A financial intelligence decision support method that integrates business and industry data

Abstract

The application relates to the technical field of enterprise financial intelligence management, and discloses a financial intelligence decision support method integrating business and industry data, which comprises the following steps: constructing a standardized time series data set, a standardized financial pressure event template library and a simplified dynamic model of an enterprise financial system; defining a set of observation index sets; running a Monte Carlo simulation containing event injection for a pre-evaluated financial strategy to obtain a benchmark evolution path and a stress response trajectory; calculating a strategy expected return score and a strategy resilience score of each financial strategy; establishing a two-dimensional coordinate system with the strategy expected return score as the horizontal axis and the strategy resilience score as the vertical axis, drawing all the financial strategies in the two-dimensional coordinate system, automatically calculating and identifying a Pareto optimal strategy set, and generating a resilience-return strategy report containing strategy characteristic analysis; and the application can improve the efficiency of financial intelligence decision of the integrated business and industry data.

Description

Technical Field

[0001] This invention relates to the field of enterprise financial intelligent management technology, and in particular to a financial intelligent decision support method that integrates business and industry data. Background Technology

[0002] Current enterprise financial intelligent decision support technologies generally focus on static return calculation and single-dimensional indicator optimization. Their core is limited to selecting the financial strategy with the highest expected return in a certain environment, without incorporating external uncertainties and sudden risk disturbances into the core decision-making framework. This leads to a disconnect between the decision-making model and the real operating environment. Existing methods mostly rely on the ex-post analysis of internal financial data, lacking in-depth integration and standardized time-series processing of business data, industry data, and macro data. Risk assessment relies on human experience judgment and has not formed a standardized financial stress event system. It is difficult to quantify the resilience and recovery ability of strategies under shocks, resulting in an imbalance between return and risk resilience in the decision-making results.

[0003] Meanwhile, existing financial simulation technologies have not achieved a deep integration of advanced engineering theories with financial decision-making. They have neither introduced the fault injection mechanism of chaos engineering to achieve dynamic simulation of random stress events, nor used robustness analysis of control theory to construct quantitative observation indicators. As a result, they cannot accurately measure the instantaneous impact offset, recovery speed, and steady-state loss of strategies. Financial strategy evaluation is solely driven by returns, lacking a dual-dimensional trade-off mechanism between returns and resilience. It cannot automatically identify Pareto optimal strategy sets, leading to low decision-making efficiency and insufficient risk prediction. This makes it difficult to support the need for stable operation of enterprises in complex and volatile environments. Therefore, how to improve the efficiency of strategy decision-making has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a financial intelligent decision support method that integrates business and industry data to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, this invention provides a financial intelligent decision support method that integrates business and industry data, comprising: S1 integrates internal business data and external industry data and preprocesses them to obtain a standardized time-series dataset. Based on historical risk cases and industry knowledge, a standardized financial stress event template library is built. S2. Based on the standardized time-series dataset, a simplified dynamic model of the enterprise financial system is constructed, and a set of observation indicators for quantifying the robustness of the system is defined. S3, for the pre-assessed financial strategy, runs a Monte Carlo simulation with event injection to obtain the baseline evolution path and stress response trajectory; S4. Based on the baseline evolution path, calculate the expected return score of each financial strategy. Based on the stress response trajectory and the observation index set, apply the strategy resilience score calculation algorithm to calculate the strategy resilience score of each financial strategy. S5 establishes a two-dimensional coordinate system with the strategy expected return score as the horizontal axis and the strategy resilience score as the vertical axis, plots all financial strategies in it, automatically calculates and identifies the Pareto optimal strategy set, and generates a resilience-return strategy report that includes strategy characteristic analysis.

[0006] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention fundamentally reconstructs the core logic of intelligent financial decision-making, upgrading traditional single-profit-oriented decision-making to a dual-objective synergistic optimization of profit and resilience, achieving a fundamental shift in the decision-making paradigm. Through a design approach oriented towards uncertainty and sudden disturbances, it simultaneously quantifies the expected returns of strategies and their robustness in terms of shock resistance, recoverability, and steady-state operation. This completely solves the shortcomings of traditional methods that only focus on profits and ignore risk resilience, allowing financial decision-making to shift from static optimal selection to dynamic adaptive selection, significantly improving the scientific nature and robustness of decision-making in complex operating environments. At the same time, relying on multi-source data integration, standardized time-series processing, and a stress event template library, it achieves unified modeling of internal and external data, greatly improving the efficiency and data support capabilities of financial decision analysis.

