A data model-based financial asset dynamic management method and system

Abstract

This invention discloses a data model-based dynamic management method and system for financial assets, belonging to the field of financial technology. The method includes: constructing a unified and trustworthy data foundation by generating data passports with confidence labels for multi-source financial data, and generating a unified digital profile integrating the confidence labels based on these passports; constructing a dynamic correlation graph using the profiles as nodes; executing hierarchical decision-making steps by generating financial asset operation instructions for different management dimensions through parallel micro-control loops, meso-control loops, and macro-control loops; and monitoring the execution results through time-lapse and continuous learning feedback steps, triggering a re-decision closed loop in response to loss events. This invention achieves full-spectrum, three-dimensional dynamic management from the instruction level to the system level, significantly improving risk identification accuracy, response speed, and system automation level.

Description

Technical Field

[0001] This invention relates to a method and system for dynamic management of financial assets based on a data model, belonging to the field of financial technology. Background Technology

[0002] With the increasing complexity of financial markets and the explosive growth of data volume, asset management institutions face unprecedented challenges. Traditional asset management methods often rely on static rules and human experience, making it difficult to cope with the rapidly changing market environment. While some data-driven risk monitoring systems have emerged, they generally suffer from the following problems: The lack of a unified trust measurement mechanism for multi-source and heterogeneous financial data (market data, public opinion data, macroeconomic data, etc.) leads to low-quality data directly entering the decision-making process, resulting in misjudgments and erroneous transactions.

[0003] Most systems can only trigger isolated responses to single risk events (such as directly selling risky assets), lacking the ability to respond to risks in a multi-layered and coordinated manner, from single orders to portfolios and then to cross-asset systems.

[0004] Risk identification, decision-making, execution, and post-event correction are often disconnected, making it impossible to form an effective learning and evolution mechanism, and the system is difficult to optimize itself.

[0005] Therefore, how to build a dynamic management system for financial assets that can automatically sense market changes, make multi-level collaborative decisions, and have the ability to self-evolve has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0006] To address the aforementioned technical issues, this invention provides a data model-based method and system for dynamic management of financial assets, enabling closed-loop management across the entire chain, from data governance to intelligent decision-making and continuous evolution.

[0007] According to one aspect of the present invention, a dynamic management method for financial assets based on a data model is provided, comprising: a hierarchical decision-making step, which generates financial asset operation instructions for different management dimensions through parallel micro-control loops, meso-control loops and macro-control loops respectively; The micro-control loop is used to dynamically regulate the execution level of a single transaction instruction; the meso-control loop is used to dynamically adjust the strategy parameters of the investment portfolio; and the macro-control loop is used to identify risk transmission paths and initiate cross-asset collaborative risk buffering operations based on the correlation network between assets.

[0008] Optionally, the hierarchical decision-making step further includes a step of constructing a unified and trusted data base, which includes: generating data passports with confidence labels for multi-source financial data; generating unified digital profiles of various types of financial assets that incorporate the confidence labels based on the data with the data passports; and constructing and updating a graph to represent the dynamic relationship between assets, using the unified digital profiles as nodes, as a relationship network between the assets.

[0009] Optionally, the unified digital profile is a homogeneous time-series feature vector across asset classes, and the confidence label is embedded in the vector as a mandatory feature dimension and serves as the input feature for subsequent artificial intelligence models.

[0010] Optionally, the micro-control loop dynamically regulates the execution level of a single transaction instruction, including: selecting one of multiple preset execution channels to execute the instruction based on the confidence level label of the asset corresponding to the transaction instruction and the real-time market status; wherein, the multiple preset execution channels include: a direct channel for high confidence scenarios, a shadow verification channel for medium confidence scenarios, and a circuit breaker channel for low confidence or high-risk scenarios.

