Intelligent fund routing system based on AI large model and optimization method thereof

Abstract

The invention relates to an intelligent fund routing system based on an AI large model and an optimization method thereof, and belongs to the field of finance, and the system comprises a data processing module which is used for automatically collecting financial market data, and carrying out the preprocessing and feature extraction; the decision forming module is used for analyzing the processed data and performing fund routing decision based on an AI model architecture; the fund routing optimization module is used for optimizing a fund routing path according to market and user behavior prediction results; the risk assessment module is used for monitoring transactions and fund flow directions in real time and assessing potential risks; and the risk management integration module is used for carrying out early warning and decision support according to the risk assessment result. The method has the advantages that the efficiency is improved; the safety is enhanced; the precision is improved, and the deep learning model can better predict the market trend and the user behavior.

Description

Technical Field

[0001] The present invention relates to an intelligent funds routing system based on an AI big model and an optimization method thereof, and belongs to the financial field. Background Art

[0002] The current financial market is facing unprecedented complexity and volatility. Traditional fund routing systems are mostly rule-based. Even with the support of some machine learning algorithms, it is often difficult to fully understand the complex financial market dynamics and user behavior patterns due to the limitations of data preprocessing and feature extraction. This leads to frequent inefficiencies and slow responses in the allocation and guidance of funds, which cannot meet the needs of the rapidly changing financial market. It has the following disadvantages:

[0003] Limited data processing capabilities: Traditional funds routing systems have limitations in automating the collection and processing of financial market data, especially when dealing with massive and diverse data, and often lack efficient data analysis automation capabilities.

[0004] Lack of deep learning capabilities: Existing systems do not fully utilize deep learning models, especially in the financial field, which can learn from historical data and predict the flow of funds, while traditional systems are relatively weak in this regard.

[0005] Slow response speed: The existing system cannot meet the demand for real-time analysis and response to market changes, resulting in untimely optimization of fund allocation and inability to achieve optimal fund allocation.

[0006] Inaccurate prediction of user behavior: Traditional systems lack in-depth analysis of user behavior patterns and preferences, resulting in the inability to accurately predict users’ fund flow trends and intentions, thus affecting the accuracy of fund routing.

[0007] In order to overcome these challenges, many institutions have begun to explore the use of artificial intelligence technology to improve the efficiency and accuracy of fund routing. Among them, AI big models, with their powerful data processing and deep learning capabilities, have shown great potential in processing incomplete, unstructured, and large amounts of dynamically changing data. Summary of the invention

[0008] In order to overcome the defects of the prior art, the present invention provides an intelligent funds routing system based on an AI big model. The technical solution of the present invention is:

[0009] An intelligent funds routing system based on an AI big model, comprising:

[0010] Data processing module, used to automatically collect financial market data, and perform preprocessing and feature extraction;

[0011] The decision-making module, based on the AI ​​model architecture, is used to analyze the processed data and make fund routing decisions;

[0012] Fund routing optimization module, which is used to optimize the fund routing path based on market and user behavior prediction results;

[0013] Risk assessment module, used to monitor transactions and capital flows in real time and assess potential risks;

[0014] The risk management integration module is used to provide early warning and decision support based on risk assessment results.

[0015] The data processing module includes a data acquisition module and a data preprocessing and feature extraction module. The data acquisition module is used to automatically collect the required market data from different data sources; the data preprocessing and feature extraction module is used to preprocess and extract features from the collected raw data; wherein, the data acquisition module includes:

[0016] Automated API retrieval module, which is used to automatically extract real-time market data, asset prices, and market news information by integrating financial market APIs;

[0017] The social media monitoring module is used to monitor and collect public data related to financial markets on social media platforms, analyze investor sentiment through natural language processing technology, and thus obtain market sentiment indicators;

[0018] The historical database module is used to obtain historical data of the financial market, provide rich historical data support for model training, and help the model learn historical market trends and patterns.

[0019] The data preprocessing and feature extraction module includes:

[0020] The outlier detection and processing module is used to identify outliers in the data and remove or correct them to ensure the quality of input data and reduce noise interference during model training. The digital feature conversion module is used to process unstructured text data and convert it into numerical features that can be processed by the model.

[0021] The time series analysis module is used to process time-sensitive financial market data to extract features that reflect market trends and periodicity.