[0007] 2. This invention creatively adapts the chaotic engineering fault injection framework and control theory robustness analysis model to the field of financial decision-making, constructing a brand-new simulation and evaluation system. Utilizing event-injected Monte Carlo simulation, it can reproduce the dynamic impact of various financial stress events on strategies, accurately generating baseline paths and stress response trajectories. Based on three observation indicators—maximum instantaneous offset, recovery time constant, and steady-state error percentage—it achieves quantifiable and comparable calculations of strategy resilience. Combined with the Pareto optimality algorithm, it automatically selects the optimal strategy set, forming a complete "simulation-calculation-selection-reporting" decision-making closed loop, effectively improving the risk resistance and decision-making efficiency of financial strategies, and providing implementable and quantifiable technical support for enterprise financial intelligent management. Attached Figure Description

[0008] Figure 1 A flowchart illustrating a financial intelligent decision support method that integrates business and industry data, provided as an embodiment of the present invention; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0009] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0010] This application provides a financial intelligent decision support method that integrates business and industry data. The executing entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the financial intelligent decision support method that integrates business and industry data can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0011] Reference Figure 1 The diagram shown is a flowchart illustrating a financial intelligent decision support method integrating business and industry data according to an embodiment of the present invention. In this embodiment, the financial intelligent decision support method integrating business and industry data includes: S1 integrates internal business data and external industry data and preprocesses them to obtain a standardized time-series dataset. Based on historical risk cases and industry knowledge, a standardized financial stress event template library is built. In this embodiment of the invention, the process of integrating internal business data and external industry data and preprocessing them to obtain a standardized time-series dataset includes: Based on preset multi-source data acquisition instructions, internal business data is obtained from the enterprise's local database, and external industry data is extracted from external industry data service provider interfaces and macroeconomic databases to obtain multi-source heterogeneous datasets. The multi-source heterogeneous dataset is subjected to unified standardization processing to obtain a standardized time series dataset, wherein the unified standardization processing includes format conversion, data cleaning, and time series alignment.

[0012] It should be noted that obtaining internal business data from the enterprise's local database refers to retrieving time-series data items from the enterprise's cash flow statement, profit and loss statement, and balance sheet through a pre-set internal enterprise API interface.

[0013] It should be noted that the operation of extracting external industry data from external industry data service provider interfaces and macroeconomic databases refers to obtaining key indicator data of the industry in which the company is located, including but not limited to the industry average profit margin and market size growth rate, as well as macroeconomic indicators including GDP growth rate and inflation rate, through authorized industry data APIs.

[0014] It should be noted that the aforementioned multi-source heterogeneous dataset refers to a collection of original data that includes both internal enterprise operational dimensions and external market environment dimensions, and that differs in time frequency and numerical format. It serves as the basic data source for subsequent data integration and cleaning.

[0015] It should be noted that the format conversion operation is based on the original data format of each data source in the multi-source heterogeneous dataset, converting non-numerical data into numerical codes, and unifying the units of all numerical data to the benchmark unit of measurement. This operation is used to eliminate the formal barriers between data and lay the foundation for subsequent calculations.

[0016] It should be noted that the data cleaning operation involves applying an anomaly detection algorithm based on the 3σ principle of statistical distribution to the format-converted dataset to identify and process missing values, outliers, and logical error values.

[0017] Furthermore, the essence of anomaly detection algorithms is to calculate the mean and standard deviation of each time series data column, and identify outliers as values ​​that deviate from the mean by more than 3 times the standard deviation. For missing values, imputation is performed using forward imputation or median imputation, depending on the data type.

[0018] It should be noted that the time-series alignment operation is performed on cleaned data with different sampling frequencies, using an interpolation algorithm to unify all data onto the same timestamp sequence.

[0019] Furthermore, the mathematical expression of the interpolation algorithm is as follows:

[0020] In the formula, Indicates target alignment date The interpolation results, and Representing the starting month and the end of the month Known data values, This indicates the date to be interpolated. Indicates the starting month. It indicates the end of the month.

[0021] It should be noted that the normalization transformation operation finds the maximum and minimum values ​​in the time series data of the same indicator; then, for each data point in the series, the minimum value of the series is subtracted; finally, the above difference is divided by the difference between the maximum and minimum values, thereby mapping all data to the interval [0,1].

[0022] It should be noted that the standardized time series dataset refers to a clean time series data set that has undergone format conversion, data cleaning, time series alignment and normalization, and has a unified time index, unified data caliber and comparable order of magnitude.

[0023] In this embodiment of the invention, the step of constructing a standardized financial stress event template library based on historical risk cases and industry knowledge includes: Based on a pre-set historical case database and industry knowledge base, extract historical risk case data containing risk descriptions, impact dimensions, and response processes; Stress features are extracted from the historical risk case data to obtain core stress feature sets for different risk types; The core pressure feature set is mathematically parameterized to obtain a parameterized pressure event template. Based on the event impact assessment algorithm, event weights are configured for each parameterized stress event template to obtain a standardized financial stress event template library.

[0024] It should be noted that the extraction of historical risk case data from the preset historical case database and industry knowledge base refers to extracting text and data information containing risk triggering factors, key financial or business indicators affected, event duration, and post-event recovery from the structured crisis case database. Its purpose is to provide original materials and analytical basis for constructing standardized stress events. The structured crisis case database includes: the company's own historical audit reports, industry research reports, and publicly available records of macroeconomic crisis events.

