Real-time all-weather risk control system based on high-frequency data and news public opinions

Abstract

The invention discloses a real-time all-weather risk control system based on high-frequency data and news public opinions, and the system comprises a quantitative intelligent transaction module, a risk income calculation module, a multi-asset risk control module, and an intelligent evaluation module. The quantitative intelligent transaction module comprises a strategy combination module, a strategy matching fund module, an automatic transaction module and a strategy compliance real-time review sub-module, the sub-module obtains financial supervision compliance requirements in real time, carries out compliance verification on financial product combination and fund putting schemes, blocks execution of uncompliance rules, and feeds back the uncompliance rules; the risk income calculation module comprises a core data collection module and a data traceability evidence storage sub-module, the sub-module marks a source identifier, collection time and personnel information for core data to form non-tampering traceability chain associated storage, and the multi-asset risk control module comprises a risk control module and a risk slow release scheme automatic generation sub-module. The sub-module receives the risk signal and analyzes the risk type and range.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of risk control, in particular to a real-time all-weather risk control system based on high-frequency data and news public opinion. BACKGROUND

[0002] With the rapid development and complexification of the financial market, high-frequency trading data and market dynamic news and other information have an increasingly significant impact on financial risks, and traditional risk control methods relying on static rules or low-frequency data have been difficult to meet the real-time and comprehensive needs of financial businesses for risk identification, compliance management and control, and decision support.

[0003] The prior art has obvious method and structure defects in the construction of the financial risk control system: in the compliance management and data evidence link, the traditional risk control scheme mostly adopts a post-compliance review or a mode of checking against a single business link, which fails to synchronize the latest compliance requirements issued by the financial regulatory department in real time, and when a non-compliant financial product combination or fund investment scheme is detected, there is no execution blocking mechanism that can be triggered immediately, and it is also impossible to clearly and timely feedback specific non-compliance items; in the risk response and asset evaluation link, the traditional multi-asset risk control relies on manual analysis of risk types and ranges to develop risk mitigation schemes, which not only has low processing efficiency, but also is prone to insufficient adaptation of the scheme to the actual risk situation due to differences in the experience of the staff; these defects directly lead to two core problems: first, the existing risk control technology cannot guarantee the real-time and completeness of compliance review, and the comprehensiveness, security and convenience of data tracing are also difficult to meet the needs, making the risk control process lack reliable real-time compliance protection and data support; second, the existing risk control technology is inefficient and lacks precision in generating risk mitigation schemes, and the comprehensiveness, consistency and accuracy of asset evaluation are lacking, which cannot provide efficient and reliable basis for the risk control decisions of financial institutions, and it is difficult to meet the business needs of real-time all-weather risk control of the financial market. Therefore, a real-time all-weather risk control system based on high-frequency data and news public opinion is proposed. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a real-time all-weather risk control system based on high-frequency data and news public opinion to solve the problem that the above-mentioned risk control technology cannot guarantee the real-time and completeness of compliance review, and the comprehensiveness, security and convenience of data tracing are also difficult to meet the needs.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical scheme: a real-time all-weather risk control system based on high-frequency data and news public opinion, comprising: a quantitative intelligent transaction module: The strategy combination module, the strategy matching fund module, the automatic transaction module and the strategy compliance real-time review submodule are comprised; the strategy compliance real-time review submodule obtains the latest compliance requirements of financial supervision in real time, and performs compliance verification on the financial product combination of the strategy combination module and the fund investment scheme of the strategy matching fund module, and blocks execution and feeds back non-compliance items if not compliant; The risk and return calculation module: The core data collection module, the order matching module, the closing position module, the closing position matching module, the position record module, the net value curve module and the data traceability storage submodule are comprised; the data traceability storage submodule marks the source identification, collection time and personnel information of the core data, forms an unalterable traceability chain and is associatedly stored, so as to facilitate the traceability of data authenticity and the collection process; The multi-asset risk control module: The risk control module, the summary report module, the performance attribution module and the risk mitigation scheme automatic generation submodule are comprised; the risk mitigation scheme automatic generation submodule receives the risk signal of the risk control module, analyzes the risk type and range, and generates an adaptive mitigation measure in combination with the historical case library; The intelligent evaluation module: The per-transaction evaluation module, the position evaluation module, the net value evaluation module, the machine learning module, the public opinion and asset correlation evaluation calibration submodule and the cross-cycle evaluation consistency verification submodule are comprised; the public opinion and asset correlation evaluation calibration submodule extracts asset-related evaluation keywords in public opinion, matches existing evaluation indicators, supplements un-covered dimensions and adjusts the weight of the indicators; the cross-cycle evaluation consistency verification submodule extracts different cycle evaluation results, analyzes the evaluation difference of the same asset, and triggers the machine learning module to correct the evaluation bias if the difference is out of range.