[0022] The decision-making module includes a deep learning module and a dynamic adjustment module. The deep learning module is used to realize prediction. The deep learning module includes:

[0023] Multi-head attention mechanism module, used to enhance the model’s ability to understand the context of market information;

[0024] Convolutional neural network module for graph feature extraction;

[0025] The long short-term memory network module is used to capture the dependencies in the time series data; the dynamic adjustment module is used to adapt to the changes in the market, and the dynamic adjustment module includes a model iteration module, which is used to update the model in real time to adapt to the changes in the financial market;

[0026] Multi-model integration module for real-time monitoring and evaluation of the performance of several models.

[0027] The fund routing optimization module includes an algorithm design module and a user behavior prediction module. The algorithm design module is used to balance several objectives using a multi-objective optimization algorithm; the user behavior prediction module is used to autonomously learn the user's historical transaction patterns and preferences and provide the user with fund routing suggestions; the specific prediction steps of the user behavior prediction module are:

[0028] 4.1 Analysis of historical transaction data:

[0029] By analyzing the user's historical transaction data, we can understand their transaction behavior, including the user's transaction frequency, transaction amount, selected capital routing path, and transaction time. This step provides basic data for subsequent user profile modeling.

[0030] 4.2 User portrait modeling: Based on historical transaction data, use AI modeling technology to build user portraits;

[0031] 4.3 Personalized recommendation: After the user portrait is built, the fund routing path that best suits the needs of different users is recommended based on their characteristics and preferences.

[0032] The risk assessment modules include:

[0033] Data input and preprocessing module, which is responsible for collecting transaction, user and market data, and cleaning and formatting them for subsequent feature extraction and risk assessment;

[0034] Feature extraction module, which is used to extract key features from raw data and use these features as input to the risk assessment model;

[0035] The risk assessment model module is used to assess different types of risks through different sub-models;

[0036] The credit risk assessment module is used to train the credit scoring model based on the user's credit data and adjust it according to the user's economic status;

[0037] Liquidity risk assessment module, which uses historical transaction and market data to predict liquidity risk and set early warning limits;

[0038] The market risk assessment module is used to predict market fluctuations using the time series analysis module; the comprehensive risk assessment module is used to combine the assessment results of each sub-model and conduct an overall risk assessment.

[0039] The specific assessment method of the comprehensive risk assessment module is as follows:

[0040] 5.1 Risk weight allocation: assign weights to different risk factors and use weighted methods to synthesize risk values;

[0041] 5.2 Risk level prediction: convert risk value into risk level and set threshold;

[0042] 5.3 Real-time feedback: Feedback risk assessment results to the transaction decision-making system in real time and issue an early warning when necessary.

[0043] The specific method for extracting key features by the feature extraction module includes:

[0044] 6.1 Transaction characteristics: analysis of transaction amount, credit scores of both parties, historical transaction behavior, and transaction frequency;

[0045] 6.2 User characteristics: evaluate the user's historical transaction behavior, credit score and behavior prediction;

[0046] 6.3 Market characteristics: monitor real-time market price fluctuations, macroeconomic indicators and market participants’ confidence indicators.

[0047] The risk management integrated module includes:

[0048] Real-time risk assessment submodule, used to predict and assess the potential risks of financial operations in real time;

[0049] The Compliance and Security submodule is used to make the operation comply with all financial regulations and security standards.

[0050] A method for optimizing the intelligent funds routing system based on the AI ​​big model comprises the following steps:

[0051] (1) Automatically collect financial market data, and perform preprocessing and feature extraction;

[0052] (2) Based on the AI ​​model architecture, it is used to analyze the processed data and make fund routing decisions;

[0053] (3) Optimize the capital routing path based on market and user behavior prediction results;

[0054] (4) Monitor transactions and capital flows in real time and assess potential risks;

[0055] (5) Provide early warning and decision support based on risk assessment results.

[0056] The advantages of the present invention are:

[0057] 1. Efficiency improvement: The use of AI technology has significantly improved the real-time response capability of the funds routing system.

[0058] 2. Security enhancement: The built-in risk assessment mechanism can reduce potential fund allocation risks.

[0059] 3. Improved accuracy: Deep learning models can better predict market trends and user behavior.

[0060] 4. Improved user experience: Provide smarter and more personalized fund management services.

[0061] 5. Reduce costs: Avoid misallocation of funds due to wrong decisions.

[0062] Overall, the present invention will greatly improve the efficiency of existing funds routing systems, meet the challenges of the financial industry, and bring significant economic and social benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 It is a system block diagram of the present invention.

[0064] Figure 2 It is a processing flow diagram of the data processing module of the present invention.