[0025] It should be noted that the historical risk case data refers to a collection of detailed information that has been compiled and labeled, describing specific financial or operational risks that have occurred in the past. This data records actual observations of how shock events affect the enterprise system and cause changes in its state.

[0026] It should be noted that the operation of extracting stress features refers to performing text analysis and data mining on the historical risk case data to identify common impact patterns and key variables of different risk events. This process first extracts the key objects affected by the risk events, including raw material costs, product demand, and financing interest rates; secondly, it combines the time-series data in the cases to summarize typical impact patterns, such as sudden drops, linear increases, and step-like increases, as well as the duration of the impact; finally, it structures this common information to form core stress features.

[0027] It should be noted that the core stress feature set refers to the set of key attributes abstracted for each type of risk event, used to reproduce the event in simulation. It typically includes affected variables, impact intensity, time function, and recovery mode, and serves as a blueprint for building a computable event template.

[0028] It should be noted that the mathematical parameterization definition operation refers to transforming the qualitative description of the core stress feature set into a mathematical expression or parameter set that can be directly called by the simulation model. For example, for a supply chain shock event, the core stress feature of raw material costs increasing linearly to 150% within 3 months is parameterized as a piecewise function, and its mathematical expression is as follows:

[0029] In the formula, Indicates time The cost of raw materials, Indicates before the event occurred The cost benchmark value at any given time. This indicates the starting time point of the random event injection.

[0030] Furthermore, the cost baseline value refers to the instantaneous value of the state variable or external input variable representing the cost of raw materials in the model when the simplified dynamic model runs to the starting time point of the random event injection.

[0031] It should be noted that the event impact assessment algorithm is used to calculate the global weight of each parameterized stress event template. This weight reflects the relative importance of the event in the overall stress test. The mathematical expression of the event impact assessment algorithm is as follows:

[0032] In the formula, Event template The weight, This refers to events calculated based on historical data. The probability of occurrence, Indicates an event Assessment value of the potential impact on the financial health of typical enterprises. This indicates the total number of event types in the template library. Indicates the event type index. It is the sum of the products of the probability of occurrence and the intensity of influence of all events.

[0033] The potential impact strength assessment value of a typical enterprise's financial health is obtained by analyzing the percentage decline of the key financial health indicators of the affected enterprise relative to the pre-event level when similar events occur in the historical risk case data. By taking the median of the decline in multiple cases and normalizing it, the relative impact strength of the event, i.e., the potential impact strength assessment value, is obtained.

[0034] It should be noted that the standardized financial stress event template library refers to a data structure consisting of multiple stress event templates that have been mathematically parameterized and weighted. Each template is an independent, configurable, and repeatable stress test script, providing a systematic, standardized, and quantifiable set of external shock sources for subsequent Monte Carlo simulations. This enables the resilience assessment of financial strategies to be based on a set of predefined extreme scenarios with clear business implications and comparability.

[0035] S2. Based on the standardized time-series dataset, a simplified dynamic model of the enterprise financial system is constructed, and a set of observation indicators for quantifying the robustness of the system is defined. In this embodiment of the invention, the simplified dynamic model of the enterprise financial system is constructed based on the standardized time-series dataset, and a set of observation indicators for quantifying the robustness of the system is defined, including: Analyze the correlation of key variables in the standardized time series dataset to determine the set of state variables and the set of external input variables; Based on the principles of system dynamics modeling and the aforementioned set of state variables and set of external input variables, state-space equations are constructed to obtain a simplified dynamic model of the enterprise financial system. Based on the robustness analysis principle in control theory, a set of observation indicators is defined to quantify the system's resistance to shocks, recovery, and steady-state deviation. The set of observation indicators includes the maximum instantaneous deviation, the recovery time constant, and the percentage of steady-state error.

[0036] It should be noted that the operation of analyzing the correlation of key variables in the standardized time series dataset refers to using the standardized time series dataset to perform correlation analysis through Pearson correlation coefficient to identify core indicators with strong correlation and causal relationship in the enterprise's financial system.

[0037] Furthermore, the set of state variables refers to the set of core financial indicators selected to characterize the internal dynamic state of the system, such as cash on hand, net accounts receivable, and net inventory value. These variables evolve over time in the model.

[0038] It should be noted that the set of external input variables refers to the set of variables that affect the set of state variables, but whose own changes are mainly determined by external factors, such as industry prosperity index or market interest rate. In simulation, these variables are usually used as input parameters or driving variables of the model.

[0039] It should be noted that the operation of constructing a state-space equation set based on the principle of system dynamics modeling refers to establishing a difference equation set describing the rate of change of the state variable set based on the business logic and statistical relationship between financial variables.

[0040] Furthermore, the essence of this modeling process is to abstract the complex financial system into a set of mathematical equations to capture the dynamic feedback mechanism between key elements such as cash flow, assets, and liabilities. The mathematical expression of this state-space equation is as follows:

[0041] In the formula, Indicates at time The state variable vector, Indicates at time The external input variable vector, Here is the state transition matrix. For the input matrix, This represents the change over time.

