Financial product risk monitoring system based on big data analysis

Abstract

The invention discloses a financial product risk monitoring system based on big data analysis, and belongs to the technical field of financial product risk monitoring, and the system specifically comprises a time axis construction module which analyzes a contract time stamp sequence, extracts a key time parameter, and uses a logic operator to construct a three-dimensional performance time axis; the game subject identification module is used for identifying subject behavior decision nodes according to a time axis, marking a claim timeliness interval, positioning a cost avoidance window and quantifying a supervision response lag period; the game tree generation module is used for mapping a time axis and a main body into a decision tree, generating a strategy branch, establishing a path dependence rule and embedding an external event trigger; the game analysis module is used for calculating a node default probability, identifying key points and quantifying a balanced vulnerability index; the risk entropy field construction module is used for marking game imbalance overlapping sections, calculating an influence attenuation curve and converting the influence attenuation curve into a visual risk gradient; and the risk map generation module is used for marking a sensitive interval and outputting a thermodynamic map to display risk dynamic distribution.

Description

Technical Field

[0001] The present invention relates to the technical field of financial product risk monitoring, and in particular to a financial product risk monitoring system based on big data analysis. Background Art

[0002] In the financial field, risk monitoring of financial products is crucial and directly related to the sound operation of financial institutions and the interests of investors.

[0003] Traditional financial product risk monitoring methods have many limitations. In the early days, they mainly relied on manual experience judgment, and financial analysts evaluated the risks of financial products based on their own professional knowledge and past experience. However, this method is highly subjective, the judgment criteria of different analysts vary, and it is inefficient, making it difficult to cope with the growing number of financial products and the complex market environment. With the development of technology, some risk monitoring methods based on simple models have emerged, such as risk assessment models based on financial indicator analysis. Although these models have improved the objectivity of risk assessment to a certain extent, they rely on limited data sources and often only focus on the financial data of financial products, ignoring a large amount of other important information, such as market dynamics, industry trends, and public opinion information. Moreover, these simple models have difficulty in handling complex nonlinear relationships, and the accuracy of risk assessment for complex financial products is insufficient.

[0004] With the rise of big data technology, new opportunities have been brought to financial product risk monitoring. Big data technology can collect and process massive amounts of multi-source data, including internal data of financial institutions, market data, regulatory data, and social media data. Many financial product risk monitoring methods based on big data analysis have emerged, but these existing methods regard contracts as fixed texts, ignoring the dynamic evolution characteristics of the time trigger mechanism and multi-party strategy game in the performance process; existing knowledge graphs only depict the spatial correlation between clauses, and cannot model the causal chain of risk transmission on the time axis and the fragility of the game equilibrium; Summary of the invention

[0005] The purpose of the present invention is to provide a financial product risk monitoring system based on big data analysis to solve the following technical problems: The contract is regarded as a fixed text, ignoring the dynamic evolution characteristics of the time trigger mechanism and multi-party strategy game in the performance process; the existing knowledge graph only describes the spatial correlation between clauses, and is unable to model the causal chain of risk transmission on the time axis and the fragility of the game equilibrium.

[0006] The purpose of the present invention can be achieved through the following technical solutions: A financial product risk monitoring system based on big data analysis, comprising: The timeline construction module is used to parse the time mark sequence in the contract text, extract the time constraint parameters of the effective time point, performance period and trigger conditions of each clause, and construct a three-dimensional performance timeline covering the entire life cycle of the contract through time logic operators; The game subject identification module is used to identify the behavioral decision nodes of each contract subject in the time dimension according to the time axis, mark the sensitive interval of the claim time of the right claiming party; locate the cost avoidance window of the obligation fulfilling party; and quantify the response lag cycle of the third-party supervision; A game tree generation module is used to map the time axis and the game subject into a spatiotemporal game decision tree, generate different strategy branches in each time node, establish path dependency rules, select a feasible path set for the preceding node strategy to dynamically constrain the subsequent node, and embed external event triggers in the key time window; A game analysis module is used to calculate the default probability of each subject at each node according to the game tree, identify the strategy deviation that causes the sudden change of the game path, mark it as a key point, and quantify the minimum external disturbance required to maintain the current game path as an equilibrium vulnerability index; The risk entropy field construction module is used to mark the time overlap segments of multiple game imbalance nodes on the time axis, calculate the attenuation curve of the impact of a certain node strategy deviation on the subsequent period, and convert the game relationship into a visual risk gradient distribution; The risk map generation module is used to inject extreme market events, observe the path breakpoints and strategy reorganization patterns of the game tree, mark the continuous periods when the equilibrium vulnerability index is greater than the preset threshold as time-sensitive risk intervals, and output risk heat maps to show the dynamic distribution of risk entropy values ​​as time, subject, and market dimensions change.