[0011] Optionally, the meso-control loop dynamically adjusts the strategy parameters of the portfolio, including: dynamically adjusting the portfolio weights through a robust optimization engine based on the topological stability index of the correlation network between the assets; the robust optimization engine dynamically sets an uncertainty interval according to the confidence label of the assets, wherein the width of the interval is inversely proportional to the confidence level.

[0012] Optionally, the macro control loop initiates a cross-asset collaborative risk buffering operation, including: in response to a risk warning signal, identifying high-risk transmission paths based on the network of relationships between the assets; selecting healthy asset nodes from the network whose topological distance from the risk source is greater than a preset threshold as collaborative buffers; and generating a configuration adjustment instruction for at least one asset in the collaborative buffer to reserve a preset proportion of liquidity in the buffer assets.

[0013] Optionally, the assets in the collaborative buffer must meet the following conditions: the topological distance between the risk source asset and the risk source asset in the association network is greater than or equal to two hops; and the current volatility is lower than the industry preset threshold.

[0014] Optionally, the method further includes: a time-travel and continuous learning feedback step, monitoring the execution results of the financial asset operation instructions, and in response to the detection of a loss event, rolling back to the state before the loss occurred and re-executing the hierarchical decision-making steps to generate hedging instructions.

[0015] According to another aspect of the present invention, a data model-based dynamic management system for financial assets is provided, comprising: The hierarchical decision control engine is configured to generate financial asset operation instructions for different management dimensions through parallel micro control loops, meso control loops, and macro control loops. The micro-control loop is configured to dynamically regulate the execution level of a single transaction instruction; the meso-control loop is configured to dynamically adjust the strategy parameters of the investment portfolio; and the macro-control loop is configured to identify risk transmission paths and initiate cross-asset collaborative risk buffering operations based on the network of relationships between assets.

[0016] Optionally, the system further includes: a unified trusted data base module, configured to generate data passports with confidence labels from multi-source financial data, generate a unified digital profile that integrates the confidence labels based on the data with the data passports, and construct a graph representing the dynamic relationship between assets using the unified digital profile as nodes; and a time-travel and continuous learning feedback module, configured to monitor execution results, and trigger the hierarchical decision control engine to re-determine and generate hedging instructions in response to loss events.

[0017] According to another aspect of the present invention, an electronic device is provided, including a processor and a memory, the memory storing a computer program that, when executed by the processor, implements the method described in any of the preceding claims.

[0018] According to another aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method described in any of the preceding claims.

[0019] Compared with the prior art, the beneficial effects of the present invention are as follows: By constructing an overall technical architecture that integrates a seven-layer logical architecture and a three-layer collaborative control loop, and using a data model with a unified digital profile and a dynamic correlation graph as the hub of the entire system, full-spectrum, three-dimensional dynamic management from the instruction level to the system level is achieved, which significantly improves the accuracy of risk identification, response speed and system automation level. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0021] Figure 1A schematic diagram of the overall architecture of a dynamic financial asset management system provided in an embodiment of the present invention; Figure 2 A flowchart of a dynamic financial asset management method provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the execution channel selection for the microcontroller loop in one embodiment of the present invention; Figure 4 This is a schematic diagram of the collaborative buffering of the macro control loop in one embodiment of the present invention; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Example 1: like Figure 1 As shown, an embodiment of the present invention provides a data model-based dynamic management system for financial assets, employing a seven-layer logical architecture, from bottom to top: a data trust access and governance layer, a unified asset digital profiling and association layer, an intelligent perception and predictive analysis engine layer, an adaptive decision-making and three-layer execution control layer, an automated strategy execution and transaction layer, a time-travel and continuous learning feedback layer, and a unified operation and maintenance monitoring and visualization layer. The adaptive decision-making and three-layer execution control layer, which embeds three parallel control loops—micro, meso, and macro—is the core decision-making unit of the present invention.

[0024] Example 2: This embodiment provides a method for dynamic management of financial assets based on a data model, such as... Figure 2 As shown, the method includes the following steps: Step S1, which establishes a unified and trusted data foundation, provides a high-quality, standardized data basis for all subsequent analyses. Specifically, it includes: S101: Receives real-time heterogeneous financial data streams from multiple sources, including market data, public opinion, macroeconomics, and supply chain data, as input data.

