A financial investment decision assistance method and system based on game theory and multi-source big data

Abstract

This invention belongs to the field of financial investment decision-making technology, specifically a financial investment decision-making assistance method and system based on game theory and multi-source big data. It collects and processes multi-source financial data, constructs a unified data foundation, and uses machine learning to generate multi-dimensional market prediction signals based on this foundation. These prediction signals are then input into a game theory model to simulate the behavior of market participants, solve for equilibrium strategies, and generate investment or risk control suggestions based on the game results, which are then visualized. This invention overcomes the limitations of single data sources by integrating multi-source big data and generating multimodal prediction signals. By combining game theory to model the complex interactive behaviors of market participants, it can deduce market evolution trends from a global perspective, thereby assisting in the formulation of more forward-looking investment or risk control strategies. It can simulate information asymmetry, strategy interaction, and dynamic evolution processes in the market, thus providing a better understanding of market fluctuations and the impact of sudden events.

Description

Technical Field

[0001] This invention belongs to the field of financial investment decision-making technology, and relates to a financial investment decision-making assistance method and system based on game theory and multi-source big data. Background Technology

[0002] Financial investment decision support technology based on game theory and multi-source big data is an interdisciplinary technology that integrates artificial intelligence, big data analysis, game theory modeling, and financial engineering. This technology aims to provide more comprehensive, dynamic, and intelligent support for investment decisions by integrating multi-source heterogeneous data (including traditional financial data, alternative data, and macro data) and using game theory to model the behavior of market participants.

[0003] Existing technologies disclose several invention patents in the field of financial investment decision-making. Among them, invention patent CN117649304A discloses an auxiliary analysis method for financial transactions. This method addresses the problem in insurance transactions where different consumers have varying levels of understanding of insurance products, making it difficult to make informed decisions. By analyzing the target individual's historical information, the method determines their level of understanding of insurance products, resulting in signals indicating low, medium, and high levels of understanding. Based on these signals, the method analyzes and processes the status information of various insurance products, detailed data on insurance products favored by the target individual, and the status information of professional service personnel in the insurance transaction. This process generates corresponding instructions and enables the execution of corresponding operations.

[0004] Financial investment decisions have long faced challenges such as high market uncertainty, incomplete information, and complex strategic interactions among participants. Traditional investment decision support methods mainly rely on historical data analysis and technical indicators, which are difficult to fully capture the nonlinear dynamics and volatile behavior of the market. In recent years, with the development of big data and artificial intelligence technologies, investment strategies based on machine learning prediction models (such as hidden Markov models and neural networks) have emerged. These methods have improved prediction accuracy to some extent, but are still mainly limited to data-driven pattern recognition. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, this invention proposes a financial investment decision-making support method and system based on game theory and multi-source big data. This invention primarily addresses the problem that traditional investment decision-making support methods rely heavily on historical data analysis and technical indicators, making it difficult to comprehensively capture the nonlinear dynamics and volatile behavior of the market. In recent years, with the development of big data and artificial intelligence technologies, investment strategies based on machine learning prediction models (such as Hidden Markov Models and neural networks) have emerged. These methods have improved prediction accuracy to some extent, but are still mainly limited to data-driven pattern recognition problems.

[0006] This invention provides a financial investment decision support method based on game theory and multi-source big data, comprising:

[0007] Data fusion steps: Collect multi-source financial data, including financial market data, alternative big data, and macro and industry data; perform natural language processing on the multi-source financial data to transform it into a structured feature set; and perform data cleaning and alignment to build a unified data foundation.

[0008] Signal generation steps: Based on the unified data foundation, a multimodal prediction signal is generated through a machine learning model. The multimodal prediction signal includes sentiment index, event impact assessment, fund flow identification, correlation network analysis, and mobile big data-specific signals.

[0009] Game deduction steps: Use the multimodal prediction signal as the input parameter of the game model, abstractly model the market participants and solve the game equilibrium, and output the optimal strategy probability distribution of each participant;

[0010] Decision application steps: Based on the solution of the game equilibrium, generate investment decision strategy signals or risk warning signals, and display them through a visual interface.

[0011] Preferably, the data fusion step further includes:

[0012] The multi-source financial data is stored using a two-tier storage architecture that combines a time-series database and a data lake.

[0013] Establish a hierarchical and categorized data management mechanism to trace the source, control access, and conduct compliance audits throughout the entire data lifecycle.