[0006] The core data collection module in the risk and return calculation module starts to collect core data; the data traceability storage submodule in the risk and return calculation module operates synchronously, marks the source identification, collection time and personnel information of the collected core data, forms an unalterable traceability chain, and stores the traceability chain and the core data in association, providing a basis for subsequent data authenticity traceability and collection process traceability; The strategy combination module in the quantitative intelligent transaction module constructs a financial product combination; the strategy matching fund module in the quantitative intelligent transaction module formulates a fund investment scheme for the above financial product combination; the strategy compliance real-time review submodule in the quantitative intelligent transaction module obtains the latest compliance requirements of financial supervision in real time, and performs compliance verification on the financial product combination constructed by the strategy combination module and the fund investment scheme formulated by the strategy matching fund module; if the compliance verification result is not compliant, the strategy compliance real-time review submodule blocks the execution of the financial product combination and the implementation of the fund investment scheme, and feeds back specific non-compliance items to the relevant operation end; if the compliance verification result is compliant, the next process is entered; The order matching module in the risk and return calculation module carries out order matching operation according to the financial product portfolio and the fund investment plan that pass the compliance verification; the position record module in the risk and return calculation module records the asset position after order matching in real time; the net value curve module in the risk and return calculation module generates the net asset value curve according to the position record of the position record module and the core data collected by the core data collection module, and intuitively presents the change state of the net asset value; The risk control module in the multi-asset risk control module monitors the risk in the process of order matching and position management in real time, and captures the risk signal; if the risk control module monitors the risk signal, the risk signal is transmitted to the risk mitigation scheme automatic generation sub-module in the multi-asset risk control module; after receiving the risk signal, the risk mitigation scheme automatic generation sub-module analyzes the risk type and risk coverage range, and generates the risk mitigation measures suitable for the current risk by combining the historical case library built in the system; the summary report module in the multi-asset risk control module generates the risk control summary report according to the risk monitoring result of the risk control module, the risk mitigation measures of the risk mitigation scheme automatic generation sub-module, the order matching and position record and other information; the performance attribution module in the multi-asset risk control module performs attribution analysis on the performance of assets during order matching and position holding, and clarifies the performance influencing factors; The transaction-by-transaction evaluation module in the intelligent evaluation and assessment module evaluates the rationality and profitability of each order matching transaction; the position evaluation module in the intelligent evaluation and assessment module evaluates the rationality and risk level of the current asset position; the net value evaluation module in the intelligent evaluation and assessment module evaluates the stability and growth of the net asset value according to the net asset value curve generated by the net value curve module; the public opinion and asset correlation evaluation calibration sub-module in the intelligent evaluation and assessment module extracts the evaluation keywords related to the current asset in the news public opinion, matches the evaluation keywords with the existing evaluation indexes, supplements the evaluation dimensions not covered by the existing evaluation indexes, and adjusts the weight of each evaluation index according to the importance of the evaluation keywords; the cross-cycle evaluation consistency verification sub-module extracts the results of transaction-by-transaction evaluation, position evaluation and net value evaluation in different time periods, such as short-term, medium-term and long-term, analyzes the evaluation difference of the same asset in different periods; if the evaluation difference exceeds the preset reasonable range, the cross-cycle evaluation consistency verification sub-module triggers the machine learning module in the intelligent evaluation and assessment module to correct the evaluation deviation and ensure the accuracy of the evaluation result.

[0007] Preferably, the strategy combination module also has a compliance requirement adaptation update function, which receives the non-compliance items fed back by the real-time compliance review submodule, analyzes the corresponding compliance requirement type, automatically adjusts the financial product combination screening conditions, and excludes financial products that do not meet the latest compliance requirements, to ensure that the combination scheme complies with regulatory rules.

[0008] Ensure that the combination scheme output by the strategy combination module always complies with regulatory rules, and avoid the risk of the combination scheme being unable to execute or violating rules due to compliance issues.

[0009] Preferably, the automatic transaction module also has a compliance blocking post-scheme recommendation function, which receives the updated compliance combination scheme of the strategy combination module after the strategy compliance real-time review submodule blocks the non-compliant scheme execution, and generates different priority automatic transaction execution schemes based on the user's current financial situation and transaction preferences, for the user to select and start the transaction operation.

[0010] After the non-compliant scheme is blocked, a compliant and user-specific transaction scheme selection can be quickly provided, avoiding the stagnation of the transaction process due to compliance blocking, and improving the continuity of transactions and user experience.

[0011] Preferably, the core data collection module also has a traceability information synchronization function, which synchronizes the source identifier, collection time, and personnel information generated by the data traceability storage submodule to the core data storage field when collecting data, ensuring that the traceability information can be directly called without additional traceability chain queries.

[0012] Synchronize data and traceability information storage, so that traceability information can be directly called without additional traceability chain queries during subsequent data calling, reducing the number of data traceability queries and time costs, and improving data usage efficiency.

[0013] Preferably, the closing module also has a traceability information associated closing decision function, which retrieves the historical data traceability information of the corresponding asset when the user initiates a closing request, and if data traceability abnormalities are found, the closing operation is suspended to verify the data abnormality reason before executing the closing process.

[0014] By associating traceability information with closing decisions, the risk of incorrect closing operations due to data traceability abnormalities is avoided, ensuring that closing decisions are based on accurate and reliable data, and reducing the operational and data risks in the closing process.

[0015] Preferably, the risk control module also has a risk signal grading function, which grades the monitored risk signals according to the impact range, occurrence probability and loss degree, synchronizes the grading results to the risk mitigation scheme automatic generation sub-module, and makes the risk mitigation scheme automatic generation sub-module generate mitigation measures of different levels of detail according to the risk level. High-level risks correspond to complete schemes containing emergency disposal steps, and low-level risks correspond to simplified optimized suggestions.

[0016] The difference between the risk signals is processed, the risk mitigation measures are accurately matched with the risk level, the high-level risks are fully and timely disposed of, the mitigation measures for low-level risks are not too complicated to waste resources, and the efficiency and pertinence of risk control are improved.

[0017] Preferably, the summary report module also has a risk mitigation effect tracking function, which records the execution process and results of the mitigation measures output by the risk mitigation scheme automatic generation sub-module, compares the execution results with the risk mitigation target, analyzes the actual effect, and if the effect is not as expected, the analysis results are fed back to the risk control module to trigger a new round of risk signal monitoring and scheme optimization.