[0025] S102. A dynamic confidence assessment is performed on each incoming data entry using the data passport module. Specifically, this assessment process includes: Two core metrics are calculated: Source Reputation Score and Statistical Anomaly Score. The Source Reputation Score is dynamically weighted based on the data provider's historical accuracy. The Statistical Anomaly Score uses the Isolation Forest algorithm to detect whether the current data value is an outlier, and then calculates it according to the formula... Generate a comprehensive confidence score. The Data Passport module outputs a standardized data passport label for each data point, which includes the data value, confidence score, and quality level based on the confidence score.

[0026] S103, the data passport module performs intelligent dynamic routing integrating stream and batch processing based on the confidence level: For data with a confidence level ≥ 0.8 and a quality level of High, it is directly sent to the high-frequency processing channel to ensure that the processing delay is less than 10ms, which is used for trading signals with extremely high timeliness requirements.

[0027] For data with a confidence level of 0.5 ≤ confidence level < 0.8 and a quality level of Mid, the data is fed into the Flink stream processing engine for real-time cleaning and enhancement, with latency controlled within 500ms.

[0028] For data with a confidence level of <0.5 and a quality level of Low, the data is transferred to the actuarial layer for in-depth processing to generate training datasets or perform complex validation.

[0029] Through the above processing, this step outputs a highly reliable data stream with a unified quality label, which can reduce the erroneous transaction rate caused by data errors by 76%.

[0030] Step S2, which generates a unified digital profile and dynamic correlation map, uses the high-reliability data stream output from S1 as input to construct a dual-core data model, specifically including: S201. Generate a unified digital profile for each type of financial asset. Specifically, fuse the micro-level three-dimensional behavioral trajectory of each asset with the confidence level label obtained in step S1. The micro-level three-dimensional behavioral trajectory includes: price changes. Transaction volume cash flow This forms a standardized, isomorphic temporal feature vector, denoted as... The confidence label, as a mandatory feature dimension embedding, endows the profile with inherent credibility attributes, providing input for subsequent AI models.

[0031] S202. Using unified digital profiles as nodes, a dynamic relationship graph is constructed. This graph is built upon multi-source relationship data, including equity linkages, supply chain dependencies, statistical correlations, and joint holding behaviors. The weight of each edge is inversely proportional to the distance calculated based on topological data analysis, and is updated in real-time via minute-level stream computing. Specifically, the output of this dynamic relationship graph is a dynamically updated network structure where nodes represent asset profiles, and edges represent the strength of associations between assets, enabling real-time mapping of potential risk transmission paths within the asset network.

[0032] The output of this step is a dual-core data base that couples profiles and graphs, where each asset is represented as a homogenized vector with a confidence label, and these vectors are interconnected by dynamically updated edges.

[0033] The hierarchical decision-making step S3, based on the dual-core data base output from step S2, generates operational instructions of different dimensions through three parallel control loops: micro, meso, and macro. Specifically, it includes: Micro-control loop: The input to this loop is a single transaction instruction about to be executed, along with a unified digital profile of the asset involved in the instruction and its real-time market status. Specifically, the processing logic of the micro-control loop is to jointly model the confidence labels and market status, and dynamically select the instruction execution channel based on the joint output results. Specifically: If the confidence level is ≥0.8 and the market is not congested, the direct channel is selected, and the instruction is sent directly to the exchange.

[0034] If the confidence level is 0.5 ≤ confidence level < 0.8 or the market is moderately congested, then the shadow verification channel is selected. The instruction will first enter the shadow account for simulated verification, and then be executed after the verification is successful.

[0035] If the confidence level is less than 0.5 or the market is severely congested, the circuit breaker channel will be selected to immediately suspend the order and dynamically lower the single transaction permission threshold for the user or the asset.