[0007] As a further solution of the present invention: the timeline construction module parses the time mark sequence in the contract text, including: Explicit time expressions, including date formats and period descriptions, are identified through a multimodal time parser. A time conflict detection matrix is ​​constructed to identify timing inversions, window overlaps, and period resonances between clauses. For clauses without explicit timing constraints, dynamic time boundaries are inferred based on historical performance data and industry rule bases.

[0008] As a further solution of the present invention: the three-dimensional performance timeline includes contract timing, market timing and game timing.

[0009] As a further solution of the present invention: the game subject identification module identifies the behavior decision nodes of each contract subject in the time dimension, including: Identify asymmetric game structures in clauses and annotate the time window cutoff mechanism set by one party through a decision tree classifier; Analyze the timing difference of information flow between entities and calculate the overlap between the latest time when the claimant obtains performance information and the deadline for performance of obligations; Construct a regulatory response lag model to quantify the impact of the average time from the occurrence of default to regulatory intervention on the game path; Generate a dynamic strategy characteristic matrix for each entity, including risk preference, historical strategy selection tendency and resource allocation speed.

[0010] As a further solution of the present invention: the game tree generation module maps the time axis and the game subject into a spatiotemporal game decision tree, including: Discretize the three-dimensional timeline into game decision stages, and define a set of atomic strategies that the subject can choose in each stage, including normal performance, delayed performance, and request for change. Establish path constraint propagation rules. When a node chooses a non-cooperative strategy, the compliance branch in the subsequent path is automatically blocked. Embed environmental monitoring triggers at key positions on the timeline to activate backup gaming paths when external market indicators exceed preset thresholds; A strategy entropy value label is attached to each node to quantify the diversity of the optional strategies of each node. The higher the entropy value, the more optional strategies there are.

[0011] As a further solution of the present invention: the game analysis module specifically includes: Monte Carlo strategy deduction is used to simulate the probability distribution of strategy selection of each subject under time pressure to obtain the probability of default; Locate the key points through a decision tree reverse traversal algorithm, i.e., nodes whose removal reduces the game path differentiation rate by more than 60%; The calculation formula of the path vulnerability index V is: ; Among them, P a represents the mean probability of default of each entity, T represents the duration of the contract, T 0 It represents the duration of the market event within the contract constraint period, N represents the number of market events, r represents the number of subsequent compliant paths of the current node, and R represents the total paths following the current node.

[0012] As a further solution of the present invention: the risk entropy field construction in the risk entropy field construction module includes: Map the time axis as a risk transmission medium and define the decay function of risk entropy at early nodes as R(t)=R 0 e -λΔt ; where R 0represents the initial value of risk entropy, R(t) is the risk entropy value at time point t, λ represents the risk attenuation coefficient, Δt is the time difference from the time when the risk event occurs to the current time, and e represents a natural number; Construct a risk ripple superposition model, and when the risk entropy peaks of multiple nodes overlap on the timeline, a risk resonance alarm is triggered.

[0013] As a further solution of the present invention: the risk map generation module outputs the risk heat map including: Set up a multi-dimensional stress test scenario library, including a sudden drop in market liquidity, gradual changes in industry policies, and changes in entity control; Path sensitivity analysis is used to calculate the disturbance coefficient of the change in the strategy weight of each node on the overall risk entropy field; Construct a butterfly effect early warning chain and locate the initial deviation event set that may trigger systemic risks through correlation analysis; Generate a spatiotemporal risk heat map, use color depth to represent the risk entropy value, and use isochrones to reflect the speed of risk transmission.

[0014] Beneficial effects of the present invention: The present invention transforms traditional static text analysis into a strategy interaction process in the space-time dimension by constructing a three-dimensional performance timeline and a dynamic game decision tree, effectively solving the core problem that the existing technology cannot capture the dynamic evolution risk of contract terms during the performance process. It innovatively introduces the game equilibrium vulnerability index and the risk entropy field quantitative model, realizes the coordinated analysis of event triggering mechanisms, multi-party strategy games and risk transmission paths, and can accurately identify the butterfly effect nodes that cause system imbalance. Through Monte Carlo strategy deduction and multi-dimensional stress testing scenario library, the present invention can simulate the sudden change of game paths under extreme market events, dynamically generate space-time risk heat maps, provide key node defense suggestions for contract terms optimization, and support real-time injection of regulatory rule changes and market data streams for dynamic calibration, which completely breaks through the prediction inaccuracy of traditional methods in dealing with time-series compound crises and strategy conflicts, and provides intelligent decision support for financial contract risk management that is both forward-looking and adaptive. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention will be further described below in conjunction with the accompanying drawings.