[0014] Preferably, the mobile big data-specific signal generated in the signal generation step includes:

[0015] An industry prosperity index based on the intensity of employees and the activity level of drivers in the chemical industry;

[0016] Enterprise operating rate and inventory index inferred from factory personnel and truck traffic;

[0017] The regional economic activity index is constructed based on population flow and freight volume, and includes a resumption of work index and a consumption activity index.

[0018] Preferably, the signal generation step further includes a credit scoring model optimization process:

[0019] A credit scoring model is constructed using deep learning or ensemble learning algorithms, and the model is iteratively trained and evaluated; the credit scoring model is used to score the creditworthiness of individuals or enterprises; this score serves as one dimension of a multimodal prediction signal.

[0020] Introduce explainable AI technology to explain the model's decision-making process;

[0021] Establish a self-learning mechanism for the model to absorb new data in real time and dynamically adjust the scoring logic.

[0022] Preferably, in the game theory deduction step, the market participants include at least four types of entities: institutional investors, retail investors, industrial capital, and regulators. The game model includes multiple models such as incomplete information game model, evolutionary game model, corporate finance and industrial investment game model, Stackelberg game model, and cooperative game model. The corporate finance and industrial investment game model constructed in the game theory deduction step takes enterprises as game participants, sets unit production cost, financing capacity, and financial market return rate as model parameters, and solves the optimal allocation ratio between financial investment and industrial investment by enterprises through Nash equilibrium.

[0023] Preferably, the decision application step further includes:

[0024] It provides a scenario simulator that supports user-defined events and can deduce market reactions and optimal strategies based on game theory models.

[0025] When the game theory model predicts that the market is evolving toward a high-risk equilibrium, a risk warning is issued in advance. The risk warning includes the risk of excessive financialization of enterprises or insufficient investment in real industries.

[0026] This invention also provides a financial investment decision support system based on game theory and multi-source big data. This system is used to implement the aforementioned financial investment decision support method based on game theory and multi-source big data. The financial investment decision support system includes:

[0027] The data layer module is used to collect multi-source financial data, including financial market data, alternative big data, and macro and industry data. Natural language processing is performed on the multi-source financial data to transform it into a structured feature set, and data cleaning and alignment are performed to build a unified data foundation.

[0028] The analysis layer module, connected to the data layer module, is used to generate multimodal prediction signals based on the unified data foundation through machine learning models. The multimodal prediction signals include sentiment index, event impact assessment, fund flow identification, correlation network analysis, and mobile big data-specific signals.

[0029] The game layer module, connected to the analysis layer module, is used to take the multimodal prediction signal as the input parameter of the game model, abstractly model the market participants and solve the game equilibrium, and output the optimal strategy probability distribution of each participant. The market participants include at least four types of entities: institutional investors, retail investors, industrial capital and regulators. The game model includes multiple types of models such as incomplete information game model, evolutionary game model, corporate finance and real investment game model, Stackelberg game model and cooperative game model.

[0030] The application layer module, connected to the game layer module, is used to generate investment decision strategy signals or risk warning signals based on the solution results of the game equilibrium, and to display them through a visual interface.

[0031] Preferably, the corporate finance and real investment game model constructed in the game layer module takes the enterprise as the game participant, sets the unit production cost, financing capacity and financial market return rate as model parameters, and solves the optimal ratio allocation between financial investment and real investment by the enterprise through Nash equilibrium.

[0032] Preferably, the application layer module further includes:

[0033] Scenario simulators are used to support user-defined events and deduce market reactions and optimal strategies based on game theory models.

[0034] The risk warning unit is used to issue early warnings of risks, including the risk of excessive financialization of enterprises or insufficient investment in real industries, when the game model predicts that the market is evolving towards a high-risk equilibrium.

[0035] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described financial investment decision support method based on game theory and multi-source big data.

[0036] The beneficial effects of this invention are as follows:

[0037] 1. In this invention, by integrating multi-source big data and generating multimodal prediction signals, the limitations of a single data source are overcome. By combining game theory to model the complex interactive behaviors of market participants, the market evolution trend can be deduced from a global perspective, thereby assisting in the formulation of more forward-looking investment or risk control strategies.

[0038] 2. In this invention, by introducing multiple game theory models, it is possible to simulate information asymmetry, strategic interaction and dynamic evolution in the market, rather than just statistical analysis based on historical data. This enables the system to capture the strategic behavior of market participants and its impact on asset prices, thereby better understanding market fluctuations and the impact of sudden events.