[0065] Figure 3 It is a processing flow diagram of the decision-making module of the present invention.

[0066] Figure 4 It is a schematic diagram of the processing flow of the fund routing optimization module of the present invention.

[0067] Figure 5 It is a schematic diagram of the processing flow of the risk assessment module of the present invention.

[0068] Figure 6 It is a schematic diagram of the processing flow of the risk management integrated module of the present invention. DETAILED DESCRIPTION

[0069] The present invention will be further described below in conjunction with specific embodiments, and the advantages and features of the present invention will become clearer as the description proceeds. However, these embodiments are exemplary only and do not constitute any limitation to the scope of the present invention. It should be understood by those skilled in the art that the details and forms of the technical solution of the present invention may be modified or replaced without departing from the spirit and scope of the present invention, but these modifications and replacements all fall within the scope of protection of the present invention.

[0070] See also Figures 1 to 6 The present invention relates to an intelligent funds routing system based on an AI big model, comprising:

[0071] Data processing module 1, used to automatically collect financial market data, and perform preprocessing and feature extraction;

[0072] Decision-making module 2, based on the AI ​​model architecture, is used to analyze the processed data and make fund routing decisions;

[0073] Fund routing optimization module 3, used to optimize the fund routing path according to the market and user behavior prediction results;

[0074] The risk assessment module 4 is used to monitor transactions and capital flows in real time and assess potential risks; the risk management integration module 5 is used to provide early warning and decision support based on the risk assessment results.

[0075] The intelligent funds routing system based on the AI ​​big model, whose structure includes data processing module, decision-making module, funds routing optimization module, risk assessment module and risk management integration module, has the following advantages:

[0076] 1. Efficient data processing capabilities:

[0077] Automate the collection and processing of financial market data and improve data processing efficiency.

[0078] Preprocessing and feature extraction capabilities enable the system to better understand the raw data and provide more accurate input for subsequent analysis.

[0079] 2. Advanced decision support:

[0080] Using AI model architecture, the system is able to learn patterns and trends from large amounts of historical data and improve the accuracy of decision-making.

[0081] The decision-making module can analyze market data in real time, respond quickly, and improve the timeliness and adaptability of capital routing.

[0082] 3. Optimized funds routing:

[0083] Based on market and user behavior prediction results, the system can dynamically adjust the funds routing path to maximize cost-effectiveness.

[0084] The optimization module is able to take into account multiple factors such as cost, time efficiency and risk to find the best routing solution.

[0085] 4. Real-time risk monitoring:

[0086] The risk assessment module can monitor transactions and capital flows in real time and detect potential risks in a timely manner.

[0087] Through real-time analysis, the system can respond quickly to market changes and reduce potential financial losses.

[0088] 5. Strengthened risk management:

[0089] The risk management integration module issues early warnings based on risk assessment results, helping managers take timely measures.

[0090] Integrated risk management provides decision support and improves the security and compliance of funds routing.

[0091] 6. Improve transparency and explainability:

[0092] The decision logic of the AI ​​model has higher transparency and explainability, enhancing the trust in the system.

[0093] Users and regulators can better understand the decision-making process, increasing the credibility of the system.

[0094] 7. Enhance user experience:

[0095] Provide smarter and more personalized fund management services to improve user satisfaction.

[0096] By accurately predicting and optimizing routing, the system is able to provide faster transaction execution and better service.

[0097] 8. Reduce operating costs:

[0098] Through automated and intelligent data processing and decision-making, the system reduces manual intervention and lowers operating costs.

[0099] Optimized funds routing and risk management help avoid unnecessary losses and further reduce costs.

[0100] The data processing module includes a data acquisition module and a data preprocessing and feature extraction module. The data acquisition module is used to automatically collect the required market data from different data sources; the data preprocessing and feature extraction module is used to preprocess and extract features from the collected raw data; wherein, the data acquisition module includes:

[0101] The automated API retrieval module is used to automatically extract real-time market data, asset prices, and market news information by integrating financial market APIs; this process can reduce manual intervention and improve the efficiency and real-time nature of data collection.

[0102] Social media monitoring module, which is used to monitor and collect public data related to financial markets on social media platforms, such as posts, comments, etc. Analyze investor sentiment through natural language processing technology to obtain market sentiment indicators;

[0103] The historical database module is used to obtain historical data of the financial market, provide rich historical data support for model training, and help the model learn historical market trends and patterns.