[0036] The output of the micro-control loop is the specific execution instruction or circuit breaker notification after channel selection.

[0037] Meso-control loop: The inputs to this loop are the current state of the entire portfolio, the dynamic correlation graph output from step S2, and the asset confidence information output from step S1. Specifically, the goal of the meso-control loop is to dynamically adjust the portfolio's strategy parameters, such as weights, leverage ratios, and stop-loss levels. The output of the meso-control loop is a new set of portfolio parameters, such as updated asset allocation ratios and leverage ratios. The specific processing includes: Initial weights for the portfolio are generated based on the Proximal Policy Optimization (PPO) algorithm; The initial weights are refined using the Covariance Matrix Adaptive Evolutionary Strategy (CMA-ES), and the fitness function incorporates the topological stability index of the graph, ensuring that the adjusted combination has stronger robustness on the graph. Dynamic robust optimization is performed by dynamically setting the uncertainty range for expected return and risk of each asset based on its confidence level label: the higher the confidence level, the narrower the range; the lower the confidence level, the wider the range. In one embodiment, for an asset with a confidence level of 0.9, its expected return range is set to... For assets with a confidence level of 0.6, the interval is set to... Then, we solve the problem of maximizing the portfolio return in the worst-case scenario to obtain the final portfolio weights and execution confidence levels.

[0038] Macro-control loop: The inputs to this loop are the dynamic correlation graph output from step S2 and risk warning signals from the intelligent perception layer. The goal of the macro-control loop is to initiate cross-asset collaborative risk buffering, rather than directly addressing the risk source itself. The output of the macro-control loop is cross-asset allocation adjustment instructions for buffer assets, specifically including: Upon receiving a risk warning for a specific risk source asset, the system simulates the transmission path of the risk in the network based on a dynamic correlation graph, identifying first-order and second-order correlation nodes that may be impacted.

[0039] Healthy assets that meet the criteria are selected from these related nodes as a collaborative buffer. The selection criteria include: the topological distance from the risk source in the graph is greater than or equal to 2 hops; and the current volatility of the asset is more than one standard deviation below the industry average.

[0040] Once buffer assets are selected, the macro control loop calculates configuration adjustment instructions for these buffer assets through an optimization model. These instructions aim to require the buffer assets to retain a certain proportion of liquidity to absorb potential shocks, thereby enhancing the system resilience of the entire asset network.

[0041] Step S4, the time rewind and continuous learning feedback process, uses the results of all executed instructions from step S3, market feedback data, and system logs as input to construct a self-correction and evolutionary closed loop for the system. Specifically, it includes: The system continuously monitors the execution results of instructions. Specifically, once a loss event caused by historical data errors, momentary model inaccuracies, or decision biases is detected, the time machine mechanism is immediately triggered. This mechanism automatically rolls back the system state to the moment before the loss occurred and calls upon the latest unified digital profile and dynamic correlation graph to re-execute the three-layer joint decision-making process in step S3. The result of the re-decision is a set of hedging instructions, which are sent to the execution layer for real-time intraday compensation.

[0042] Simultaneously, all execution results and compensation effect data are collected to form a high-quality feedback dataset. This data is used for online training and iterative optimization of the profile update algorithm in step S2, the AI ​​model in step S3, and the decision rules of the control loop, enabling the system to have continuous self-learning capabilities. The AI ​​model can be PPO or CMA-ES.

[0043] The outputs of this step include: real-time generated hedging compensation instructions and feedback data for model updates. Through this mechanism, the automatic compensation rate for erroneous instructions can reach 87%, enabling the system to dynamically adapt to market changes and continuously improve management efficiency.

[0044] Example 3: Figure 3 A specific embodiment of the micro-control loop is illustrated: The system receives a buy order for stock A, amounting to 1 million yuan. At this time, the confidence level label in the unified digital profile of this stock is 0.9, and the market congestion prediction module outputs a normal market status. The micro-control loop takes the confidence level of 0.9 and the normal status as inputs, and through joint modeling, determines that the pass-through condition is met. Therefore, the order is marked as pass-through and sent directly to the transaction execution system.