[0016] Figure 1 It is a module schematic diagram of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] See also Figure 1 As shown, the present invention is a financial product risk monitoring system based on big data analysis, comprising: Timeline building blocks: This module is the foundation of the entire system. It undertakes the important task of parsing the complex time information in the contract text. The contract text usually contains many clauses, each of which may involve key elements such as the effective time point, the fulfillment period, and the time constraints of the triggering conditions. The timeline construction module will carefully analyze these time-stamped sequences and use advanced natural language processing technology and text mining algorithms to accurately extract the time-related parameters of each clause.

[0019] After obtaining these parameters, the module will use a variety of time logic operators, such as the common "and / or" combination, as well as complex window period nesting, to construct a three-dimensional performance timeline covering the entire life cycle of the contract. This three-dimensional timeline includes T1 contract timeline, which mainly reflects the time sequence and logical relationship stipulated in the contract terms themselves; T2 market timeline, which is closely related to the dynamic changes in the market environment, such as the fluctuations of market indicators and the adjustment of macroeconomic policies, which have an impact on the time of contract performance; T3 game timeline, which focuses on the strategic interaction and decision-making process of the various contract participants in the time dimension. By constructing such a three-dimensional timeline, the complex structure and dynamic changes of the contract in the time dimension can be fully and accurately presented, providing a solid foundation for subsequent analysis.

[0020] Game subject identification module: Based on the three-dimensional performance timeline generated by the timeline construction module, this module deeply identifies the behavioral decision nodes of each subject in the time dimension. For the rights claimant, the module will accurately mark the sensitive interval of its claim validity period. This interval is usually the overlapping period of the latest claim deadline and the validity period of the evidence. During this period, the claim behavior of the rights claimant has a higher legal effect and success probability. By accurately marking this interval, it helps the rights claimant to reasonably arrange the timing of the claim and improve the success rate of the claim.

[0021] For the obligee, the module will locate its cost avoidance window. The determination of this window requires comprehensive consideration of the game relationship between the low-risk performance period and the market volatility period. During periods of low market volatility and low risk, the obligee may be more inclined to perform its obligations in accordance with the contract to avoid additional costs due to breach of contract; while during periods of high market volatility, the obligee may adjust its strategy according to market conditions and look for a lower-cost way to perform. By locating the cost avoidance window, the obligee can optimize its performance strategy and reduce performance costs.

[0022] In addition, the module will also quantify the response lag cycle of third-party supervision. In the financial market, the intervention of third-party regulators is crucial to maintaining market order and fair performance of contracts. However, there is often a certain time lag from the occurrence of violations to regulatory intervention, which may have an important impact on the game balance between the parties to the contract. The module will collect and analyze a large amount of regulatory data, combined with the actual market situation, accurately quantify this response lag cycle, and provide valuable reference information for regulators and market participants.

[0023] Game tree generation module: This module organically maps the timeline generated by the timeline construction module with the game roles determined by the game subject identification module to generate a spatiotemporal game decision tree. In this decision tree, each time node represents a decision point, which will generate multiple strategy branches, including common decision options such as normal performance, request for change of terms, and dispute resolution.

[0024] In order to accurately reflect the dynamic changes and path dependencies in the contract performance process, the module will establish path dependency rules. These rules indicate that the strategy selection of the previous node will dynamically constrain the feasible path set of the subsequent node. For example, if a request for a change of terms is selected at a certain time node, the subsequent decision path will be restricted by this choice, which may involve a series of decisions such as negotiating with the other party on the specific content of the change of terms and re-determining the performance period.

[0025] In critical time windows, the module embeds external event triggers. These triggers can be breakthroughs in market indicator thresholds, such as stock prices, interest rates, exchange rates, etc. reaching specific thresholds; they can also be policy release dates, such as the government's introduction of new financial regulatory policies, tax policies, etc. When these external events occur, they will have a significant impact on the decision-making of the various parties in the contract, thereby changing the path of the game tree. By embedding external event triggers, the game tree can more realistically reflect the impact of changes in the market environment on contract performance.