[0018] The whole process of tracking and evaluating the risk mitigation effect is realized, the risk mitigation situation that does not meet the expectation is found in time and the optimization mechanism is triggered, forming a closed loop of risk control, and continuously improving the effectiveness of the risk mitigation scheme and the level of risk control.

[0019] Preferably, the public opinion and asset correlation evaluation calibration sub-module also has a keyword priority sorting function, which sorts the extracted public opinion keywords according to the correlation degree with the asset and the public opinion propagation range, and preferentially supplements the evaluation dimensions corresponding to the high-priority keywords to the transaction-by-transaction evaluation module and the holding evaluation module. Low-priority keywords are used as alternatives and are gradually supplemented after the high-priority dimensions are stable.

[0020] Ensure that the relevant evaluation module preferentially obtains the public opinion evaluation dimensions with high correlation degree and wide influence range, guarantee the core effectiveness of the evaluation results, and at the same time, through the phased supplementation of the evaluation dimensions, avoid the evaluation module running in disorder due to too many and miscellaneous dimensions, and improve the stability and evaluation accuracy of the evaluation module.

[0021] Preferably, the cross-period evaluation consistency verification sub-module also has a difference reason analysis function, which, when detecting that the difference degree of the evaluation results of different periods is out of range, retrieves the high-frequency data changes, news public opinion content and transaction operation records in the corresponding period, analyzes the specific reasons for the difference, classifies and labels the reasons and feeds them back to the machine learning module.

[0022] The root of the difference in the evaluation results of different periods can be accurately positioned, the machine learning module is provided with clear optimization basis, the blindness of evaluation logic optimization caused by the inability to determine the difference reason is avoided, and the consistency and reliability of cross-period evaluation are improved.

[0023] Preferably, the machine learning module also has an evaluation logic iterative optimization function, which receives the difference reason classification result fed back by the cross-period evaluation consistency checking submodule, adjusts the input parameters of the analysis model for different reason types, and continuously iterates to make the evaluation logic more suitable for actual application scenarios.

[0024] The machine learning module can continuously optimize the evaluation logic based on the actual evaluation difference reason, make the evaluation logic close to the actual application demand, improve the applicability of the evaluation logic and the accuracy of the evaluation result, and provide more reliable support for the decision of the subsequent related modules.

[0025] In summary, compared with the prior art, the present application provides a real-time all-weather risk control system based on high-frequency data and news public opinion, which has the following beneficial effects: In the present application, the strategy compliance real-time review submodule obtains the latest compliance requirements of financial supervision in real time, and simultaneously performs compliance verification on the financial product combination of the strategy combination module and the fund investment scheme of the strategy matching fund module, and simultaneously implements execution blocking and non-compliance item feedback when it is not compliant, breaking the limitations of lagging compliance review, only checking a single link or lacking timely intervention and clear feedback after violation in traditional risk control, and improving the real-time and integrity of compliance risk control; The data traceability storage submodule also labels the source identification, collection time and personnel information of the core data, and constructs an unalterable traceability chain and stores it in association with the core data, compared with the traditional data storage method which only labels part of the information, lacks the unalterable feature and association management, greatly improves the comprehensiveness, security and convenience of data traceability, and ensures the data traceability in the whole process. The risk mitigation scheme automatic generation submodule automatically analyzes the risk type and range after receiving the risk signal, and generates the appropriate risk mitigation measures combined with the historical case library, changes the status of traditional multi-asset risk control relying on manual preparation of mitigation schemes, low efficiency and experience limitation, and realizes the automatic and accurate generation of risk mitigation schemes. The opinion and asset correlation evaluation calibration submodule extracts asset-related evaluation keywords in the opinion, matches existing evaluation indexes, supplements evaluation dimensions not covered, adjusts index weights, and verifies consistency across cycles. The difference in evaluation of the same asset is analyzed, and when the difference exceeds the range, the machine learning module is triggered to correct the evaluation bias, solving the problems of traditional asset evaluation ignoring the influence of public opinion, lacking cross-cycle verification and automatic bias correction, significantly improving the comprehensiveness, consistency and accuracy of asset evaluation, and providing a more reliable basis for risk control decisions. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 is a schematic diagram of the system of the present application. DETAILED DESCRIPTION

[0027] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0028] Please refer to Figure 1 The present application provides a technical solution, a real-time all-weather risk control system based on high-frequency data and news public opinion, comprising: Quantitative intelligent transaction module: Including strategy combination module, strategy matching fund module, automatic transaction module and strategy compliance real-time review submodule; the strategy compliance real-time review submodule obtains the latest compliance requirements of financial supervision in real time, and performs compliance verification on the financial product combination of the strategy combination module and the fund investment scheme of the strategy matching fund module, and blocks the execution and feeds back the non-compliance items if it is not compliant; Risk and return calculation module: Including core data collection module, order matching module, closing position module, closing position matching module, position record module, net value curve module and data traceability storage submodule; the data traceability storage submodule marks the source identification, collection time and personnel information of the core data, forms an unalterable traceability chain and is associated with storage, so as to facilitate the traceability of data authenticity and collection process; Multi-asset risk control module: Including risk control module, summary report module, performance attribution module and risk mitigation scheme automatic generation submodule; the risk mitigation scheme automatic generation submodule receives the risk signal of the risk control module, analyzes the risk type and range, and generates adaptive mitigation measures combined with the historical case library; Intelligent evaluation module: The evaluation system comprises a transaction-by-transaction evaluation module, a position evaluation module, a net value evaluation module, a machine learning module, an opinion and asset correlation evaluation calibration submodule, and a cross-cycle evaluation consistency verification submodule. The opinion and asset correlation evaluation calibration submodule extracts asset-related evaluation keywords in public opinions, matches existing evaluation indicators, supplements un-covered dimensions, and adjusts the weights of the indicators. The cross-cycle evaluation consistency verification submodule extracts evaluation results of different cycles, analyzes the difference in the evaluation of the same asset, and triggers the machine learning module to correct the evaluation bias if the difference is out of range.