[0045] In another embodiment, the confidence level of the same instruction is 0.6, and the market condition is moderately congested, meaning that within 5 consecutive minutes, the number of assets with an order book depth of less than 5 levels and a large order execution rate exceeding 90% has reached 50. In this case, the joint modeling result is to enable the shadow verification channel. The instruction is sent to a shadow account for simulated trading, while the system monitors the simulation results. If no anomalies occur in the simulated trading within a preset time, the instruction is automatically converted to live trading; otherwise, the instruction will be terminated and an alert will be sent to the user.

[0046] In other embodiments, if the confidence level label is only 0.3 and the market is severely congested, the circuit breaker is triggered directly. The instruction is rejected, and the system automatically and temporarily lowers the user's single transaction limit from 1 million yuan to 700,000 yuan, and records this event in the log for subsequent auditing.

[0047] Example 4: Cooperative Buffering of Macro-Control Loop Figure 4 A specific embodiment of the macro-control loop is illustrated, in which node A is identified as a risk source in the dynamic correlation graph. The macro-control loop takes this graph and early warning signals as inputs.

[0048] The system first calculates nodes B, C, and D that are 1 hop away from node A, and nodes E, F, and G that are 2 hop away. Next, it filters buffer assets: nodes B and C, although 1 hop away, have significantly increased volatility and do not meet the criteria; node D, although 1 hop away, has normal volatility, but considering its direct correlation, it is not selected to avoid overreaction; nodes E and F are selected as collaborative buffers.

[0049] Subsequently, the optimization model calculates configuration adjustment instructions: requiring the asset manager at node E to increase its cash holdings from 1% to 3%, and node F to increase its cash holdings from 0% to 2%. These two instructions are issued through an automated execution pipeline. When a risk shock propagates from A to E and F, the liquidity reserved by E and F can be used to absorb part of the shock, specifically for purchasing affected assets at a low price or as margin, thereby preventing the entire system from collapsing due to a chain reaction.

[0050] Example 5: This embodiment provides an electronic device, such as... Figure 5 As shown, the electronic device includes a central processing unit (CPU), which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) or loaded from storage into random access memory (RAM). The RAM also stores various programs and data required for the operation of the electronic device. The CPU, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0051] The following components are connected to the I / O interface: input sections including keyboards, mice, etc.; output sections including cathode ray tubes (CRTs), liquid crystal displays (LCDs), and speakers; storage sections including hard disks; and communication sections including network interface cards such as LAN cards and modems. The communication sections perform communication processing via networks such as the Internet. Drives are also connected to the I / O interface as needed. Removable media, such as disks, optical disks, magneto-optical disks, semiconductor memories, etc., are installed on the drive as needed so that computer programs read from them can be installed into the storage section as required.

[0052] In particular, according to the embodiments disclosed in this invention, the above references Figure 2The described process can be implemented as a computer software program. Specifically, embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowchart. In such embodiments, the computer program can be downloaded and installed from a network via a communication component, and / or installed from a removable medium. When the computer program is executed by a central processing unit, it performs the functions defined in the system of this application.

[0053] It should be noted that the computer-readable medium shown in this invention can be a computer-readable signal medium, a computer-readable storage medium, or any combination thereof. Computer-readable storage media include, but are not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specifically, computer-readable storage media include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including, but not limited to, electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0054] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for dynamic management of financial assets based on a data model, characterized in that, include: The hierarchical decision-making process generates financial asset operation instructions for different management dimensions through parallel micro-control loops, meso-control loops, and macro-control loops. The micro-control loop is used to dynamically regulate the execution level of a single transaction instruction. The meso-level control loop is used to dynamically adjust the strategy parameters of the investment portfolio; The macro-control loop is used to identify risk transmission paths and initiate cross-asset collaborative risk buffering operations based on the network of relationships between assets.

