, the initial prediction model is constructed using machine learning algorithms to analyze the characteristics of financial data, including the time series characteristics, distribution of data, and correlation between variables. For data with obvious time series characteristics (such as income, cost, etc.), the linear regression technology in the machine learning algorithm is used to define the model, that is, y _i = β ₀ + β ₁ X ₁ + β ₂ X ₂ + … + β _{n Xi} + ∈. The trained income and cost models are evaluated using the test set data, and the performance indicators of different models are compared. The mean square error is used as the loss function to evaluate the difference between the model prediction value and the true value. The formula is: Apply the selected optimal model to the new time data to predict income and cost. Calculate the derived features for the six basic financial data (investment, financing, income, cost, tax, and profit distribution), such as investment growth rate, financing volatility, income trend line, cost volatility, tax change rate, and profit distribution ratio. Use feature dimension reduction methods, such as PCA, to standardize the feature data so that each feature has zero mean and unit variance. That is, the original feature matrix is X, where each row represents a sample and each column represents a feature. The standardized feature matrix is recorded as X ^′ , and the standardization formula is: Where μ _j is the mean of the j-th column feature, σ _j is the standard deviation of the j-th column feature, and the covariance matrix C of the standardized feature matrix is calculated, that is, Where n is the number of samples. The covariance matrix C is decomposed by eigenvalues to obtain eigenvalues λ _i and corresponding eigenvectors _vi . The eigenvalues are arranged in descending order, and the eigenvectors corresponding to the first k eigenvalues are selected to form a projection matrix P. The original derived feature matrix X ^′ is multiplied by the projection matrix P to obtain the reduced feature matrix Y = X ^′ P, thereby achieving feature dimensionality reduction. Through feature importance evaluation, the key features that have a significant impact on the prediction results in the investment, financing, income, cost, tax, and profit distribution data are determined. New basic financial data on investment, financing, income, cost, tax, and profit distribution are continuously received and input into the trained model for prediction. The performance indicators of the model, such as mean square error, accuracy, recall, etc., are calculated to evaluate the model prediction performance. When the model does not reach the preset threshold in terms of prediction performance, the model retraining or parameter adjustment process is automatically triggered, so that the model can adapt to market changes in real time and maintain prediction accuracy.

Embodiment 2

This embodiment aims to elaborate on the specific application of a financial data analysis system, which integrates multi-source financial data, constructs a relationship map for comprehensive analysis, combines decision trees and rule engines to generate personalized suggestions, and allows users to intuitively evaluate the risks and benefits of different strategies through scenario simulation.

First, define investment, financing, income, cost, tax and profit distribution as types of financial entities. The attributes of investment entities include investment amount, investment period, investment project type, etc.; the attributes of income entities include income source, income time, income amount, etc. Define the relationship between financial entities, such as the relationship between investment and income (investment may bring income growth), the relationship between financing and cost (financing may generate costs), the relationship between income and profit distribution, etc. Obtain the processed relevant financial data from the basic data collection and preprocessing module, and organize it according to the defined entity and relationship types, collect investment project information, financing channels and amounts, income sources and cost structures of each subsidiary, use the graph database, create nodes to represent financial entities, create edges to represent the relationship between entities, and add corresponding attribute data to nodes and edges, create investment nodes and income nodes, and establish the association edges between them, and record the relevant attribute information. Use the PageRank algorithm to assign an initial PageRank value to each financial entity node. For each node i, the update formula of its PageRank value PR(j) is, that is, If the PageRank score of an investment project node is high, it means that it has a significant impact on the financial status of the enterprise. It can identify the financing channels or income source nodes that have a key influence on the flow of funds. It provides a wealth of visualization tools to present complex financial data relationships to users in an intuitive and easy-to-understand way, and display the relationships between entities such as investment, financing, and income, display the distribution of important financial indicators, and present the flow of funds between different entities. Users can quickly locate data points and relationships of interest through customized query and filtering conditions.