[0029] The core data collection module in the risk and return calculation module is started to collect core data. The data traceability and storage submodule in the risk and return calculation module is operated synchronously to mark the source identification, collection time, and personnel information of the collected core data, form an unalterable traceability chain, and store the traceability chain and the core data in association, providing a basis for subsequent data authenticity tracing and collection process tracing. The strategy combination module in the quantitative intelligent transaction module constructs a financial product combination. The strategy matching fund module in the quantitative intelligent transaction module formulates a fund investment plan for the above-mentioned financial product combination. The strategy compliance real-time review submodule in the quantitative intelligent transaction module obtains the latest compliance requirements of financial supervision in real time, and performs compliance verification on the financial product combination constructed by the strategy combination module and the fund investment plan formulated by the strategy matching fund module, respectively. If the compliance verification result is non-compliant, the strategy compliance real-time review submodule blocks the execution of the financial product combination and the implementation of the fund investment plan, and feeds back the specific non-compliant items to the relevant operation end. If the compliance verification result is compliant, the next process is entered. The risk and return calculation module comprises a core data collection module, a data traceability and storage submodule, a strategy combination module, a strategy matching fund module, and a strategy compliance real-time review submodule. The core data collection module in the risk and return calculation module is started to collect core data. The data traceability and storage submodule in the risk and return calculation module is operated synchronously to mark the source identification, collection time, and personnel information of the collected core data, form an unalterable traceability chain, and store the traceability chain and the core data in association, providing a basis for subsequent data authenticity tracing and collection process tracing. The risk control module in the multi-asset risk control module monitors the risk situation in the process of order matching and position management in real time, and captures risk signals; if the risk control module monitors the risk signal, the risk signal is transmitted to the risk mitigation scheme automatic generation sub-module in the multi-asset risk control module; after the risk mitigation scheme automatic generation sub-module receives the risk signal, the risk type and risk coverage range are analyzed, and combined with the built-in historical case library of the system, the risk mitigation measures suitable for the current risk are generated; the summary report module in the multi-asset risk control module generates a risk control summary report according to the risk monitoring results of the risk control module, the risk mitigation measures of the risk mitigation scheme automatic generation sub-module, the order matching and position record information, etc.; the performance attribution module in the multi-asset risk control module performs attribution analysis on the performance of assets during order matching and position, and determines the performance influencing factors; The transaction-by-transaction evaluation module in the intelligent evaluation and assessment module evaluates the rationality and profitability of each order matching transaction; the position evaluation module in the intelligent evaluation and assessment module evaluates the rationality and risk level of the current asset position; the net value evaluation module in the intelligent evaluation and assessment module evaluates the stability and growth of asset net value according to the asset net value curve generated by the net value curve module; the public opinion and asset correlation evaluation calibration sub-module in the intelligent evaluation and assessment module extracts evaluation keywords related to the current asset from news public opinion, matches the evaluation keywords with existing evaluation indicators, supplements the evaluation dimensions not covered by existing evaluation indicators, and adjusts the weights of each evaluation indicator according to the importance of the evaluation keywords; the cross-cycle evaluation consistency verification sub-module extracts the results of transaction-by-transaction evaluation, position evaluation, and net value evaluation in different time periods, such as short-term, medium-term, and long-term, and analyzes the evaluation difference of the same asset in different cycles; if the evaluation difference exceeds the preset reasonable range, the cross-cycle evaluation consistency verification sub-module triggers the machine learning module in the intelligent evaluation and assessment module to correct the evaluation bias and ensure the accuracy of the evaluation results; The strategy compliance real-time review sub-module in the quantitative intelligent transaction module can obtain the latest compliance requirements of financial supervision in real time, perform pre-compliance verification on the financial product portfolio and fund investment scheme, block non-compliant operations before transaction execution and feedback non-compliant items, avoid the system from facing regulatory penalties due to illegal transactions, and ensure that the system transaction behavior always complies with regulatory specifications; The data traceability and storage sub-module in the risk and return calculation module marks the source, time, and personnel of the core data and forms an unalterable traceability chain, which can clearly trace the data collection source and process, effectively prevent data falsification or tampering, and ensure the authenticity and reliability of the core data. At the same time, when there is a dispute over data, the data traceability can be quickly completed, improving the problem troubleshooting efficiency; The risk control module in the multi-asset risk control module can monitor risks in real time, and the risk mitigation scheme automatic generation sub-module can quickly generate adaptive risk mitigation measures in combination with the historical case library, without relying on manual long-time analysis and decision-making, so as to timely contain risk spread and reduce economic losses caused by risks, and significantly improve the dynamic management and control capability of the system on multi-asset risks. The intelligent evaluation and assessment module can avoid one-sidedness of evaluation caused by a single evaluation perspective by supplementing evaluation dimensions and adjusting index weights through the public opinion and asset correlation evaluation calibration sub-module; the cross-cycle evaluation consistency verification sub-module can correct evaluation bias in combination with the machine learning module, so as to guarantee consistency of evaluation results of the same asset in different cycles and improve accuracy of the evaluation results, and provide reliable basis for strategy optimization and asset allocation adjustment and other decisions. The system forms a complete risk control closed loop from core data collection, strategy compliance verification, to transaction execution, risk monitoring, and multi-dimensional evaluation, and each module supports real-time operation, so that uninterrupted risk management and control can be realized all day long without manual intervention, meeting the real-time and continuity requirements of the financial field on risk control.