[0039] 3. In this invention, by constructing a game model of corporate finance and real investment, it is possible to analyze the trade-off between corporate returns in the financial market and real investment, and to predict the risks of "decoupling from the real economy" or insufficient real investment. This helps investors understand the transmission mechanism of macroeconomic policies to micro-corporate behavior and also supports regulatory agencies in identifying systemic risks.

[0040] 4. In this invention, the signal generation step combines a machine learning model and introduces interpretable AI technology, which makes the originally "black box" model prediction process more transparent. The game deduction step finally outputs the optimal strategy probability distribution of each participant, providing a relatively quantitative scientific basis for investment decisions or risk warnings.

[0041] 5. In this invention, the system can absorb new data in real time and dynamically adjust the scoring logic by establishing a self-learning mechanism for the model in order to adapt to the rapidly changing market environment. At the same time, the system can simulate market reactions under different events through scenario simulators, and issue early warnings of risks, including excessive financialization of enterprises, when the market evolves towards a high-risk equilibrium in the game theory deduction, thus realizing the transformation from passive response to active defense.

[0042] 6. In this invention, a two-tier storage architecture combining time-series databases and data lakes, along with a hierarchical and classified data management mechanism, is adopted. This not only solves the problem of efficient storage of multi-source heterogeneous data, but also ensures the traceability, access control, and compliance auditing of data throughout its entire lifecycle, providing underlying support for the robust operation of the system. Attached Figure Description

[0043] Figure 1 This is a flowchart of the method of the present invention;

[0044] Figure 2 This is the system architecture diagram of the present invention. Detailed Implementation

[0045] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0046] Example 1: Financial Anomalies and Risk Warnings for "Blue-Chip Stocks"

[0047] Background: The stock price of a well-known listed company (referred to as "Company A") has been growing steadily for a long time and is known in the market as a "blue-chip stock". Analysts hope to identify in advance the potential risks of financial fraud or deterioration of its fundamental business.

[0048] Implementation steps:

[0049] 1. Data fusion steps:

[0050] Financial market data: High-frequency data such as stock price, trading volume, volatility, and margin trading balance of Company A and its peers are collected;

[0051] Alternative Big Data:

[0052] Public opinion data: Scrape financial news, stock forum posts, and social media (such as Weibo and WeChat public accounts) discussions about Company A, and analyze sentiment and focus through natural language processing (NLP);

[0053] Mobile Big Data Special Signal: Obtain data on "factory pedestrian flow" and "truck flow" entering and leaving the factory area around Company A's main production base, construct enterprise operating rate and inventory index. If it is found that pedestrian flow and vehicle flow continue to decline, but this is inconsistent with the "booming production and sales" situation disclosed in the company's financial report, an abnormal signal is formed.

[0054] Related network data: Analyze changes in business registration information, legal proceedings, guarantee relationships, etc. between Company A and its subsidiaries, major customers, and suppliers to construct a related network;

[0055] Macroeconomic and industry data: Collect data such as the business climate index and upstream and downstream price changes of the industry in which Company A operates;

[0056] All data is cleaned, aligned, and stored in a unified data platform.

[0057] 2. Signal generation steps:

[0058] The above data is processed using machine learning models (such as LSTM and Transformer) to generate multimodal prediction signals:

[0059] Sentiment Index: Based on public opinion analysis, a negative sentiment index of the market towards Company A is generated;

[0060] Impact Assessment: Assess the potential impact of recent events such as executive share reductions and regulatory inquiries;

[0061] Fund Flow Identification: Identifying whether major funds (institutions) are continuously and covertly flowing out;

[0062] Network analysis revealed that a supplier associated with Company A was experiencing operational irregularities, raising concerns about potential risk transmission.

[0063] Mobile Big Data Special Signal: Based on the flow of people and trucks in the factory, output a special early warning signal that "the enterprise's operating rate deviates significantly from the expected rate".

[0064] 3. Game theory deduction steps:

[0065] Using the above signals as input, a game theory model is constructed, with participants including:

[0066] Institutional investors: It is known that some institutions may have discovered the problem through research or data. Their strategy is to "continue to hold" or "sell first".

[0067] Industrial capital: Will competitors or upstream and downstream companies in the same industry take the opportunity to acquire or adjust their cooperation strategies?

