Method and system for constructing general financial business system based on artificial intelligence

Abstract

The invention provides a method and a system for constructing a common financial business system based on artificial intelligence, and relates to the field of financial system construction. Comprising the following steps: converting financial rule data into analog modulation signal feature vectors containing value intensity, risk association degree and aging attenuation rate through a three-dimensional feature conversion module, and executing market behavior capture, rule space positioning and strategy generation operations in combination with a multi-stage state machine; the supervision text is converted into a continuous control signal injection feature generation logic through a dynamic constraint feedback loop, and a closed-loop adjustment mechanism of value gain calibration, risk attenuation compensation and aging synchronous correction is formed; a physical signal cross validation module is adopted to perform dual coupling verification of main channel waveform characteristics and auxiliary channel time sequence characteristics on the strategy; the time sequence deviation of characteristic conversion, state transition and constraint feedback is synchronously coordinated through a reference clock, and it is ensured that the time sequence deviation of the operation rhythm of all modules of the system is lower than a preset tolerance.

Description

Technical Field

[0001] The present invention relates to the field of financial system construction, and specifically to a method and system for constructing an inclusive financial business system based on artificial intelligence. Background Art

[0002] As a core area of financial technology innovation, inclusive finance is committed to lowering service barriers and improving the efficiency of financial resource allocation through technological means. The integration of artificial intelligence technologies has promoted the rapid development of risk assessment, intelligent decision-making and other fields. However, the complex and ever-changing financial market environment and real-time regulatory needs have placed higher demands on the system's dynamic adaptability. How to achieve efficient rule analysis, accurate strategy generation and reliable execution has become a difficult problem that the industry urgently needs to overcome.

[0003] In existing technologies, some systems process financial data based on traditional machine learning models and generate business strategies by matching preset conditions through rule engines; other solutions use natural language processing technology to extract keywords from regulatory texts, form static constraints and embed them into the decision-making process; in addition, some studies use time series analysis algorithms to monitor market behavior fluctuations and manually adjust strategy parameters based on expert experience; the above methods mostly rely on centralized architecture and software-level logical control, and maintain system operation by updating the rule base at a fixed period.

[0004] However, traditional machine learning models have limited accuracy in modeling nonlinear relationships among multi-dimensional dynamic features, resulting in insufficient ability to capture implicit correlations between rules; static rule engines lack real-time feedback mechanisms, making it difficult to respond to market fluctuations and policy adjustments in a timely manner; data processing and strategy generation are subject to delays under centralized architectures, and high-concurrency scenarios can easily lead to system performance bottlenecks; in addition, the asynchrony between regulatory constraints and market changes may cause policy lags or deviations, affecting the accuracy and compliance of business execution. Summary of the Invention

[0005] (1) Technical problems solved In response to the shortcomings of the existing technology, the present invention provides a method and system for building an inclusive financial business system based on artificial intelligence to solve the problems proposed in the above background technology, such as the traditional machine learning model's insufficient accuracy in modeling the nonlinear relationship of multi-dimensional dynamic features, resulting in weak ability to capture rule correlations; the static rule engine is difficult to respond to market fluctuations and policy changes in a timely manner due to the lack of a real-time feedback mechanism; the data processing delay and high-concurrency scenario performance bottlenecks under the centralized architecture are significant; and the strategy lag or execution deviation caused by the asynchrony of regulatory constraints and market changes.

[0006] (2) Technical solution To achieve the above objectives, the present invention is implemented through the following technical solutions: a method and system for building an inclusive financial business system based on artificial intelligence, comprising the following steps: S1 receives financial business rule data and generates a rule feature vector containing value intensity, risk correlation, and time decay rate through a three-dimensional feature conversion module. The rule feature vector expresses the dynamic change relationship between the three dimensions of value intensity, risk correlation, and time decay rate in the form of an analog modulation signal; S2 uses a multi-stage state machine to process the feature vector, sequentially executing market behavior capture, rule space positioning, association rule activation, topology structure adjustment, business policy generation, and policy persistence operations. Each stage triggers state transitions through hardware-level events. Market behavior capture includes collecting user transaction behavior features. S3 receives regulatory text data and builds a dynamic constraint feedback loop, converts the requirements in the regulatory text data into continuous control signals and injects the control signals into the feature vector generation logic of the three-dimensional feature conversion module, forming a closed-loop regulation mechanism for value gain calibration, risk attenuation compensation, and time synchronization correction; S4 performs dual verification on the generated strategy through the physical signal cross-verification module, and activates the strategy execution when the waveform characteristics of the primary verification channel and the timing characteristics of the secondary verification channel meet the preset coupling conditions; S5 maintains system timing synchronization and coordinates the operating rhythm of each module through a reference clock, so that the timing deviation of feature conversion, state transfer, and constraint feedback is lower than the preset tolerance; the reference clock shares the clock source with the physical signal cross-validation module and synchronizes the operating rhythm of each module through a phase-locked loop circuit.