Then, obtain the revenue forecast, cost forecast and other data from the intelligent prediction model construction and optimization module, and obtain the relationship data related to investment, financing, revenue, cost, tax and profit distribution from the relationship map construction and analysis module. For example, obtain the revenue forecast data of a certain investment project and the relationship between the project and other financial entities, and clarify the calculation scope of total revenue and total cost. The total revenue includes all sources of revenue related to the project, such as sales revenue of products or services and investment income; the total cost covers direct costs (such as raw material procurement and labor costs), indirect costs (such as management expenses and equipment depreciation), fixed costs (such as rent and equipment purchase) and variable costs (such as raw material consumption and wages calculated based on output). Summarize the various sources of revenue. Assuming that there are n sources of revenue, namely I ₁ ,I ₂ ,…,I _n , the total revenue TI is: Add up all kinds of costs, that is, there are m types of cost categories, namely C ₁ , C ₂ ,…, C _m , then the total cost TC is: The calculation formula for ROI is: Predict the net cash flow of the project in the future periods, including cash inflows and cash outflows. That is, the forecast period is T period, the net cash flow in period t is CF _t , and the net present value NPV is calculated, that is, The value of the internal rate of return r is continuously adjusted through iteration so that the net present value NPV approaches zero, which is the required internal rate of return of the project. According to historical data and industry experience, the risk warning thresholds of the input-output ratio and the internal rate of return are set. When the ROI is lower than 10% or the internal rate of return is lower than the industry average, a risk warning is triggered. When the forecast result deviates from the expected range, an alarm is automatically issued to remind users to pay attention to potential risks, and an alarm message is automatically sent to relevant personnel to prompt possible risks.

Finally, collect and organize domain knowledge and historical experience, including previous financial decision-making cases, successful and failed project experiences, industry standards and best practices, analyze and refine the collected knowledge and experience, and transform them into clear decision-making rules and conditions. Based on these rules and conditions, define and manage rule sets, design scenario simulation input interfaces, and allow users to input different market condition parameters (such as market growth rate, competition intensity, price fluctuation range, etc.), policy environment parameters (such as tax policy changes, industry regulatory policy adjustments, etc.) and other assumptions that may affect financial decisions. Based on the assumptions entered by users, call corresponding models and algorithms to simulate financial results under different decision paths, including changes in key indicators such as revenue, cost, profit, and cash flow. Present the simulation results to users in a clear and intuitive manner, such as through charts, data tables, etc., to help users compare and evaluate the potential risks and benefits of different decision-making strategies, and plan the layout and functional divisions of the interface, including The data input area, result display area, and operation buttons enable users to easily configure analysis parameters by dragging and clicking, such as selecting financial indicators to be analyzed, setting time ranges, and selecting data sources. In the result display area, the analysis results are presented in real time, allowing users to quickly view key data and analysis conclusions. A report export function is provided, and support is provided for generating reports in multiple formats (such as PDF, Excel, etc.), which is convenient for users to share and discuss with team members or superiors. The results of financial analysis are input into the decision tree or rule engine, and reasoning and judgment are performed based on preset rules and conditions. Personalized financial decision recommendations are automatically generated, including investment strategies, financing plans, cost control measures, and profit distribution recommendations. For example, if the analysis results show that a project has high potential, but funds are currently tight, the system may recommend a reasonable financing plan to support the development of the project. The generated decision recommendations are evaluated for rationality and feasibility to ensure that they are in line with the actual situation and development goals of the enterprise.

To sum up, through the above implementation steps, we can use the relationship map construction and analysis module, project capital investment and output analysis module and decision support and suggestion generation module in the financial data analysis system to achieve in-depth analysis and mining of financial data, provide a scientific basis for the company's financial decision-making, optimize resource allocation, reduce risks, and enhance the company's competitiveness and sustainable development capabilities.

It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims (10)