2. The method for dynamic management of financial assets based on a data model according to claim 1, characterized in that, The hierarchical decision-making step is preceded by a step of building a unified and trusted data foundation, which includes: Generate data passports with confidence labels for multi-source financial data; Based on the data with the aforementioned data passport, a unified digital profile integrating the aforementioned confidence labels is generated for various types of financial assets; Using the unified digital profile as nodes, a graph is constructed and updated to represent the dynamic relationships between assets, serving as a network of relationships between the assets.

3. The method for dynamic management of financial assets based on a data model according to claim 2, characterized in that, The unified digital profile is a homogeneous time-series feature vector across asset classes. The confidence label is embedded in this vector as a mandatory feature dimension and serves as the input feature for subsequent artificial intelligence models.

4. The method for dynamic management of financial assets based on a data model according to claim 1, characterized in that, The micro-control loop dynamically regulates the execution level of a single transaction instruction, including: Based on the confidence level label of the asset corresponding to the trading instruction and the real-time market status, one of the multiple preset execution channels is selected to execute the instruction; The multiple preset execution channels include: a direct channel for high-confidence scenarios, a shadow verification channel for medium-confidence scenarios, and a circuit breaker channel for low-confidence or high-risk scenarios.

5. The method for dynamic management of financial assets based on a data model according to claim 1, characterized in that, The meso-level control loop dynamically adjusts the strategy parameters of the investment portfolio, including: Based on the topological stability index of the relationship network between the assets, the weights of the portfolio are dynamically adjusted through a robust optimization engine. The robust optimization engine dynamically sets an uncertainty range based on the asset's confidence level label, wherein the width of the range is inversely proportional to the confidence level.

6. The method for dynamic management of financial assets based on a data model according to claim 1, characterized in that, The macro-control loop initiates cross-asset coordinated risk buffering operations, including: In response to risk warning signals, high-risk transmission paths are identified based on the network of relationships between the assets; Healthy asset nodes that are more than a preset threshold in topological distance from the risk source are selected from the network as collaborative buffers; Generate a configuration adjustment instruction for at least one asset in the collaborative buffer to reserve a preset proportion of liquidity in the buffer assets.

7. The method for dynamic management of financial assets based on a data model according to claim 6, characterized in that, The assets in the collaborative buffer must meet the following conditions: The topological distance between the risk source asset and the asset in the aforementioned network of relationships is greater than or equal to two hops; and Current volatility is below the industry's preset threshold.

8. The method for dynamic management of financial assets based on a data model according to claim 1, characterized in that, Also includes: The time-rewinding and continuous learning feedback steps monitor the execution results of the financial asset operation instructions. In response to the detection of a loss event, the system rolls back to the state before the loss occurred and re-executes the hierarchical decision-making steps to generate hedging instructions.

9. A dynamic management system for financial assets based on a data model, characterized in that, include: The hierarchical decision control engine is configured to generate financial asset operation instructions for different management dimensions through parallel micro control loops, meso control loops, and macro control loops. The micro-control loop is configured to dynamically regulate the execution level of a single transaction instruction. The meso-level control loop is configured to dynamically adjust the strategy parameters of the investment portfolio; The macro-control loop is configured based on a network of relationships between assets to identify risk transmission paths and initiate cross-asset collaborative risk buffering operations.

10. A data model-based dynamic management system for financial assets according to claim 9, characterized in that, Also includes: The unified trusted data base module is configured to generate data passports with confidence labels from multi-source financial data, generate a unified digital profile that integrates the confidence labels based on the data with the data passports, and construct a graph to represent the dynamic relationship between assets using the unified digital profile as nodes. as well as The time rewind and continuous learning feedback module is configured to monitor execution results and, in response to loss events, trigger the hierarchical decision control engine to re-determine and generate hedging instructions.

Images