[0042] It should be noted that the simplified dynamic model refers to a computational model composed of the state-space equations that can simulate the evolution of a company's key financial indicators over time under given external inputs and initial conditions.

[0043] It should be noted that the maximum instantaneous offset is an indicator used to quantify the system's shock resistance capability. It is defined as the maximum instantaneous absolute magnitude of the deviation of the system's key output variable from its reference path under the impact of a stress event. It is obtained by calculating the absolute value of the difference between the impacted path and the reference path at each time point, and then taking the maximum value among all these absolute values.

[0044] It should be noted that the recovery time constant is an indicator used to quantify the system recovery speed. Drawing on the concept of first-order system response, it is defined as the time required for the system output variable to recover to a value approximately 63.2% different from the new steady-state value from the moment it deviates from the baseline path after being impacted.

[0045] Furthermore, the recovery time constant measures how quickly a system recovers from a disturbance to normal operation; a smaller time constant indicates a faster recovery.

[0046] It should be noted that the steady-state error percentage is an indicator used to quantify the permanent performance loss caused by shocks. It is defined as the percentage of the relative deviation of the key output variable from the original steady-state value before the shock when the system reaches a new steady state after the shock. First, the difference between the new steady-state value and the original steady-state value is calculated. Then, the difference is divided by the original steady-state value and the absolute value is taken. Finally, it is multiplied by 100% to convert it into a percentage form.

[0047] It should be noted that the observation index set is a set of three specific indicators: maximum instantaneous offset, recovery time constant, and steady-state error percentage. It represents a calculable and comparable metric for quantifying the robustness of a company's financial system in the face of stress events from three dimensions: impact intensity, recovery speed, and permanent loss.

[0048] S3, for the pre-assessed financial strategy, runs a Monte Carlo simulation with event injection to obtain the baseline evolution path and stress response trajectory; In this embodiment of the invention, the step of running a Monte Carlo simulation with event injection for the pre-assessed financial strategy to obtain the baseline evolution path and stress response trajectory includes: Run the simplified dynamic model under the baseline scenario to generate the baseline evolution path of the strategy's key financial indicators; At random time points, events are selected from the financial stress event template library and injected into the simplified dynamic model. The simulation continues to run and the system response is recorded to generate a stress response trajectory.

[0049] It should be noted that running the simplified dynamic model under the baseline scenario means setting the external input variable parameters of the simplified dynamic model to values ​​that reflect the historical average level of the industry or a preset baseline level, and initializing the internal state of the model to typical values ​​based on historical data, thereby constructing a normal operating environment without external sudden stress events. The simplified dynamic model is driven to run iteratively under the baseline scenario. Under the above configuration, starting from the initial simulation time, the state variable values ​​at each future time point are calculated step by step according to the state space equation defined by the model, with a fixed step size (e.g., 1 month), until the preset total simulation time is reached, thereby simulating the natural evolution process of the enterprise financial system under a shock-free environment.

[0050] It should be noted that the benchmark evolution path of the key financial indicators of the strategy refers to the time series data sequence formed by the changes of core evaluation indicators such as return on net assets, net profit, and net operating cash flow over time in a single benchmark scenario simulation. It reflects the expected evolution trajectory of the core financial performance of the enterprise over time after adopting the financial strategy under an ideal external environment, and provides a reference benchmark for subsequent evaluation of the static return potential of the strategy.

[0051] It should be noted that, at random time points, the event injection time point is randomly determined within a preset time window. By using a uniformly distributed random number generator, a specific simulation time point is randomly selected within the simulation time period of the injectable event (e.g., from the 6th month to the 24th month) to simulate the uncertainty of the timing of stress events in the real world.

[0052] It should be noted that selecting an event from the financial stress event template library to inject into the simplified dynamic model means that when the simulation time reaches the event injection time point, one or more external input variables or key internal parameters of the simplified dynamic model are dynamically modified according to the mathematically parameterized rules in the selected template. For example, the raw material cost coefficient may be increased instantaneously or the market demand coefficient may be decreased, thereby reproducing the impact of the stress event in the model.

[0053] It should be noted that the phrase "continue to drive the model to run and record the temporal response of the variables corresponding to the observation index set" means that from the time point of event injection, under the continuous influence of the event, the iterative calculation of the simplified dynamic model continues until the simulation ends, and during this process, the values ​​of the original observation variables used to calculate the maximum instantaneous offset, recovery time constant, and steady-state error percentage are continuously tracked and recorded.

[0054] It should be noted that the stress response trajectory refers to the complete sequence of changes of the original observed variables over time recorded during the simulation phase after the event injection, as well as the initial values ​​of the maximum instantaneous offset, recovery time constant, and steady-state error percentage that can be calculated from it. This trajectory physically records the specific dynamic response of the enterprise's financial system under a single specific stress event test, and is the original data unit for quantitatively assessing the resilience of the strategy.