[0068] Regulators: The probability of regulatory agencies launching an investigation based on related networks and financial irregularities;

[0069] Retail investors: mainly influenced by public information and sentiment, their strategies are "buying on the sidelines" or "panic selling";

[0070] Using an incomplete information game model, we simulated a scenario where institutions have an information advantage and retail investors are at an information disadvantage. By solving the game equilibrium, we found that the market is very likely to evolve into a high-risk equilibrium where "institutions sell first, causing stock prices to fall, and retail investors then panic and sell off."

[0071] 4. Decision-making application steps:

[0072] Risk Warning: On the visual interface, the system issues a red warning to Company A for “excessive financialization of the enterprise based on game theory simulation / fundamental risk”. The warning signal indicates that there is a discrepancy between the company’s actual operating data (mobile big data) and financial reports, and the game model simulation results show that the stock price is at risk of a “stampede” decline due to information asymmetry.

[0073] Decision Recommendation: It is recommended to reduce the position of Company A in the portfolio, or to buy put options for hedging.

[0074] Example 2: Industry Investment Strategies under Regional Industrial Policy Adjustments

[0075] Background: A local government announced production restrictions and rectification measures for a chemical industrial park. Investors want to assess the impact of this policy on different chemical companies and find potential investment targets.

[0076] Implementation steps:

[0077] 1. Data fusion steps:

[0078] Financial market data: Stock prices and bond yields of chemical companies in the affected region;

[0079] Alternative Big Data:

[0080] Mobile Big Data Special Signal: Retrieve data on "employee intensity" and "driver activity" of various chemical enterprises in the affected area to construct an industry prosperity index. At the same time, construct a regional economic activity index by analyzing "population flow" and "freight situation" in the surrounding areas to assess the impact of production restrictions on local employment and logistics.

[0081] Event data: Collect texts of government announcements and press conferences, conduct NLP analysis, and extract key information such as policy strength, implementation period, and regulatory intensity;

[0082] Macroeconomic and industry data: national inventory, price trends, and import / export data for chemical products;

[0083] Build a unified data foundation;

[0084] 2. Signal generation steps:

[0085] Generate multimodal prediction signals:

[0086] Impact assessment: Quantitatively assess the degree of impact of production restriction policies on the total supply of regional chemical products;

[0087] Network analysis: Identify the position of affected chemical companies in the industrial chain, which downstream companies (such as coatings and plastic products) may be harmed by raw material shortages, and which unaffected competitors in other regions may benefit from them;

[0088] Mobile Big Data Special Signal: Dynamically monitor the "operational rate index" of affected enterprises to assess the actual impact on enterprises in real time, rather than relying solely on official estimates;

[0089] 3. Game theory deduction steps:

[0090] Construct a Stackelberg game model (leader-follower model);

[0091] Regulators (leaders): Take the lead and announce production restriction policies;

[0092] Industrial capital (chemical companies - followers): As participants in the game, companies need to decide on strategies such as "reducing production to maintain prices", "illegible production", "converting production" or "increasing production in subsidiaries in other locations".

[0093] Meanwhile, a game model of corporate finance and industrial investment is introduced: For affected companies, it is necessary to reassess the allocation of resources between "industrial investment" (such as upgrading environmental protection equipment to comply with new regulations) and "financial investment" (such as using idle funds to purchase wealth management products). The model uses unit production cost (which may increase due to production restrictions), financing capacity (banks may tighten credit due to production restrictions) and financial market return rate as parameters to solve the company's optimal strategy.

[0094] Decision application steps:

[0095] Investment decision strategy signals: Game theory results show that leading companies, with their advantages in capital and technology, are more likely to choose the "transformation and upgrading" strategy, thereby gaining a larger market share in the future, and their stock prices have the potential to rise. On the other hand, for some small companies with high debt, the game equilibrium solution points to high-risk strategies such as "illegal production" or "financial speculation".

[0096] Scenario Simulation: Users (investors) can use the scenario simulator to customize different scenarios such as "relaxation of policy implementation" or "unexpected decline in downstream demand," rerun the game model, observe the changes in the optimal strategies of each company, and thus formulate more robust portfolio strategies, such as buying stocks of leading companies while selling or avoiding bonds of high-risk small companies.

[0097] Example 3: Pricing and Trading Strategies on the First Day of New Stock Listing

[0098] Background: A technology innovation company is about to be listed on the Science and Technology Innovation Board. Due to the lack of historical trading data, traditional valuation methods are facing challenges. Quantitative investment fund managers hope to use this patented method to help formulate trading strategies for the first day of listing.