[0030] The strategy combination module also has a compliance requirement adaptive updating function, which receives non-compliant items fed back by the strategy compliance real-time review sub-module, analyzes the corresponding compliance requirement type, automatically adjusts the financial product combination screening conditions, and excludes financial products that do not meet the latest compliance requirements, so as to ensure that the combination scheme complies with regulatory rules.

[0031] The compliance requirement adaptive updating function receives non-compliant items fed back by the strategy compliance real-time review sub-module; analyzes the received non-compliant items to determine the corresponding compliance requirement type; automatically adjusts the financial product combination screening conditions according to the compliance requirement type obtained through the analysis; and excludes financial products that do not meet the latest compliance requirements according to the adjusted screening conditions. This ensures that the combination scheme output by the strategy combination module always complies with regulatory rules, avoiding the risk of non-execution or violation of rules due to compliance problems.

[0032] The automatic transaction module also has a compliance blocking post-scheme recommendation function, which receives the updated compliance combination scheme of the strategy combination module after the strategy compliance real-time review sub-module blocks the execution of the non-compliant scheme, generates different priority automatic transaction execution schemes in combination with the current fund situation and transaction preferences of the user, and provides the user with the option to start the transaction operation.

[0033] After the strategy compliance real-time review submodule blocks the non-compliant scheme execution, the compliance blocking scheme recommendation function is started; the updated compliance portfolio scheme of the strategy combination module is received; the compliance portfolio scheme is processed in combination with the current fund situation and transaction preferences of the user; different priority automatic transaction execution schemes are generated; the generated automatic transaction execution schemes are provided to the user, and the transaction operation is started after the user selects; After the non-compliant scheme is blocked, a compliant and user-specific transaction scheme selection can be quickly provided, avoiding transaction process stagnation due to compliance blocking and improving transaction continuity and user experience.

[0034] The core data collection module also has a traceability information synchronization function. When collecting data, the traceability information synchronization function synchronizes the source identifier, collection time, and personnel information generated by the data traceability storage submodule to the core data storage field, ensuring that the traceability information can be directly called when the data is called without additional traceability chain queries.

[0035] The traceability information synchronization function is started synchronously during the data collection process of the core data collection module; the source identifier, collection time, and personnel information generated by the data traceability storage submodule are obtained; the obtained source identifier, collection time, and personnel information are stored in the core data storage field; the association storage of data and traceability information is completed, ensuring that the traceability information can be directly obtained when the data is called; The synchronous storage of data and traceability information is realized, and in the subsequent data calling process, the traceability information can be directly called without additional traceability chain queries, reducing the number of data traceability queries and time cost, and improving data usage efficiency.

[0036] The liquidation module also has a traceability information associated liquidation decision function. When the user initiates a liquidation request, the traceability information associated liquidation decision function calls the historical data traceability information of the assets corresponding to the liquidation. If data traceability abnormalities are found, the liquidation operation is suspended, and the data abnormality reason is verified first. After confirming that the data is correct, the liquidation process is executed.

[0037] When the user initiates a liquidation request, the traceability information associated liquidation decision function is triggered; the historical data traceability information of the assets corresponding to the liquidation operation is called; the called historical data traceability information is verified to determine whether there are traceability abnormalities; if data traceability abnormalities are found, the liquidation operation is suspended, and the data abnormality reason is verified first; after the data abnormality reason is verified and it is confirmed that the data is correct, the liquidation process is executed; By associating traceability information with liquidation decisions, incorrect liquidation operations caused by data traceability abnormalities are avoided, ensuring that liquidation decisions are based on accurate and reliable data, and reducing the operational and data risks in the liquidation process.

[0038] The risk control module also has a risk signal grading function. The risk signal grading function grades the monitored risk signals according to the influence range, occurrence probability and loss degree, synchronizes the grading results to the risk mitigation scheme automatic generation sub-module, and enables the risk mitigation scheme automatic generation sub-module to generate mitigation measures of different levels of detail according to the risk level. High-level risks correspond to complete schemes containing emergency disposal steps, and low-level risks correspond to simplified optimization suggestions.

[0039] The risk signal grading function processes the risk signals monitored by the risk control module; grades the risk signals according to the influence range, occurrence probability and loss degree; synchronizes the grading results to the risk mitigation scheme automatic generation sub-module; and enables the risk mitigation scheme automatic generation sub-module to generate mitigation measures of different levels of detail according to the risk level, wherein high-level risks correspond to complete schemes containing emergency disposal steps, and low-level risks correspond to simplified optimization suggestions; The risk signal grading function processes the risk signals monitored by the risk control module; grades the risk signals according to the influence range, occurrence probability and loss degree; synchronizes the grading results to the risk mitigation scheme automatic generation sub-module; and enables the risk mitigation scheme automatic generation sub-module to generate mitigation measures of different levels of detail according to the risk level, wherein high-level risks correspond to complete schemes containing emergency disposal steps, and low-level risks correspond to simplified optimization suggestions;

[0040] The summary report module also has a risk mitigation effect tracking function. The risk mitigation effect tracking function records the execution process and results of the mitigation measures output by the risk mitigation scheme automatic generation sub-module, compares the execution results with the risk mitigation target, analyzes the actual effect, and if the effect is not as expected, feeds back the analysis results to the risk control module to trigger a new round of risk signal monitoring and scheme optimization.