[0055] In this embodiment of the invention, the step of selecting an event injection model from the financial stress event template library at a random time point, continuing to run the simulation and record the system response to generate a stress response trajectory includes: A specific simulation time point is randomly selected within the simulation time period of the injectable event using a uniformly distributed random number generator; an event is selected from the financial stress event template library and injected into the simplified dynamic model; the parameters of the simplified dynamic model are then modified. Continue driving the model to run and record the time-series response of the variables corresponding to the observed index set to obtain the pressure response trajectory.

[0056] It should be noted that, at random time points, the event injection time point is randomly determined within a preset time window. By using a uniformly distributed random number generator, a specific simulation time point is randomly selected within the simulation time period of the injectable event (e.g., from the 6th month to the 24th month) to simulate the uncertainty of the timing of stress events in the real world.

[0057] It should be noted that selecting an event from the financial stress event template library to inject into the simplified dynamic model means that when the simulation time reaches the event injection time point, one or more external input variables or key internal parameters of the simplified dynamic model are dynamically modified according to the mathematically parameterized rules in the selected template. For example, the raw material cost coefficient may be increased instantaneously or the market demand coefficient may be decreased, thereby reproducing the impact of the stress event in the model.

[0058] It should be noted that the phrase "continue to drive the model to run and record the temporal response of the variables corresponding to the observation index set" means that from the time point of event injection, under the continuous influence of the event, the iterative calculation of the simplified dynamic model continues until the simulation ends, and during this process, the values ​​of the original observation variables used to calculate the maximum instantaneous offset, recovery time constant, and steady-state error percentage are continuously tracked and recorded.

[0059] It should be noted that the stress response trajectory refers to the complete sequence of changes of the original observed variables over time recorded during the simulation phase after the event injection, as well as the initial values ​​of the maximum instantaneous offset, recovery time constant, and steady-state error percentage that can be calculated from it. This trajectory physically records the specific dynamic response of the enterprise's financial system under a single specific stress event test, and is the original data unit for quantitatively assessing the resilience of the strategy.

[0060] S4. Based on the baseline evolution path, calculate the expected return score of each financial strategy. Based on the stress response trajectory and the observation index set, apply the strategy resilience score calculation algorithm to calculate the strategy resilience score of each financial strategy. In this embodiment of the invention, calculating the expected return score of each financial strategy based on the benchmark evolution path includes: Based on the aforementioned benchmark evolution path dataset, the key financial indicator values ​​for each benchmark evolution path in the final simulation cycle are extracted to obtain a sequence of key financial indicator values. Aggregate and statistically analyze the series of key financial indicator values ​​to obtain the benchmark expected return of the financial strategy; The benchmark expected return value is standardized and transformed based on a preset linear mapping function to obtain the expected return score of the financial strategy.

[0061] It should be noted that the operation of extracting the key financial indicator values ​​of each benchmark evolution path at the final simulation period refers to taking out the specific financial indicator values ​​at the end of the simulation from each time-series path of the benchmark evolution path dataset, such as return on net assets, net profit, or economic value added.

[0062] It should be noted that the key financial indicator value sequence refers to a set of key financial indicator values ​​for the corresponding number of final periods extracted from several independent benchmark simulations, where each value represents the final performance of the strategy under a risk-free shock scenario. This sequence reflects the statistical distribution of the strategy's results under normal conditions.

[0063] It should be noted that performing aggregated statistical calculations on the key financial indicator value series means calculating the statistics of the series to obtain a comprehensive representative value.

[0064] Furthermore, the aggregate statistical calculation typically employs the method of calculating mathematical expectation, and the formula for calculating mathematical expectation is as follows:

[0065] In the formula, This represents the expected value of the benchmark return. Indicates the first Key financial metrics for the final cycle of the second-benchmark simulation. This represents the total number of baseline simulations.

[0066] It should be noted that the benchmark expected return is a single value that integrates the results of multiple simulations. It reflects the statistical expected level of the key financial performance of the financial strategy under the baseline scenario without stress events, and reflects the average or typical return potential of the strategy.

[0067] It should be noted that the mathematical expression of the preset linear mapping function is as follows:

[0068] In the formula, This represents the calculated expected return score of the strategy. This represents the expected value of the benchmark return. and These are the maximum and minimum values ​​of the preset profit reference range, respectively. and These are the upper and lower limits of the preset score range, respectively.

[0069] It should be noted that the expected return score of the strategy is a standardized scalar value located in the preset range [0,10]. It represents the relative profitability of each financial strategy in a stable and shock-free environment in a uniform, unitless score form. The higher the score, the higher the expected financial return of the strategy under normal conditions.