[0099] Implementation steps:

[0100] 1. Data fusion steps:

[0101] Financial market data: Valuation levels (price-to-earnings ratio, price-to-sales ratio) of comparable listed companies (in the same industry, with similar revenue scale), and the performance of recent newly listed stocks on their first day of trading (price change, turnover rate).

[0102] Alternative Big Data:

[0103] Public opinion data: Analyze prospectuses, roadshow materials, brokerage research reports, and financial media reports to extract market attention and consensus expectations for the company;

[0104] Mobile Big Data Special Signal: If the company has a To C business, its consumer activity index can be constructed using its APP downloads, user activity, e-commerce platform sales data, etc., to verify its business growth potential;

[0105] Network analysis: Analyze its shareholder structure (whether there are prominent venture capitalists), the concentration and strength of its core customers and suppliers;

[0106] Macroeconomic and industry data: Current overall market risk appetite, the level of policy support for the technology sector, etc.

[0107] Build a unified data foundation;

[0108] 2. Signal generation steps:

[0109] Generate multimodal prediction signals:

[0110] Sentiment Index: Based on public opinion during the roadshow and IPO subscription phases, a market enthusiasm index for IPO subscriptions is generated.

[0111] Fund flow identification: Although there is no historical fund flow on the first day of listing, the proportion of "long-term funds" and "short-term speculative funds" can be identified by the type of offline placement objects (public funds, private funds, social security funds, etc.) and the lock-up period.

[0112] Mobile Big Data Special Signal: If user activity data is impressive, it will generate a strong positive fundamental signal;

[0113] 3. Game theory deduction steps:

[0114] An evolutionary game model was constructed to simulate the dynamic game process between the bulls and bears on the first day of listing.

[0115] Institutional investors: The strategy is to buy or sell within a certain price range based on valuation and fundamentals. Some institutions that have been allocated shares have the incentive to sell and realize profits.

[0116] Retail investors: Their strategy is to buy high and sell low, and they are significantly influenced by the stock price movement after the market opens and the order book.

[0117] Industrial capital (such as company executives / employees): Due to the lock-up period, they cannot sell on the first day, but their holding mentality will indirectly affect the game through market sentiment;

[0118] At the same time, a cooperative game theory model is used to analyze which institutional investors may form a "price cartel" to maintain the stock price, and the stability of such a cartel;

[0119] Decision application steps:

[0120] Investment decision strategy signal: The system outputs a set of strategy suggestions based on game equilibrium;

[0121] If the opening price is lower than the equilibrium range: Game theory analysis shows that "value discovery" funds based on fundamental signals will gradually enter the market, followed by retail investors, and the price is likely to rise. Strategy: Buy at an opportune time after the market opens.

[0122] If the opening price is much higher than the equilibrium range: Game theory analysis shows that the institutions that have been allocated shares have a strong desire to cash out, while the fundamental signals are not enough to support such a high premium. The market is very likely to form a "prisoner's dilemma" type of long squeeze. Strategy: Remain on the sidelines, or if short selling is allowed, take the opportunity to short sell when the price rises and then weakens.

[0123] Risk Warning: When stock prices are irrationally driven to dangerous levels, the system issues a price bubble risk warning.

[0124] Examples 1, 2, and 3 demonstrate how by integrating multi-source big data with complex game theory models, traditional, historical data-based analysis can be transformed into a dynamic projection of future market interactions, thereby providing users with more forward-looking and insightful decision-making support.

[0125] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

[0126] The above description is only a preferred embodiment of the present invention and is not intended to limit the ideas of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A financial investment decision support method based on game theory and multi-source big data, characterized in that, include: Data fusion steps: Collect multi-source financial data, including financial market data, alternative big data, and macro and industry data; perform natural language processing on the multi-source financial data to transform it into a structured feature set; and perform data cleaning and alignment to build a unified data foundation. Signal generation steps: Based on the unified data foundation, a multimodal prediction signal is generated through a machine learning model. The multimodal prediction signal includes sentiment index, event impact assessment, fund flow identification, correlation network analysis, and mobile big data-specific signals. Game deduction steps: Use the multimodal prediction signal as the input parameter of the game model, abstractly model the market participants and solve the game equilibrium, and output the optimal strategy probability distribution of each participant; Decision application steps: Based on the solution of the game equilibrium, generate investment decision strategy signals or risk warning signals, and display them through a visual interface.