[0026] Game analysis module: This module conducts multi-level strategy deduction based on the spatiotemporal game decision tree generated by the game tree generation module. First, the module calculates the strategy deviation of each node, that is, the probability that the subject chooses a non-agreed strategy. By analyzing and modeling a large amount of historical data, combined with the market environment and the behavioral characteristics of the subject, the module predicts the possibility of the subject choosing a non-agreed strategy under different circumstances. This indicator can help financial institutions and regulatory authorities assess the uncertainty and risk level in the contract performance process.

[0027] Secondly, the module will identify butterfly effect nodes. During the game process, some minor deviations in strategy may cause a sudden change in the game path. These nodes are butterfly effect nodes. Through in-depth analysis and simulation of the game tree, these key nodes are found and focused on and monitored. Once abnormal conditions are found in these nodes, timely measures can be taken to avoid the expansion and spread of risks.

[0028] Finally, the module will quantify the equilibrium fragility index. This index is used to measure the minimum external disturbance energy required to maintain the current game path. The higher the index, the more fragile the current game equilibrium is, and it is easily affected by external factors and changed. By quantifying the equilibrium fragility index, it can provide financial institutions and regulatory authorities with an intuitive indicator for evaluating the stability and risk level of the contract, so that appropriate measures can be taken in a timely manner to prevent and control risks.

[0029] Risk entropy field building blocks: This module integrates the time constraint strength and game stability parameters to construct a space-time risk field. On the time axis, the module will mark the risk concentration area, which is usually the time overlap of multiple game imbalance nodes. In the risk concentration area, the uncertainty and risk level in the contract performance process are high, which requires special attention and management.

[0030] The module calculates the risk transmission ripple, that is, the attenuation curve of the impact of a certain node strategy deviation on subsequent periods. By analyzing and modeling a large amount of historical data, combined with the structure and characteristics of the game tree, the spread and impact of a certain node's strategy deviation in different time and space dimensions are predicted. Based on the calculation results of the risk transmission ripple, the risk entropy contour is generated, and the abstract game relationship is converted into a visualized risk gradient distribution. This visualized risk field can help financial institutions and regulatory authorities intuitively understand the risk distribution in the contract performance process and provide strong support for risk decision-making.

[0031] Risk map generation module: This module simulates the evolution of game paths under external environmental disturbances. By injecting extreme market events, such as financial crises, major natural disasters, and major policy adjustments, the path breakpoints and strategy reorganization patterns of the game tree are observed. Under the influence of extreme market events, the decisions and behaviors of the parties to the contract may change significantly, resulting in the path of the game tree breaking or reorganizing. The module will record and analyze these changes in detail to identify possible risk points and response strategies.

[0032] The module will mark time-sensitive risk intervals, which are continuous periods when the game equilibrium vulnerability index is greater than the threshold. During these periods, the stability of the contract is poor and it is easy to be affected by external factors and cause risk events. By marking time-sensitive risk intervals, it can help financial institutions and regulatory authorities to prepare for risk prevention and response in advance.

[0033] Finally, the module will output a risk heat map, showing the dynamic distribution of risk entropy values ​​over time, subject, and market dimensions. This map can intuitively reflect the changes in risks during the performance of the contract, providing financial institutions and regulators with a comprehensive and dynamic risk monitoring tool. At the same time, the module will also generate a monitoring report containing elements such as the game imbalance index, strategy sandbox simulation interface, and key node defense recommendations. The game imbalance index is used to quantify the inherent instability of the contract and provide a comprehensive score for the risk assessment of the contract; the strategy sandbox simulation interface supports the modification of time parameters and real-time observation of changes in risk entropy values, helping financial institutions and regulators to conduct risk scenario analysis and decision-making simulations; the key node defense recommendations propose specific clause optimization plans for the butterfly effect nodes, providing specific operational guidelines for risk prevention and control of the contract.

[0034] In another preferred embodiment of the present invention, the timeline construction module parses the time mark sequence in the contract text, including: Recognize explicit time expressions such as date formats and period descriptions, as well as implicit event triggers such as market indicator thresholds and behavioral triggers through a multimodal time parser; Construct a time conflict detection matrix to identify timing inversions between clauses, such as payment before acceptance; window overlaps, such as the hedging period and the account freeze period; and cycle resonances, such as the common denominator time period of multiple cycle clauses; For clauses without clear timing constraints, dynamic time boundaries are inferred based on historical performance data and industry rule bases, and the default grace period = average response time of similar clauses × 1.5 standard deviations; In another preferred embodiment of the present invention, the game subject identification module identifies the behavior decision nodes of each contract subject in the time dimension, including: Identify asymmetric game structures in the terms and annotate the strategic control points of the dominant party through a decision tree classifier, such as the time window cutoff mechanism set by one party; Analyze the timing difference of information flow between entities and calculate the overlap between the latest time when the claimant obtains performance information and the deadline for performance of obligations; Construct a regulatory response lag model to quantify the impact of the average time from the occurrence of default to regulatory intervention on the game path; Generate a dynamic strategy feature matrix for each entity, including three dimensions: risk preference, historical strategy selection tendency, and resource allocation speed.