[0070] In this embodiment of the invention, the step of calculating the strategy resilience score of each financial strategy based on the stress response trajectory and the observation index set, using a strategy resilience score calculation algorithm, includes: Based on the stress response trajectory dataset, for each financial strategy and each stress event type, the original values ​​of the observation indicator set are extracted and calculated from all stress response trajectories injected with the current stress event type to obtain the corresponding original value sequence of observation indicators. The expected value of each observation indicator is calculated by performing statistical calculation on the original value sequence of the observation indicators to obtain the expected value of each observation indicator for the event under the current strategy. Based on the strategy resilience score calculation algorithm, combined with the weights of each event in the financial stress event template library and the preset penalty coefficient, the expected values ​​of the observed indicators are weighted, aggregated, and converted inversely, and then normalized to obtain the strategy resilience score.

[0071] It should be noted that the operation of extracting and calculating the original values ​​of the observation indicators from all stress response trajectories injected with events refers to traversing the stress response trajectory dataset, filtering out all stress response trajectories injected with event types, and calculating the original values ​​of the maximum instantaneous offset, recovery time constant, and steady-state error percentage for a single simulation based on the time series data of the original observation variables recorded for each trajectory.

[0072] It should be noted that the original value sequence of the observed indicators refers to three data sequences consisting of a set of values, a set of values, and a set of values ​​obtained from multiple independent simulations for a specific strategy and event; these sequences reflect the statistical distribution of the robustness indicators of the strategy when facing multiple impacts of the same type of event.

[0073] It should be noted that the statistical expectation calculation of the original value sequence of the observed indicators refers to calculating the arithmetic mean of the original value sequence of each observed indicator to obtain a statistically stable expected value that characterizes the typical vulnerability level of the strategy to the event.

[0074] Furthermore, taking the maximum instantaneous offset as an example, the formula for calculating its expected value is as follows:

[0075] In the formula, Representation Strategy In the face of the event At that time, the expected value of the maximum instantaneous offset, Indicates the first The event was injected. The original value of the maximum instantaneous offset calculated in the simulation. This represents the total number of simulations that have been injected with event $e$.

[0076] Similarly, the expected values ​​of the recovery time constant and the percentage of steady-state error are obtained.

[0077] It should be noted that the expected values ​​of the observation indicators are three scalars that quantify the average vulnerability of a strategy in the face of different types of stress events across three dimensions: impact intensity, recovery speed, and permanent loss. The larger the expected value, the more vulnerable the strategy is on average in that dimension.

[0078] It should be noted that the normalization of the strategy resilience score means mapping the original values ​​calculated by all strategies to a preset standard interval (interval of [0,10]) that is comparable to the expected return score of the strategy.

[0079] In this embodiment of the invention, the mathematical expression of the strategy resilience score calculation algorithm is as follows:

[0080] In the formula, Representation Strategy Strategy resilience score, This represents a collection of templates for all financial stress events. Indicates a single stress event type. Indicates an event The preset weights satisfy , , , Representing strategies In the face of the event The expected values ​​of the maximum instantaneous offset, recovery time constant, and steady-state error percentage at that time. , , The preset penalty coefficients are used to adjust the decision-maker's relative aversion to the three vulnerability dimensions of instantaneous impact, recovery speed, and permanent loss, with values ​​of 0.4, 0.4, and 0.2, respectively.

[0081] S5 establishes a two-dimensional coordinate system with the strategy expected return score as the horizontal axis and the strategy resilience score as the vertical axis, plots all financial strategies in it, automatically calculates and identifies the Pareto optimal strategy set, and generates a resilience-return strategy report that includes strategy characteristic analysis.

[0082] In this embodiment of the invention, the establishment of a two-dimensional coordinate system with the expected return score of the strategy as the horizontal axis and the strategy resilience score as the vertical axis, plotting all financial strategies within it, automatically calculating and identifying the Pareto optimal strategy set, and generating a resilience-return strategy report containing strategy characteristic analysis includes: Establish a two-dimensional coordinate system, with the standardized expected return score of each financial strategy as the horizontal axis and the standardized strategy resilience score as the vertical axis, and plot the coordinate points corresponding to each strategy to obtain a scatter plot of the strategy distribution of all financial strategies on a two-dimensional plane. Based on the Pareto optimal front calculation algorithm, all coordinate points in the policy distribution scatter plot are analyzed to identify all coordinate points that are not dominated by other policies, so as to obtain the Pareto optimal policy point set and mark it in the policy distribution scatter plot. For each strategy in the Pareto optimal strategy point set, its corresponding baseline evolution path, stress response trajectory, key strategy parameters, and expected values ​​of observation indicators for various stress events are extracted. Based on a preset report template, a resilience-return strategy report containing revenue and resilience characteristic analysis is generated.

[0083] It should be noted that the operation of establishing a two-dimensional coordinate system provides a visual analysis plane for multi-objective decision-making. Each axis is standardized to the same dimension to ensure that all strategy points are comparable on this plane.