[0104] The data preprocessing and feature extraction module includes:

[0105] The outlier detection and processing module is used to identify outliers in the data and remove or correct them to ensure the quality of the input data and reduce noise interference during the model training process; the digital feature conversion module is used to use natural language processing (NLP) technology to process unstructured text data and convert it into numerical features that the model can process; this may include word frequency statistics, sentiment analysis, topic modeling, etc.

[0106] The time series analysis module is used to apply time series analysis techniques, such as moving average and autoregressive models, to process time-sensitive financial market data in order to extract features that reflect market trends and cyclicalities.

[0107] The decision-making module includes a deep learning module and a dynamic adjustment module. The deep learning module is used to realize prediction. The deep learning module includes:

[0108] Multi-head attention mechanism module, used to enhance the model’s ability to understand the context of market information;

[0109] Convolutional neural network module, used for graph feature extraction; for example, analyzing and processing transaction amount trend charts in financial markets, capable of capturing local features and spatial structure information.

[0110] Long short-term memory network module, used to capture dependencies in time series data; suitable for in-depth analysis of historical data in financial market forecasting, so as to identify and predict market trends;

[0111] In order to adapt to the rapid changes in the financial market, our model needs to have the ability to dynamically adjust. The dynamic adjustment module is used to adapt to market changes. The dynamic adjustment module includes

[0112] Model iteration module, which is used to update the model in real time to adapt to changes in the financial market. This involves continuously monitoring market data, automatically adjusting model parameters, and retraining the model when new data is available, ensuring that the model is always up to date, thereby improving the accuracy of predictions and the ability to respond to market changes;

[0113] The multi-model integration module is used to monitor and evaluate the performance of several models in real time, and select the best model for routing decision by monitoring the performance of multiple models in real time.

[0114] The fund routing optimization module includes an algorithm design module and a user behavior prediction module. The algorithm design module is used to balance several objectives using a multi-objective optimization algorithm:

[0115] Cost minimization: Reduce all resource costs transferred by the funding party.

[0116] Time optimization: Improve the timeliness of funds arrival.

[0117] Risk Assessment: By integrating specialized risk assessment models.

[0118] The user behavior prediction module is used to autonomously learn the user's historical transaction patterns and preferences and provide the user with fund routing suggestions; the specific prediction steps of the user behavior prediction module are:

[0119] 4.1 Analysis of historical transaction data:

[0120] By analyzing the user's historical transaction data, we can understand their transaction behavior, including the user's transaction frequency, transaction amount, selected capital routing path, and transaction time. This step provides basic data for subsequent user profile modeling.

[0121] 4.2 User portrait modeling: Based on historical transaction data, AI modeling technology (such as machine learning or deep learning) is used to build user portraits. User portraits are detailed descriptions of user behavior patterns and preferences, helping the system to deeply understand user needs and habits.

[0122] 4.3 Personalized recommendation: After the user portrait is built, the fund routing path that best suits the needs of different users is recommended based on their characteristics and preferences.

[0123] For example: For time-sensitive users, the system can recommend a quick but potentially slightly more expensive path to ensure funds arrive quickly.

[0124] For cost-sensitive users, the system can recommend the path with the lowest cost, even if this may mean a slightly longer arrival time.

[0125] The risk assessment modules include:

[0126] Data input and preprocessing module, which is responsible for collecting transaction, user and market data, and cleaning and formatting them for subsequent feature extraction and risk assessment;

[0127] Feature extraction module, which is used to extract key features from raw data and use these features as input to the risk assessment model;

[0128] The risk assessment model module is used to assess different types of risks through different sub-models;

[0129] The credit risk assessment module is used to train the credit scoring model based on the user's credit data and adjust it according to the user's economic status;

[0130] Liquidity risk assessment module, which uses historical transaction and market data to predict liquidity risk and set early warning limits;

[0131] Market risk assessment module, used to predict market fluctuations using the time series analysis module, taking into account macroeconomic and policy factors.

[0132] The comprehensive risk assessment module is used to combine the assessment results of each sub-model and conduct an overall risk assessment.

[0133] The specific assessment method of the comprehensive risk assessment module is as follows:

[0134] 5.1 Risk weight allocation: assign weights to different risk factors and use weighted methods to synthesize risk values;

[0135] 5.2 Risk level prediction: convert risk values ​​into risk levels and set thresholds (e.g. low risk, medium risk, high risk);

[0136] 5.3 Real-time feedback: Feedback risk assessment results to the transaction decision-making system in real time and issue an early warning when necessary.