[0035] In another preferred embodiment of the present invention, mapping the time axis and the game subject into a spatiotemporal game decision tree in the game tree generation module includes: Discretize the three-dimensional timeline into game decision stages, and define a set of atomic strategies that the subject can choose in each stage, such as normal performance / delayed performance / request for change; Establish path constraint propagation rules. When a node chooses a non-cooperative strategy, the compliance branch in the subsequent path is automatically blocked. Embed environmental monitoring triggers at key positions on the timeline to activate backup gaming paths when external market indicators exceed preset thresholds; A strategy entropy value label is attached to each node to quantify the diversity of optional strategies at that moment. The higher the entropy value, the greater the game uncertainty.

[0036] In another preferred embodiment of the present invention, the game analysis module specifically includes: Monte Carlo strategy deduction is used to simulate the probability distribution of strategy selection of each subject under time pressure; Identify strategic resonance points where multiple entities tend to deviate from the agreed strategy in the same period; The butterfly effect node is located through the decision tree reverse traversal algorithm, that is, the node whose removal can reduce the game path differentiation rate by more than 60%.

[0037] The calculation formula of the path vulnerability index V is: ; Among them, P a represents the mean probability of default of each entity, T represents the duration of the contract, T 0 It represents the duration of the market event within the contract constraint period, N represents the number of market events, r represents the number of subsequent compliant paths of the current node, and R represents the total paths following the current node.

[0038] In another preferred embodiment of the present invention, the risk entropy field construction in the risk entropy field construction module includes: Map the time axis as a risk transmission medium and define the decay function of risk entropy at early nodes as R(t)=R 0 e -λΔt ; where R 0 represents the initial value of risk entropy, R(t) is the risk entropy value at time point t, λ represents the risk attenuation coefficient, Δt is the time difference from the time when the risk event occurs to the current time, and e represents a natural number; Calculate the risk concentration gradient in a three-dimensional space-time coordinate system, integrating three parameters: game imbalance, market volatility, and subject credit attenuation factor; Construct a risk ripple superposition model. When the risk entropy peaks of multiple nodes overlap on the time axis, a risk resonance alarm is triggered. Generate risk equipotential surfaces that identify spatiotemporally continuous regions with the same risk entropy values.

[0039] In another preferred embodiment of the present invention, the risk map generation module outputs the risk heat map including: Set up a multi-dimensional stress test scenario library, including "black swan" events (sudden drop in market liquidity), "gray rhino" events (gradual changes in industry policies), and sudden changes in game rules (changes in control of the subject); Path sensitivity analysis is used to calculate the disturbance coefficient of the change in the strategy weight of each node on the overall risk entropy field; Construct a butterfly effect early warning chain and locate the initial deviation event set that may trigger systemic risks through correlation analysis; Generate a spatiotemporal risk heat map, use color depth to represent the risk entropy value, and use isochrones to reflect the speed of risk transmission.

[0040] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0041] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by any technician familiar with the technical field within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

[0042] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A financial product risk monitoring system based on big data analysis, characterized in that: include: The timeline construction module is used to parse the time mark sequence in the contract text, extract the time constraint parameters of the effective time point, performance period and trigger conditions of each clause, and construct a three-dimensional performance timeline covering the entire life cycle of the contract through time logic operators; A game subject identification module is used to identify the behavior decision nodes of each contract subject in the time dimension according to the time axis, and mark the time-sensitive interval of the claim of the right claiming party; Locate the cost avoidance window of the obligor; quantify the response lag period of third-party supervision; A game tree generation module is used to map the time axis and the game subject into a spatiotemporal game decision tree, generate different strategy branches in each time node, establish path dependency rules, select a feasible path set for the preceding node strategy to dynamically constrain the subsequent node, and embed external event triggers in the key time window; A game analysis module is used to calculate the default probability of each subject at each node according to the game tree, identify the strategy deviation that causes the sudden change of the game path, mark it as a key point, and quantify the minimum external disturbance required to maintain the current game path as an equilibrium vulnerability index; The risk entropy field construction module is used to mark the time overlap segments of multiple game imbalance nodes on the time axis, calculate the attenuation curve of the impact of a certain node strategy deviation on the subsequent period, and convert the game relationship into a visual risk gradient distribution; The risk map generation module is used to inject extreme market events, observe the path breakpoints and strategy reorganization patterns of the game tree, mark the continuous periods when the equilibrium vulnerability index is greater than the preset threshold as time-sensitive risk intervals, and output risk heat maps to show the dynamic distribution of risk entropy values ​​as time, subject, and market dimensions change.