[0084] It should be noted that the strategy distribution scatter plot refers to a graphical representation in which each financial strategy is positioned as a discrete point in the two-dimensional coordinate system based on its calculated return score and resilience score. It intuitively shows the overall distribution pattern of all strategies to be evaluated in the two dimensions of return and resilience, helping decision-makers to quickly identify strategy clusters with high returns, high resilience, or specific trade-offs.

[0085] It should be noted that the Pareto optimal front calculation algorithm is an algorithm for identifying the optimal trade-off solution set in multi-objective optimization. In this two-dimensional plane, the criterion for determining whether one policy point dominates another policy point is: the payoff score of the dominant policy point is not lower than that of the dominated policy point and the resilience score is also not lower than that of the dominated policy point, and it is strictly superior to the dominated policy point in at least one of these aspects. Specifically, by traversing all policy points in the scatter plot, for each policy point, we check whether there is another policy point that can dominate it. We then filter out all policy points that are not dominated by any other points. The set of these points is the Pareto optimal policy point set. These points usually form a leading curve or boundary in the upper right of the distribution area on the graph.

[0086] It should be noted that the Pareto optimal strategy set refers to the set of strategies that have achieved the optimal trade-off between the two objectives of benefit and resilience among all current evaluation strategies, and cannot be improved simultaneously.

[0087] It should be noted that the operation of extracting the corresponding data and generating a resilience-return strategy report refers to extracting key information from the original simulation data and evaluation results of each strategy on the Pareto front and performing structured analysis.

[0088] Furthermore, the report typically includes the following core feature analysis: Benchmark performance analysis: Based on its benchmark evolution path, this section describes the typical financial performance of the strategy in a no-shock environment (such as expected return level and growth trend). Resilience Source Analysis: Based on its stress response trajectory and the expected values ​​of observation indicators for various events, this study specifically explains which dimension of the strategy stands out in terms of shock resistance, rapid recovery, and low loss when facing different types of stress events such as supply chain shocks and credit tightening, revealing the specific sources of its resilience advantages. Risk warning: Also based on the expected value of the observed indicators, this article points out the relative weaknesses that this strategy may have when facing certain types of stress events.

[0089] It should be noted that the resilience-benefit strategy report is a decision support document that integrates quantitative scores, visualized distributions, Pareto optimal set identification, and in-depth characteristic analysis. It is used to transform the aforementioned complex simulation, calculation, and multi-objective analysis results into insights and suggestions that decision-makers can directly understand and use for final strategy selection, thus completing a closed loop from data analysis to decision support.

[0090] In the several embodiments provided by this invention, it should be understood that the disclosed method can be implemented in other ways.

[0091] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0092] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, and technology that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A financial intelligent decision support method that integrates business and industry data, characterized in that, The method includes: S1 integrates internal business data and external industry data and preprocesses them to obtain a standardized time-series dataset. Based on historical risk cases and industry knowledge, a standardized financial stress event template library is built. S2, Based on the standardized time-series dataset, construct a simplified dynamic model of the enterprise financial system and define a set of observation indicators for quantifying the robustness of the system; S3, for the pre-assessed financial strategy, runs a Monte Carlo simulation with event injection to obtain the baseline evolution path and stress response trajectory; S4. Based on the baseline evolution path, calculate the expected return score of each financial strategy. Based on the stress response trajectory and the observation index set, apply the strategy resilience score calculation algorithm to calculate the strategy resilience score of each financial strategy. S5 establishes a two-dimensional coordinate system with the strategy expected return score as the horizontal axis and the strategy resilience score as the vertical axis, plots all financial strategies in it, automatically calculates and identifies the Pareto optimal strategy set, and generates a resilience-return strategy report that includes strategy characteristic analysis.

2. The financial intelligent decision support method integrating business and industry data as described in claim 1, characterized in that, The process of integrating internal business data and external industry data and preprocessing them yields a standardized time-series dataset, including: Based on preset multi-source data acquisition instructions, internal business data is obtained from the enterprise's local database, and external industry data is extracted from external industry data service provider interfaces and macroeconomic databases to obtain multi-source heterogeneous datasets. The multi-source heterogeneous dataset is subjected to unified standardization processing to obtain a standardized time series dataset, wherein the unified standardization processing includes format conversion, data cleaning, and time series alignment.

3. The financial intelligent decision support method integrating business and industry data as described in claim 1, characterized in that, The standardized financial stress event template library, built based on historical risk cases and industry knowledge, includes: Based on a pre-set historical case database and industry knowledge base, extract historical risk case data containing risk descriptions, impact dimensions, and response processes; Stress features are extracted from the historical risk case data to obtain core stress feature sets for different risk types; The core pressure feature set is mathematically parameterized to obtain a parameterized pressure event template. Based on the event impact assessment algorithm, event weights are configured for each parameterized stress event template to obtain a standardized financial stress event template library.