2. The financial investment decision support method based on game theory and multi-source big data as described in claim 1, characterized in that: The data fusion step further includes: The multi-source financial data is stored using a two-tier storage architecture that combines a time-series database and a data lake. Establish a hierarchical and categorized data management mechanism to trace the source, control access, and conduct compliance audits throughout the entire data lifecycle.

3. The financial investment decision support method based on game theory and multi-source big data as described in claim 1, characterized in that: The mobile big data-specific signals generated in the signal generation step include: An industry prosperity index based on the intensity of employees and the activity level of drivers in the chemical industry; Enterprise operating rate and inventory index inferred from factory personnel and truck traffic; The regional economic activity index is constructed based on population flow and freight volume, and includes a resumption of work index and a consumption activity index.

4. The financial investment decision support method based on game theory and multi-source big data according to claim 3, characterized in that: The signal generation step also includes a credit scoring model optimization process: A credit scoring model is constructed using deep learning or ensemble learning algorithms, and the model is iteratively trained and evaluated; the credit scoring model is used to score the creditworthiness of individuals or enterprises; this score serves as one dimension of a multimodal prediction signal. Introduce explainable AI technology to explain the model's decision-making process; Establish a self-learning mechanism for the model to absorb new data in real time and dynamically adjust the scoring logic.

5. The financial investment decision support method based on game theory and multi-source big data as described in claim 1, characterized in that: In the game theory deduction steps, the market participants include at least four types of entities: institutional investors, retail investors, industrial capital, and regulators. The game model includes various types such as incomplete information game model, evolutionary game model, corporate finance and industrial investment game model, Stackelberg game model, and cooperative game model. The corporate finance and industrial investment game model constructed in the game theory deduction steps takes enterprises as game participants, sets unit production cost, financing capacity, and financial market return rate as model parameters, and solves the optimal allocation ratio between financial investment and industrial investment for enterprises through Nash equilibrium.

6. The financial investment decision support method based on game theory and multi-source big data as described in claim 5, characterized in that: The decision application steps further include: It provides a scenario simulator that supports user-defined events and can deduce market reactions and optimal strategies based on game theory models. When the game theory model predicts that the market is evolving toward a high-risk equilibrium, a risk warning is issued in advance. The risk warning includes the risk of excessive financialization of enterprises or insufficient investment in real industries.

7. A financial investment decision support system based on game theory and multi-source big data, characterized in that, The financial investment decision support system based on game theory and multi-source big data is used to implement the financial investment decision support method based on game theory and multi-source big data as described in any one of claims 1-6. The financial investment decision support system includes: The data layer module is used to collect multi-source financial data, including financial market data, alternative big data, and macro and industry data. Natural language processing is performed on the multi-source financial data to transform it into a structured feature set, and data cleaning and alignment are performed to build a unified data foundation. The analysis layer module, connected to the data layer module, is used to generate multimodal prediction signals based on the unified data foundation through machine learning models. The multimodal prediction signals include sentiment index, event impact assessment, fund flow identification, correlation network analysis, and mobile big data-specific signals. The game layer module, connected to the analysis layer module, is used to take the multimodal prediction signal as the input parameter of the game model, abstractly model the market participants and solve the game equilibrium, and output the optimal strategy probability distribution of each participant. The market participants include at least four types of entities: institutional investors, retail investors, industrial capital and regulators. The game model includes multiple types of models such as incomplete information game model, evolutionary game model, corporate finance and real investment game model, Stackelberg game model and cooperative game model. The application layer module, connected to the game layer module, is used to generate investment decision strategy signals or risk warning signals based on the solution results of the game equilibrium, and to display them through a visual interface.

8. A financial investment decision support system based on game theory and multi-source big data as described in claim 7, characterized in that: The game theory model for corporate finance and real investment constructed in the game theory layer module takes the enterprise as the game participant, sets the unit production cost, financing capacity and financial market return rate as model parameters, and solves the optimal allocation ratio between financial investment and real investment by the enterprise through Nash equilibrium.

9. A financial investment decision support system based on game theory and multi-source big data as described in claim 7, characterized in that: The application layer module also includes: Scenario simulators are used to support user-defined events and deduce market reactions and optimal strategies based on game theory models. The risk warning unit is used to issue early warnings of risks, including the risk of excessive financialization of enterprises or insufficient investment in real industries, when the game model predicts that the market is evolving towards a high-risk equilibrium.

10. A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 6.

Images