[0137] Among other things, the system will provide real-time risk assessment results and issue early warnings if high risks are detected:

[0138] Warning generation: Automatically generate warnings based on risk levels; provide detailed risk reports and recommendations.

[0139] Recommendations and measures: Provide trading strategy recommendations based on risk assessment results; if necessary, prompt transaction adjustments or transaction verification.

[0140] The specific method for extracting key features by the feature extraction module includes:

[0141] 6.1 Transaction characteristics: analysis of transaction amount, credit scores of both parties, historical transaction behavior, and transaction frequency;

[0142] 6.2 User characteristics: evaluate the user's historical transaction behavior (the user's past transaction patterns and habits), credit score (the user's credit status) and behavior prediction (the predicted flow of funds in the user's profile);

[0143] 6.3 Market characteristics: monitor real-time market price fluctuations (current market volatility), economic indicators (macroeconomic indicators such as stock indexes, interest rates, etc.) and market participant confidence indicators (through market news and other signal analysis).

[0144] The risk management integrated module includes:

[0145] Real-time risk assessment submodule, used to predict and assess the potential risks of financial operations in real time;

[0146] The Compliance and Security submodule is used to make the operation comply with all financial regulations and security standards.

[0147] It also includes model optimization and iteration

[0148] It also includes continuously collecting new data and updating models to improve the accuracy of risk assessment, including model retraining: regularly retraining models with new data; evaluating model performance and making necessary adjustments.

[0149] It also includes online learning: supporting online learning mechanisms and timely updating of evaluation models; dynamically adjusting weights and algorithms to adapt to market changes.

[0150] Based on the above steps, the risk assessment module can comprehensively cover risk factors, monitor the flow of funds in real time, and provide high-precision and reliable risk management support for the intelligent fund routing system.

[0151] The present invention also relates to an optimization method for realizing the intelligent funds routing system based on the AI ​​big model, characterized in that it comprises the following steps:

[0152] (1) Automatically collect financial market data, and perform preprocessing and feature extraction;

[0153] (2) Based on the AI ​​model architecture, it is used to analyze the processed data and make fund routing decisions;

[0154] (3) Optimize the capital routing path based on market and user behavior prediction results;

[0155] (4) Monitor transactions and capital flows in real time and assess potential risks;

[0156] (5) Provide early warning and decision support based on risk assessment results.

[0157] The core problem to be solved by this invention is how to create a technical system that can perform deep learning and prediction based on big data, realize intelligent decision-making, and optimize the flow and routing of funds. The main contents include:

[0158] 1. Data analysis automation: Automatically collect and process massive and diverse financial market data, including transaction data, user behavior data, macroeconomic indicators, etc.

[0159] 2. Deep learning architecture: Build a deep learning model suitable for the financial field that can learn from historical data and predict the flow of funds.

[0160] 3. Real-time prediction and response: Analyze small changes in the market in real time, and respond and adjust quickly to achieve optimal capital allocation.

[0161] 4. User behavior prediction: By analyzing users’ behavior patterns and preferences, predict the trends and intentions of their capital flows.

[0162] 5. Risk control: Integrate risk assessment mechanisms into fund routing strategies to minimize potential losses.

[0163] 6. Algorithm transparency and explainability: Ensure the transparency and explainability of decision logic to increase the credibility of the system and the trust of users.

[0164] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. An intelligent funds routing system based on AI big model, characterized in that: include: Data processing module, used to automatically collect financial market data, and perform preprocessing and feature extraction; The decision-making module, based on the AI ​​model architecture, is used to analyze the processed data and make fund routing decisions; Fund routing optimization module, which is used to optimize the fund routing path based on market and user behavior prediction results; Risk assessment module, used to monitor transactions and capital flows in real time and assess potential risks; The risk management integration module is used to provide early warning and decision support based on risk assessment results.

2. The intelligent funds routing system based on the AI ​​big model according to claim 1 is characterized in that: The data processing module includes a data acquisition module and a data preprocessing and feature extraction module. The data acquisition module is used to automatically collect the required market data from different data sources; the data preprocessing and feature extraction module is used to preprocess and extract features from the collected raw data; wherein, the data acquisition module includes: Automated API retrieval module, which is used to automatically extract real-time market data, asset prices, and market news information by integrating financial market APIs; The social media monitoring module is used to monitor and collect public data related to financial markets on social media platforms, analyze investor sentiment through natural language processing technology, and thus obtain market sentiment indicators; The historical database module is used to obtain historical data of the financial market, provide rich historical data support for model training, and help the model learn historical market trends and patterns. The data preprocessing and feature extraction module includes: The outlier detection and processing module is used to identify outliers in the data and remove or correct them to ensure the quality of input data and reduce noise interference during model training; The digital feature conversion module is used to process unstructured text data and convert it into numerical features that can be processed by the model; The time series analysis module is used to process time-sensitive financial market data to extract features that reflect market trends and periodicity.