1. A financial data analysis system, characterized in that the components of the system include: basic data collection and preprocessing module, intelligent prediction model construction and optimization module, relationship map construction and analysis module, project capital investment and output analysis module, decision support and suggestion generation module; The basic data collection and preprocessing module is used to integrate multi-source heterogeneous data, including internal data sources and external data sources of the enterprise, and clean, standardize and convert the collected data; The intelligent prediction model construction and optimization module uses machine learning algorithms to build an initial prediction model, perform feature engineering, train and verify the model, and self-optimize the model through automatic feedback and adjustment mechanisms; The relationship map construction and analysis module is used to construct financial entities including investment, financing, income, cost, tax, profit distribution and their associated information, and to mine hidden data associations and potential risk points; The project capital input and output analysis module is used to combine the output results of the intelligent prediction model with the relevant data in the relationship map, calculate the key indicators of the input-output ratio, and perform sensitivity analysis and risk warning; The decision support and suggestion generation module automatically generates personalized financial decision suggestions based on the analysis results, including fund allocation strategies, cost control measures, and tax planning schemes, and provides an interactive interface to support users to adjust assumptions according to actual needs and conduct scenario simulation analysis. 2. A financial data analysis system according to claim 1, characterized in that the basic data collection and preprocessing module collects multi-source heterogeneous data, wherein the multi-source data respectively include: External data sources include, but are not limited to, public databases, industry reports, and policy documents; Internal data sources collect six types of basic financial data including investment, financing, income, cost, tax and profit distribution from the company's internal system. 3. According to claim 1, a financial data analysis system is characterized in that the specific steps of the intelligent prediction model construction and optimization module are: Model selection: Analyze the characteristics of financial data and select machine learning algorithms based on the forecasting objectives, i.e., predicting the trend of revenue and assessing the risk of profits. Perform a preliminary analysis of the financial data, i.e., determine the type of problem, and conduct time series forecasting, i.e., predict the future trend of revenue, i.e., determine the financial risk level of the enterprise, and predict the profit amount. Feature Engineering: For six types of basic financial data, PCA, a feature dimension reduction method, is used to screen and optimize the extracted features, highlighting the features that have an important impact on the prediction results; Model training and validation: Using cross validation, the dataset containing six types of basic financial data is divided into a training set and a test set. The model is trained using the training set. The model parameters and model structure are adjusted to optimize the model's prediction accuracy and generalization ability in investment, financing, income, cost, taxation, and profit distribution. The model performance is validated using the test set. Self-optimization mechanism: Continuously receive new basic financial data on investment, financing, income, cost, tax, and profit distribution, and input them into the trained model for prediction, calculate the model's performance indicators, and maintain accurate predictions for the six types of basic financial data. 4. A financial data analysis system according to claim 3, characterized in that the intelligent prediction model construction and optimization module uses the linear regression technology in the machine learning algorithm to define the model for data with obvious time series characteristics, namely, income and cost, that is, y _i = β ₀ + β ₁ X ₁ + β ₂ X ₂ + ... + β _{n Xi} + ∈, wherein y _i represents the target variable for prediction, namely, income and cost, X ₁ to _Xi are independent variables, including but not limited to time, previous income and cost values, and related economic indicators, β ₀ is the intercept, β ₁ to β _n are regression coefficients, ∈ is the error term, and the mean square error is used as the loss function for evaluating the difference between the predicted value and the true value of the model, and the formula is: Where m is the number of samples, _yi is the true value, To predict the model value, use the test set data to evaluate the trained income and cost models, compare the performance indicators of different models, and apply the selected optimal model to the new time data to predict the income and cost. 5. According to claim 3, a financial data analysis system is characterized in that the intelligent prediction model construction and optimization module uses the feature dimension reduction method PCA for six types of basic financial data to standardize the feature data so that each feature has zero mean and unit variance, that is, the original feature matrix is X, where each row represents a sample and each column represents a feature. The standardized feature matrix is recorded as X ^′ , and the standardization formula is: Where μ _j is the mean of the j-th column feature, σ _j is the standard deviation of the j-th column feature, and the covariance matrix C of the standardized feature matrix is calculated, that is, Where n is the number of samples. The covariance matrix C is decomposed by eigenvalue to obtain the eigenvalue λ _i and the corresponding eigenvector _vi . The eigenvalues are arranged in descending order, and the eigenvectors corresponding to the first k eigenvalues are selected to form the projection matrix P. The original derived feature matrix X ^′ is multiplied by the projection matrix P to obtain the reduced dimension feature matrix Y = X ^′ P, thereby achieving feature dimensionality reduction. Through feature importance evaluation, the key features that have a significant impact on the prediction results in the investment, financing, income, cost, tax, and profit distribution data are determined. 