[0041] The risk mitigation effect tracking function records the execution process and results of the mitigation measures output by the risk mitigation scheme automatic generation sub-module; compares the execution results of the mitigation measures with the preset risk mitigation target; analyzes the actual effect of the risk mitigation according to the comparison results; if the analysis finds that the actual effect of the risk mitigation does not meet the expectation, feeds back the analysis results to the risk control module; and triggers the risk control module to start a new round of risk signal monitoring and risk mitigation scheme optimization process; The risk mitigation effect tracking function records the execution process and results of the mitigation measures output by the risk mitigation scheme automatic generation sub-module; compares the execution results of the mitigation measures with the preset risk mitigation target; analyzes the actual effect of the risk mitigation according to the comparison results; if the analysis finds that the actual effect of the risk mitigation does not meet the expectation, feeds back the analysis results to the risk control module; and triggers the risk control module to start a new round of risk signal monitoring and risk mitigation scheme optimization process;

[0042] The public opinion and asset correlation evaluation calibration sub-module also has a keyword priority sorting function. The keyword priority sorting function sorts the extracted public opinion keywords according to the degree of correlation with the asset and the range of public opinion dissemination, and preferentially supplements the evaluation dimensions corresponding to high-priority keywords to the per-transaction evaluation module and the holding evaluation module, and the evaluation dimensions corresponding to low-priority keywords are used as alternatives and are gradually supplemented after the high-priority dimensions are stable.

[0043] The keyword priority sorting function processes the public opinion keywords extracted by the public opinion and asset correlation evaluation and calibration submodule; it prioritizes the public opinion keywords according to the closeness of their correlation with assets and the scope of their dissemination; it prioritizes the evaluation dimensions corresponding to high-priority keywords to the transaction evaluation module and the position evaluation module; and it sets the evaluation dimensions corresponding to low-priority keywords as alternatives, which will be gradually added to the relevant evaluation modules after the high-priority dimensions have been running stably. Ensure that relevant evaluation modules prioritize public opinion evaluation dimensions that are highly relevant to assets and have a wide impact, thereby guaranteeing the core validity of the evaluation results. At the same time, by supplementing evaluation dimensions in stages, avoid the evaluation module from operating chaotically due to too many or too complex dimensions, and improve the stability and accuracy of the evaluation module.

[0044] The cross-cycle evaluation consistency verification submodule also has a difference cause analysis function. When the difference between evaluation results of different cycles exceeds the range, the difference cause analysis function retrieves the high-frequency data changes, news and public opinion content and transaction operation records in the corresponding cycle, analyzes the specific reasons for the difference, classifies and labels the reasons and feeds them back to the machine learning module.

[0045] When the cross-cycle evaluation consistency verification submodule detects that the difference between evaluation results of different cycles exceeds the set range, the difference cause analysis function is activated; it retrieves the high-frequency data changes, news and public opinion content, and transaction operation records within the corresponding cycle in which the difference occurred; it performs a comprehensive analysis on the retrieved high-frequency data changes, news and public opinion content, and transaction operation records to determine the specific reasons for the difference in evaluation results; it classifies and labels the reasons for the difference obtained from the analysis, and feeds the classified and labeled reasons for the difference back to the machine learning module; It can accurately pinpoint the root causes of differences in evaluation results across different periods, providing clear optimization basis for the machine learning module, avoiding blind optimization of evaluation logic due to the inability to determine the cause of the differences, and improving the consistency and reliability of cross-period evaluation.

[0046] The machine learning module also has an evaluation logic iterative optimization function. This function receives the difference cause classification results from the cross-cycle evaluation consistency verification submodule, adjusts the input parameters of the analysis model for different cause types, and makes the evaluation logic more suitable for actual application scenarios through continuous iteration.

[0047] The evaluation logic iterative optimization function receives the classification results of the differences from the cross-cycle evaluation consistency verification submodule; for different types of differences, the input parameters of the analysis model are adjusted accordingly; through continuous parameter adjustment and effect verification, the evaluation logic is iteratively updated, making the iterative evaluation logic more suitable for actual application scenarios. This allows the machine learning module to continuously optimize the evaluation logic based on the actual reasons for evaluation differences, making the evaluation logic closer to the needs of actual applications, improving the applicability of the evaluation logic and the accuracy of the evaluation results, and providing more reliable support for the decision-making of subsequent related modules.