3. The intelligent funds routing system based on the AI ​​big model according to claim 1 is characterized in that: The decision-making module includes a deep learning module and a dynamic adjustment module. The deep learning module is used to realize prediction. The deep learning module includes: Multi-head attention mechanism module, used to enhance the model’s ability to understand the context of market information; Convolutional neural network module for graph feature extraction; Long short-term memory network module, used to capture dependencies in time series data; The dynamic adjustment module is used to adapt to market changes. The dynamic adjustment module includes a model iteration module, which is used to update the model in real time to adapt to changes in the financial market; Multi-model integration module for real-time monitoring and evaluation of the performance of several models.

4. The intelligent funds routing system based on the AI ​​big model according to claim 1 is characterized in that: The fund routing optimization module includes an algorithm design module and a user behavior prediction module. The algorithm design module is used to balance several objectives using a multi-objective optimization algorithm; the user behavior prediction module is used to autonomously learn the user's historical transaction patterns and preferences and provide the user with fund routing suggestions; the specific prediction steps of the user behavior prediction module are: 4.1 Analysis of historical transaction data: By analyzing the user's historical transaction data, we can understand their transaction behavior, including the user's transaction frequency, transaction amount, selected capital routing path, and transaction time. This step provides basic data for subsequent user profile modeling. 4.2 User portrait modeling: Based on historical transaction data, use AI modeling technology to build user portraits; 4.3 Personalized recommendation: After the user portrait is built, the fund routing path that best suits the needs of different users is recommended based on their characteristics and preferences.

5. The intelligent funds routing system based on the AI ​​big model according to claim 4 is characterized in that: The risk assessment modules include: Data input and preprocessing module, which is responsible for collecting transaction, user and market data, and cleaning and formatting them for subsequent feature extraction and risk assessment; Feature extraction module, which is used to extract key features from raw data and use these features as input to the risk assessment model; The risk assessment model module is used to assess different types of risks through different sub-models; The credit risk assessment module is used to train the credit scoring model based on the user's credit data and adjust it according to the user's economic status; Liquidity risk assessment module, which uses historical transaction and market data to predict liquidity risk and set early warning limits; Market risk assessment module, used to predict market fluctuations using time series analysis module; The comprehensive risk assessment module is used to combine the assessment results of each sub-model and conduct an overall risk assessment.

6. The intelligent funds routing system based on the AI ​​big model according to claim 5 is characterized in that: The specific assessment method of the comprehensive risk assessment module is as follows: 5.1 Risk weight allocation: assign weights to different risk factors and use weighted methods to synthesize risk values; 5.2 Risk level prediction: convert risk value into risk level and set threshold; 5.3 Real-time feedback: Feedback risk assessment results to the transaction decision-making system in real time and issue an early warning when necessary.

7. The intelligent funds routing system based on the AI ​​big model according to claim 5 is characterized in that: The specific method for extracting key features by the feature extraction module includes: 6.1 Transaction characteristics: analysis of transaction amount, credit scores of both parties, historical transaction behavior, and transaction frequency; 6.2 User characteristics: evaluate the user's historical transaction behavior, credit score and behavior prediction; 6.3 Market characteristics: monitor real-time market price fluctuations, macroeconomic indicators and market participants’ confidence indicators.

8. The intelligent funds routing system based on AI big model according to claim 5 is characterized in that: The risk management integrated module includes: Real-time risk assessment submodule, used to predict and assess the potential risks of financial operations in real time; The Compliance and Security submodule is used to make the operation comply with all financial regulations and security standards.

9. A method for optimizing the intelligent funds routing system based on the AI ​​big model according to any one of claims 1 to 8, characterized in that: The following steps are involved: (1) Automatically collect financial market data, and perform preprocessing and feature extraction; (2) Based on the AI ​​model architecture, it is used to analyze the processed data and make fund routing decisions; (3) Optimize the capital routing path based on market and user behavior prediction results; (4) Monitor transactions and capital flows in real time and assess potential risks; (5) Provide early warning and decision support based on risk assessment results.

Images