[0007] Preferably, the three-dimensional feature conversion module parses the text information in the financial business rule data through a natural language processing model to extract economic value parameters, risk transmission parameters, and time impact parameters corresponding to value intensity, risk correlation, and time decay rate; the economic value parameters are logarithmically converted through a hardware multiplier and a logarithmic amplifier circuit, including inputting the normalized economic value parameters into the logarithmic amplifier to generate an amplitude modulation signal, whose dynamic range is controlled by the volatility of the financial market. When an increase in the variance of the price series is detected, the signal amplitude gradient is compressed to suppress excessive fluctuations, and at the same time, the analog-to-digital converter writes the amplitude signal to a specified address of the register, wherein the formula for generating the amplitude modulation signal through logarithmic conversion of the economic value parameters is: , where k1 is the dynamic adjustment coefficient, V max is the maximum normalized value of the economic value parameter. When the price series variance σ is detected 2 When it increases, the dynamic range of the signal is compressed, and the adjustment coefficient k1 is updated as follows: , where α is the compression factor; the risk conduction parameter is inversely converted through a voltage-controlled oscillator. After the input parameter is scaled, it drives the oscillator to output a frequency modulation signal, whose frequency value is inversely proportional to the risk conduction speed. The specific formula is: , where k2 is the proportionality coefficient, To prevent the smallest constant from dividing by zero; and to calibrate the deviation in real time through the frequency counter to ensure that the signal frequency strictly matches the risk parameter; the time-effect parameter is converted into an arc tangent through the phase modulator, and the digital-to-analog converter converts the time-effect parameter into a phase modulation signal. The specific formula is: , where β is the time scaling factor, which is nonlinearly mapped to the remaining validity period; and drives the phase modulator to generate a phase offset, whose phase angle is nonlinearly mapped to the remaining validity period. At the same time, the non-volatile memory stores the historical data of the phase modulation signal to support retrospective analysis; finally, the three-dimensional modulated signals are fused into a regular feature vector through the analog superposition circuit: , where A V ,f R ,φ T The signals are modulated for amplitude, frequency and phase respectively, and the chip temperature changes are monitored by a temperature sensor. The signal reference value is corrected in real time by the second-order filtering algorithm in the compensation circuit to eliminate the signal drift caused by the ambient temperature and ensure the stability of the feature vector generation.

[0008] Preferably, the market behavior capture stage of the multi-stage state machine collects user transaction behavior characteristics at a fixed sampling frequency, converts the transaction behavior into a digital signal through an analog-to-digital converter, and uses a circular buffer and a sliding window algorithm to count the transaction frequency per unit time in real time. In specific implementation, each sampling updates the tail data of the buffer and triggers the accumulator to calculate the mean and variance of the data in the window, and the statistical results are transmitted to the rule space positioning module through the bus; the rule space positioning stage uses a hardware accelerator to optimize the nearest neighbor search algorithm to calculate the Euclidean distance between the current market state and all rule nodes in the three-dimensional feature space. The three-dimensional coordinates are based on value strength, risk correlation, and time decay rate. The error tolerance is set to a preset threshold to reduce the calculation complexity. The neighboring node information is cached in a high-speed cache and a fast access channel is established, wherein the nearest neighbor search algorithm is used to calculate the current market state coordinates With all rule nodes The Euclidean distance is: , where V, R, and T are value intensity, risk correlation, and time decay rate, respectively. In the association rule activation phase, the feature similarity of adjacent nodes is compared through a parallel comparator array. When the similarity exceeds the activation threshold, the crossbar switch matrix establishes an association link. The link weight is generated by the analog-to-digital converter according to the similarity value to generate an analog voltage signal, which drives the variable resistor to adjust the connection strength. In the topology structure adjustment phase, the algebraic sum of the connection strengths of all nodes is calculated through an accumulator. If the algebraic sum is not zero, the weight of each node is scaled proportionally. Specifically, , where w iis the original connection strength of node i, and N is the total number of nodes. The top three links are retained through the priority queue, and the strengths of the remaining links are set to zero. The updated topology is written into the register. At the same time, the dynamic balance constraint rule monitors the changes in connection strength in real time through the feedback loop to ensure system stability.