[0048] This solution involves the following steps: The core data collection module within the risk-return calculation module is activated to collect core data. Simultaneously, the data traceability and evidence preservation submodule within the risk-return calculation module operates concurrently, labeling the collected core data with source identifiers, collection times, and personnel information to form an immutable traceability chain. This traceability chain is then associated with and stored in relation to the core data. During this process, the traceability information synchronization function within the core data collection module is also activated, acquiring the source identifiers, collection times, and personnel information generated by the data traceability and evidence preservation submodule and storing them synchronously in the core data storage field, thus completing the association between the data and the traceability information. The strategy portfolio module in the quantitative intelligent trading module constructs a portfolio of financial products; the strategy matching fund module in the quantitative intelligent trading module formulates a fund allocation plan for the above-mentioned portfolio of financial products. The real-time compliance review submodule in the quantitative intelligent trading module obtains the latest financial regulatory compliance requirements in real time and performs compliance verification on the financial product portfolios constructed by the strategy portfolio module and the capital allocation plans formulated by the strategy matching fund module. If the compliance verification result is non-compliant, the real-time compliance review submodule blocks the execution of the financial product portfolios and the implementation of the capital allocation plans, and reports the specific non-compliance items to the relevant operating terminals. At this time, the compliance requirement adaptation and update function in the strategy portfolio module receives the non-compliance items reported by the real-time compliance review submodule, analyzes the received non-compliance items, clarifies the corresponding compliance requirement type, and automatically adjusts the financial product portfolio screening conditions according to the analysis of the compliance requirement type. Based on the adjusted screening conditions, financial products that do not meet the latest compliance requirements are eliminated to ensure that the portfolio plan output by the strategy portfolio module complies with regulatory rules. At the same time, the compliance blocking solution recommendation function in the automatic trading module is activated. It receives the updated compliant portfolio plan from the strategy portfolio module, processes the compliant portfolio plan according to the user's current capital status and trading preferences, generates automatic trading execution plans with different priorities, and provides the generated automatic trading execution plans to the user. After the user selects, the trading operation is started. If the compliance verification result is compliant, the process proceeds to the next step. The order matching module in the risk and return calculation module conducts order matching operations based on the financial product portfolio and capital allocation plan that have passed compliance verification; the position record module in the risk and return calculation module records the asset position status in real time after the order matching; the net asset value curve module in the risk and return calculation module generates the asset net asset value curve based on the position record module and the core data collected by the core data collection module; if a user initiates a closing request, the traceability information associated with the closing decision function in the closing module of the risk and return calculation module is triggered, and the historical data traceability information of the asset corresponding to the closing operation is retrieved. The retrieved historical data traceability information is checked to determine whether there is any traceability anomaly. If a traceability anomaly is found, the closing operation is suspended, and the cause of the data anomaly is checked first. After the cause of the data anomaly is checked and the data is confirmed to be correct, the closing process is executed. The risk control module within the multi-asset risk control module monitors the risk situation in real time during order matching and position management, capturing risk signals. The risk signal grading function within the risk control module categorizes monitored risk signals according to their impact range, probability of occurrence, and degree of loss, and synchronizes the grading results to the risk mitigation plan automatic generation submodule within the multi-asset risk control module. If the risk control module detects a risk signal, it transmits the risk signal and grading results to the risk mitigation plan automatic generation submodule. Upon receiving the risk signal and grading results, the risk mitigation plan automatic generation submodule analyzes the risk type and risk coverage, and, combined with the system's built-in historical case library, generates risk mitigation measures of varying levels of detail according to risk level. High-level risks correspond to a complete plan including emergency response steps, while low-level risks correspond to simplified, optimized suggestions. The summary report module in the multi-asset risk control module generates a risk management summary report based on the risk monitoring results of the risk control module, the risk mitigation measures automatically generated by the risk mitigation plan sub-module, and information such as order matching and position records. The risk mitigation effect tracking function in the summary report module records the execution process and results of the mitigation measures output by the risk mitigation plan automatic generation sub-module, compares the execution results of the mitigation measures with the preset risk mitigation targets, analyzes the actual effect of risk mitigation based on the comparison results, and if the analysis finds that the actual effect of risk mitigation has not met expectations, the analysis results are fed back to the risk control module, triggering the risk control module to start a new round of risk signal monitoring and risk mitigation plan optimization process. The performance attribution module in the multi-asset risk control module performs attribution analysis on the performance of assets during order matching and position holding, and identifies the factors affecting performance. The transaction evaluation module in the intelligent assessment and evaluation module evaluates the rationality and profitability of each order matching transaction; the position evaluation module in the intelligent assessment and evaluation module evaluates the rationality and risk level of the current asset position; the net asset value evaluation module in the intelligent assessment and evaluation module evaluates the stability and growth of the net asset value based on the net asset value curve generated by the net asset value curve module. The public opinion and asset correlation evaluation calibration submodule in the intelligent evaluation module extracts evaluation keywords related to the current asset from news and public opinion, matches the evaluation keywords with existing evaluation indicators, supplements evaluation dimensions not covered by existing evaluation indicators, and adjusts the weight of each evaluation indicator according to the importance of the evaluation keywords. The keyword priority sorting function in the public opinion and asset correlation evaluation calibration submodule sorts the extracted public opinion keywords according to the closeness of their correlation with the asset and the scope of public opinion dissemination. It prioritizes the evaluation dimensions corresponding to high-priority keywords to be added to the transaction evaluation module and the position evaluation module, and sets the evaluation dimensions corresponding to low-priority keywords as alternatives. After the high-priority dimensions are running stably, they will be gradually added to the relevant evaluation modules. The cross-period evaluation consistency verification submodule in the intelligent evaluation module extracts the results of transaction-by-transaction evaluation, position evaluation, and net asset value evaluation under different time periods, and analyzes the evaluation difference of the same asset under different periods. If the evaluation difference exceeds the preset reasonable range, the difference cause analysis function in the cross-period evaluation consistency verification submodule is activated. It retrieves high-frequency data changes, news and public opinion content, and transaction operation records in the corresponding period where the difference occurred, comprehensively analyzes the retrieved information, determines the specific reasons for the difference in evaluation results, classifies and labels the analyzed difference causes, and feeds the classified and labeled difference causes back to the machine learning module in the intelligent evaluation module. At the same time, the cross-period evaluation consistency verification submodule triggers the machine learning module, which corrects the evaluation deviation. The evaluation logic iterative optimization function in the machine learning module receives the difference cause classification results fed back by the cross-period evaluation consistency verification submodule, adjusts the input parameters of the analysis model for different types of difference causes, and achieves iterative updates of the evaluation logic through continuous parameter adjustment and effect verification, making the iterative evaluation logic more suitable for actual application scenarios.