2. A financial product risk monitoring system based on big data analysis according to claim 1, characterized in that: The timeline construction module parses the time mark sequence in the contract text, including: Explicit time expressions, including date formats and period descriptions, are identified through a multimodal time parser. A time conflict detection matrix is ​​constructed to identify timing inversions, window overlaps, and period resonances between clauses. For clauses without explicit timing constraints, dynamic time boundaries are inferred based on historical performance data and industry rule bases.

3. The financial product risk monitoring system based on big data analysis according to claim 1 is characterized in that: The three-dimensional performance timeline includes contract timing, market timing and game timing.

4. A financial product risk monitoring system based on big data analysis according to claim 1, characterized in that: The game subject identification module identifies the behavior decision nodes of each contract subject in the time dimension, including: Identify asymmetric game structures in clauses and annotate the time window cutoff mechanism set by one party through a decision tree classifier; Analyze the timing difference of information flow between entities and calculate the overlap between the latest time when the claimant obtains performance information and the deadline for performance of obligations; Construct a regulatory response lag model to quantify the impact of the average time from the occurrence of default to regulatory intervention on the game path; Generate a dynamic strategy characteristic matrix for each entity, including risk preference, historical strategy selection tendency and resource allocation speed.

5. A financial product risk monitoring system based on big data analysis according to claim 1, characterized in that: In the game tree generation module, mapping the time axis and the game subject into a spatiotemporal game decision tree includes: Discretize the three-dimensional timeline into game decision stages, and define a set of atomic strategies that the subject can choose in each stage, including normal performance, delayed performance, and request for change. Establish path constraint propagation rules. When a node chooses a non-cooperative strategy, the compliance branch in the subsequent path is automatically blocked. Embed environmental monitoring triggers at key positions on the timeline to activate backup gaming paths when external market indicators exceed preset thresholds; A strategy entropy value label is attached to each node to quantify the diversity of the optional strategies of each node. The higher the entropy value, the more optional strategies there are.

6. A financial product risk monitoring system based on big data analysis according to claim 1, characterized in that: The game analysis module specifically includes: Monte Carlo strategy deduction is used to simulate the probability distribution of strategy selection of each subject under time pressure to obtain the probability of default; Locate the key points through a decision tree reverse traversal algorithm, i.e., nodes whose removal reduces the game path differentiation rate by more than 60%; The calculation formula of the path vulnerability index V is: ； Among them, P a represents the mean probability of default of each entity, T represents the duration of the contract constraint, T0 represents the duration of the market event within the contract constraint, N represents the number of market events, r represents the number of subsequent compliance paths of the current node, and R represents the total path following the current node.

7. A financial product risk monitoring system based on big data analysis according to claim 1, characterized in that: The risk entropy field construction module includes: Map the time axis as a risk transmission medium and define the attenuation function of risk entropy at early nodes as R(t)=R0e -λΔt ; Where R0 represents the initial value of risk entropy, R(t) is the risk entropy value at time point t, λ represents the risk attenuation coefficient, Δt is the time difference from the time when the risk event occurs to the current time, and e represents a natural number; Construct a risk ripple superposition model, and when the risk entropy peaks of multiple nodes overlap on the timeline, a risk resonance alarm is triggered.

8. The financial product risk monitoring system based on big data analysis according to claim 1 is characterized in that: The risk map generation module outputs the risk heat map including: Set up a multi-dimensional stress test scenario library, including a sudden drop in market liquidity, gradual changes in industry policies, and changes in entity control; Path sensitivity analysis is used to calculate the disturbance coefficient of the change in the strategy weight of each node on the overall risk entropy field; Construct a butterfly effect early warning chain and locate the initial deviation event set that may trigger systemic risks through correlation analysis; Generate a spatiotemporal risk heat map, use color depth to represent the risk entropy value, and use isochrones to reflect the speed of risk transmission.

Images