4. The financial intelligent decision support method integrating business and industry data as described in claim 1, characterized in that, Based on the standardized time-series dataset, a simplified dynamic model of the enterprise financial system is constructed, and a set of observation indicators for quantifying the robustness of the system is defined, including: Analyze the correlation of key variables in the standardized time series dataset to determine the set of state variables and the set of external input variables; Based on the principles of system dynamics modeling and the aforementioned set of state variables and set of external input variables, state-space equations are constructed to obtain a simplified dynamic model of the enterprise financial system. Based on the robustness analysis principle in control theory, a set of observation indicators is defined to quantify the system's resistance to shocks, recovery, and steady-state deviation. The set of observation indicators includes the maximum instantaneous deviation, the recovery time constant, and the percentage of steady-state error.

5. The financial intelligent decision support method integrating business and industry data as described in claim 1, characterized in that, The aforementioned pre-assessed financial strategy is subjected to a Monte Carlo simulation with event injection to obtain a baseline evolution path and stress response trajectory, including: Run the simplified dynamic model under the baseline scenario to generate the baseline evolution path of the strategy's key financial indicators; At random time points, events are selected from the financial stress event template library and injected into the simplified dynamic model. The simulation continues to run and the system response is recorded to generate a stress response trajectory.

6. The financial intelligent decision support method integrating business and industry data as described in claim 5, characterized in that, At random time points, an event injection model is selected from the financial stress event template library, the simulation continues to run and the system response is recorded to generate a stress response trajectory, including: A specific simulation time point is randomly selected within the simulation time period of the injectable event using a uniformly distributed random number generator; an event is selected from the financial stress event template library and injected into the simplified dynamic model; the parameters of the simplified dynamic model are then modified. Continue driving the model to run and record the time-series response of the variables corresponding to the observed index set to obtain the pressure response trajectory.

7. The financial intelligent decision support method integrating business and industry data as described in claim 1, characterized in that, The calculation of the expected return score for each financial strategy based on the benchmark evolution path includes: Based on the aforementioned benchmark evolution path dataset, the key financial indicator values ​​for each benchmark evolution path in the final simulation cycle are extracted to obtain a sequence of key financial indicator values. Aggregate and statistically analyze the series of key financial indicator values ​​to obtain the benchmark expected return of the financial strategy; The benchmark expected return value is standardized and transformed based on a preset linear mapping function to obtain the expected return score of the financial strategy.

8. The financial intelligent decision support method integrating business and industry data as described in claim 1, characterized in that, Based on the stress response trajectory and the set of observed indicators, the strategy resilience score of each financial strategy is calculated using a strategy resilience score calculation algorithm, including: Based on the stress response trajectory dataset, for each financial strategy and each stress event type, the original values ​​of the observation indicator set are extracted and calculated from all stress response trajectories injected with the current stress event type to obtain the corresponding original value sequence of observation indicators. The expected value of each observation indicator is calculated by performing statistical calculation on the original value sequence of the observation indicators to obtain the expected value of each observation indicator for the event under the current strategy. Based on the strategy resilience score calculation algorithm, combined with the weights of each event in the financial stress event template library and the preset penalty coefficient, the expected values ​​of the observed indicators are weighted, aggregated, and converted inversely, and then normalized to obtain the strategy resilience score.

9. A financial intelligent decision support method integrating business and industry data as described in claim 4, characterized in that, The mathematical expression for the strategy resilience score calculation algorithm is as follows: In the formula, Representation Strategy Strategy resilience score, This represents a collection of templates for all financial stress events. Indicates a single stress event type. Indicates an event The preset weights satisfy , , , Representing strategies In the face of the event The expected values ​​of the maximum instantaneous offset, recovery time constant, and steady-state error percentage at that time. , , The preset penalty coefficients are used to adjust the decision-maker's relative aversion to the three vulnerability dimensions of instantaneous impact, recovery speed, and permanent loss, with values ​​of 0.4, 0.4, and 0.2, respectively.

10. The financial intelligent decision support method integrating business and industry data as described in claim 1, characterized in that, The system establishes a two-dimensional coordinate system with the expected return score of the strategy as the horizontal axis and the strategy resilience score as the vertical axis. All financial strategies are plotted within this system, and the Pareto-optimal strategy set is automatically calculated and identified. A resilience-return strategy report containing strategy characteristic analysis is generated, including: Establish a two-dimensional coordinate system, with the standardized expected return score of each financial strategy as the horizontal axis and the standardized strategy resilience score as the vertical axis, and plot the coordinate points corresponding to each strategy to obtain a scatter plot of the strategy distribution of all financial strategies on a two-dimensional plane. Based on the Pareto optimal front calculation algorithm, all coordinate points in the policy distribution scatter plot are analyzed to identify all coordinate points that are not dominated by other policies, so as to obtain the Pareto optimal policy point set and mark it in the policy distribution scatter plot. For each strategy in the Pareto optimal strategy point set, its corresponding baseline evolution path, stress response trajectory, key strategy parameters, and expected values ​​of observation indicators for various stress events are extracted. Based on a preset report template, a resilience-return strategy report containing revenue and resilience characteristic analysis is generated.

Images