6. A financial data analysis system according to claim 1, characterized in that the specific steps of the relationship map construction and analysis module are: Define entity types: clearly define investment, financing, income, cost, tax and profit distribution as types of financial entities, and define relevant attributes for each type, i.e., investment amount and term, financing method and interest rate, income source and time, cost type and amount, tax type and tax rate, profit distribution ratio and object; Determine the relationship type: define the relationship between financial entities, i.e. the connection between investment and income, between financing and costs, and between income and profit distribution; Collect and organize data: Obtain processed financial data from the basic data collection and preprocessing module and organize it according to the defined entity and relationship types; Use graph database storage: Use a graph database to create nodes to represent financial entities, edges to represent the relationships between entities, and add corresponding attribute data to nodes and edges; Application of graph analysis algorithm: Calculate the importance score of each financial entity. For investment projects, evaluate their importance in the entire financial system based on their connection relationship and strength with other entities. 7. A financial data analysis system according to claim 6, characterized in that the relationship graph construction and analysis module uses the PageRank algorithm in the graph analysis algorithm to assign an initial PageRank value to each financial entity node. For each node i, the update formula of its PageRank value PR(j) is, that is, d is the damping coefficient, M(i) is the set of nodes pointing to node i, L(j) is the number of links pointed out by node j, for investment nodes, M(i) is the other financial nodes associated with it, namely income nodes and cost nodes, and L(j) is the number of links pointed out by these related nodes. 8. According to claim 1, a financial data analysis system is characterized in that the project capital input and output analysis module obtains the data of income forecast and cost forecast from the intelligent forecast model construction and optimization module, obtains the relationship data related to investment, financing, income, cost, tax, and profit distribution from the relationship map construction and analysis module, associates and integrates the output results of the intelligent forecast model with the relevant data in the relationship map, establishes a unified analysis data set, calculates the input-output ratio ROI according to the predicted income data and the actual income situation, as well as the predicted cost data and the actual cost input, that is, clarifies the calculation scope of total income and total cost, the total income includes the sales income of products and services, investment income and all income sources related to the project, the total cost covers direct cost, indirect cost, fixed cost and variable cost, and summarizes each income source, that is, there are n income sources, namely I ₁ , I ₂ , ..., I _n , then the total income TI is: Add up all kinds of costs, that is, there are m types of cost categories, namely C ₁ , C ₂ ,…, C _m , then the total cost TC is: The calculation formula for ROI is: Predict the net cash flow of the project in the future periods, including cash inflows and cash outflows. That is, the forecast period is T period, the net cash flow in period t is CF _t , and the net present value NPV is calculated, that is, r is the estimated value of internal rate of return. The value of internal rate of return r is adjusted through iteration so that the net present value NPV approaches zero, which is the required internal rate of return of the project. 9. A financial data analysis method, characterized in that the method comprises the following specific steps: S100, basic data collection and preprocessing: used to integrate multi-source heterogeneous data, including internal and external data sources, and clean, standardize and convert the collected data; S200, intelligent prediction model construction and optimization: use machine learning algorithms to build initial prediction models, perform feature engineering, train and validate models, and self-optimize models through automatic feedback and adjustment mechanisms; S300, relationship map construction and analysis: used to construct financial entities and their associated information including investment, financing, income, cost, tax, profit distribution, and to mine hidden data associations and potential risk points; S400, project capital input and output analysis: used to combine the output results of the intelligent prediction model with the relevant data in the relationship map, calculate the key indicators of the input-output ratio, and conduct sensitivity analysis and risk warning; S500, decision support and suggestion generation: Based on the analysis results, it automatically generates personalized financial decision suggestions, including fund allocation strategies, cost control measures, and tax planning schemes. It provides an interactive interface to support users to adjust assumptions according to actual needs and conduct scenario simulation analysis. 10. A financial data analysis method according to claim 9, characterized in that the scenario simulation analysis in S500 designs an input interface for scenario simulation, inputs different market condition parameters, namely, market growth rate, degree of competition, price fluctuation range, and according to the input conditions, calls the prediction model to simulate the financial results under different decision paths, including changes in key indicators of revenue, cost, profit, and cash flow, and presents the simulation results in the form of a data table to clearly and intuitively present the potential risks and benefits of different decision strategies.