[0049] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0050] 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 real-time, 24 / 7 risk control system based on high-frequency data and news sentiment, characterized in that, include: Quantitative Intelligent Trading Module: It includes a strategy portfolio module, a strategy matching fund module, an automated trading module, and a strategy compliance real-time review sub-module. The strategy compliance real-time review sub-module obtains the latest financial regulatory compliance requirements in real time, performs compliance verification on the financial product portfolio of the strategy portfolio module and the fund allocation plan of the strategy matching fund module, and blocks execution and reports non-compliance items if they do not comply with the rules. Risk-return calculation module: It includes a core data collection module, an order matching module, a closing module, a closing matching module, a position record module, a net value curve module, and a data traceability and evidence storage sub-module. The data traceability and evidence storage sub-module marks the source identifier, collection time, and personnel information of the core data, forming an immutable traceability chain and storing it in association, which facilitates the traceability of data authenticity and collection process. Multi-asset risk control module: It includes a risk control module, a summary report module, a performance attribution module, and a risk mitigation plan automatic generation sub-module; the risk mitigation plan automatic generation sub-module receives risk signals from the risk control module, analyzes the risk type and scope, and generates appropriate mitigation measures in combination with the historical case library; Intelligent evaluation module: It includes a transaction-by-transaction evaluation module, a portfolio evaluation module, a net asset value evaluation module, a machine learning module, a public opinion and asset correlation evaluation calibration submodule, and a cross-period evaluation consistency verification submodule. The public opinion and asset correlation evaluation calibration submodule extracts asset-related evaluation keywords from public opinion, matches them with existing evaluation indicators, supplements uncovered dimensions, and adjusts indicator weights. The cross-period evaluation consistency verification submodule extracts evaluation results from different periods, analyzes the evaluation differences of the same asset, and triggers the machine learning module to correct evaluation deviations if the differences exceed the specified range.

2. The real-time, all-weather risk control system based on high-frequency data and news sentiment as described in claim 1, characterized in that: The strategy portfolio module also has a compliance requirement adaptation and update function. This function receives non-compliance items from the strategy compliance real-time review submodule, analyzes the corresponding compliance requirement types, automatically adjusts the financial product portfolio screening conditions, and removes financial products that do not meet the latest compliance requirements to ensure that the portfolio scheme complies with regulatory rules.

3. The real-time, all-weather risk control system based on high-frequency data and news sentiment as described in claim 2, characterized in that: The automated trading module also has a compliance blocking solution recommendation function. After the strategy compliance real-time review submodule blocks the execution of non-compliant solutions, the compliance blocking solution recommendation function receives the updated compliant combination solution from the strategy combination module, and generates automated trading execution solutions with different priorities based on the user's current capital status and trading preferences, allowing the user to select and start the trading operation.

4. The real-time, all-weather risk control system based on high-frequency data and news sentiment as described in claim 1, characterized in that: The core data collection module also has a traceability information synchronization function. When collecting data, the traceability information synchronization function synchronizes the source identifier, collection time and personnel information generated by the data traceability and evidence storage submodule to the core data storage field, ensuring that the traceability information can be directly retrieved when the data is called without the need for additional querying of the traceability chain.

5. The real-time, all-weather risk control system based on high-frequency data and news sentiment as described in claim 4, characterized in that: The closing module also has a function of linking closing decisions with traceability information. When a user initiates a closing request, the function retrieves the historical data traceability information of the corresponding asset. If an anomaly is found in the data, the closing operation is suspended, the cause of the data anomaly is checked first, and the closing process is executed only after the data is confirmed to be correct.

6. The real-time, all-weather risk control system based on high-frequency data and news sentiment as described in claim 1, characterized in that: The risk control module also has a risk signal classification function. The risk signal classification function classifies the monitored risk signals according to their impact range, probability of occurrence and degree of loss. The classification results are synchronized to the risk mitigation plan automatic generation sub-module, which generates mitigation measures of different levels of detail according to the risk level. For high-level risks, a complete plan including emergency response steps is generated, while for low-level risks, a simplified and optimized version is generated.

7. A real-time, all-weather risk control system based on high-frequency data and news sentiment as described in claim 6, characterized in that: The summary report module also has a risk mitigation effect tracking function. The risk mitigation effect tracking function records the execution process and results of mitigation measures output by the risk mitigation plan automatic generation submodule, compares the execution results with the risk mitigation target, analyzes the actual effect, and if the effect does not meet expectations, the analysis results are fed back to the risk control module to trigger a new round of risk signal monitoring and plan optimization.

8. The real-time, all-weather risk control system based on high-frequency data and news sentiment as described in claim 1, characterized in that: The public opinion and asset correlation evaluation and calibration submodule also has a keyword priority sorting function. The keyword priority sorting function sorts the extracted public opinion keywords according to their closeness to the asset and the scope of public opinion dissemination. The evaluation dimensions corresponding to high-priority keywords are added to the transaction evaluation module and the position evaluation module first, while the evaluation dimensions corresponding to low-priority keywords are used as alternatives and are gradually added after the high-priority dimensions are running stably.

9. A real-time, all-weather risk control system based on high-frequency data and news sentiment as described in claim 8, characterized in that: The cross-cycle evaluation consistency verification submodule also has a difference cause analysis function. When the difference between evaluation results of different cycles exceeds the range, the difference cause analysis function retrieves the high-frequency data changes, news and public opinion content and transaction operation records in the corresponding cycle, analyzes the specific reasons for the difference, classifies and labels the reasons and feeds them back to the machine learning module.

10. A real-time, all-weather risk control system based on high-frequency data and news sentiment as described in claim 9, characterized in that: The machine learning module also has an evaluation logic iterative optimization function. The evaluation logic iterative optimization function receives the difference cause classification results fed back by the cross-cycle evaluation consistency verification submodule, adjusts the input parameters of the analysis model for different cause types, and makes the evaluation logic more suitable for actual application scenarios through continuous iteration.

Images