[0009] Preferably, the dynamic constraint feedback loop parses the regulatory text data into constraint elements, and encodes them into time-continuous control signals through a finite state machine, with the pulse width being proportional to the constraint strength; the control signal is filtered through a programmable low-pass filter to eliminate high-frequency noise, and the filter cutoff frequency is dynamically adjusted according to the rule update frequency. Specifically, the digital signal processor calculates the rule update rate in real time and drives the voltage-controlled filter to adjust the parameters. When the user transaction frequency statistic is higher than the preset threshold, the cutoff frequency is inversely proportional to the frequency value to suppress high-frequency interference; the processed control signal is decomposed into three independent channels: the value gain calibration channel superimposes the control signal on the reference voltage of the amplitude modulation signal through an adder, the risk attenuation compensation channel adjusts the duty cycle of the frequency modulation signal through a subtractor, and the time synchronization correction channel aligns the phase modulation signal with the system clock through a phase comparator; the closed-loop adjustment mechanism monitors the corrected eigenvector deviation in real time, and triggers the control signal to regenerate the process if the deviation exceeds the threshold. At the same time, the feedback loop ensures that the adjustment signal is strictly synchronized with the market rhythm through a phase-locked loop circuit.

[0010] Preferably, the main verification channel of the physical signal cross-verification module detects the phase difference of the characteristic vector through a phase-locked loop circuit. If the absolute value of the phase difference exceeds a threshold, an alarm signal is triggered and the strategy execution is interrupted; the auxiliary verification channel uses a counter to count the duty cycle distribution of the pulse signal. If the ratio of the high-level duration to the period does not meet the preset logical conditions, the strategy is marked as invalid; the double verification result is judged by a logic AND gate. Only when the main channel does not alarm and the auxiliary channel passes the review, the strategy execution port is activated and a valid mark is written to the non-volatile memory; during the verification process, the reference clock synchronizes the timing of the main and auxiliary channels through the phase-locked loop circuit to ensure the time consistency of the verification results. At the same time, the clock deviation monitoring circuit detects the clock signals of each module in real time. If the deviation exceeds the tolerance, the calibration process is triggered and data writing is suspended to ensure data integrity.

[0011] Preferably, the reference clock generates a master clock signal from a high-precision crystal oscillator and is distributed to each module through a phase-locked loop circuit; the clock division coefficient of the three-dimensional feature conversion module is set to an integer multiple of the data sampling rate to ensure synchronization between analog-to-digital conversion and signal modulation; the clock division coefficient of the multi-stage state machine matches the sliding window update period to seamlessly connect the market behavior capture and rule positioning stages; the timing monitoring circuit detects the clock deviation of each module in real time, and if the deviation exceeds the preset tolerance, the phase-locked loop is triggered to resynchronize the clock signal; during the calibration process, the non-volatile memory suspends the write operation and resumes after the clock stabilizes to avoid data misalignment; at the same time, the temperature sensor monitors environmental changes and adjusts the clock signal reference value through the compensation circuit to further improve the timing accuracy.

[0012] Preferably, the three-dimensional feature conversion module detects the difference in amplitude of value intensity signals of different rules in real time through a comparator. If the difference exceeds a conflict threshold, the amplitude limiting mechanism is triggered; the limiting mechanism is implemented by a clamping circuit, which limits the signal amplitude to not exceed the system safety threshold through a diode array, and at the same time, the dynamic range compression mechanism is suspended to avoid signal distortion; the limited signal is resampled by an analog-to-digital converter and updated to a non-volatile memory, and the status register records the limiting status for reference in the topology adjustment stage; if a temperature anomaly is detected, the second-order filtering algorithm in the compensation circuit automatically adjusts the signal reference value, and combines with the limiting mechanism to jointly suppress environmental interference.

[0013] Preferably, the sliding window algorithm stores user transaction behavior data in a circular queue, where the queue length is consistent with the window size. Each time a sample is taken, the new data overwrites the oldest data at the end of the queue. The accumulator calculates the sum of the data in the window in real time, and accumulates the square of the difference between each data point and the mean, and then divides it by the window size to generate the variance. The variance of the transaction frequency in the sliding window is calculated as: ,in is the window mean, and N is the window size; the statistical results are transmitted to the rule space positioning module through the bus, and the error tolerance threshold of the nearest neighbor search algorithm is dynamically adjusted; the update period of the sliding window is controlled by a hardware timer to ensure strict synchronization with the sampling frequency of the market behavior capture stage, and the statistical results are written to the shared memory for calling by the multi-stage state machine.

[0014] Preferably, the association rule activation stage accelerates feature similarity calculation by a parallel comparator array, each comparator independently processes the feature vectors of a pair of rule nodes, and outputs cosine similarity to the priority encoder; if the similarity exceeds the activation threshold, the encoder generates a link establishment instruction, wherein the association rule activation stage calculates the node feature vector F i With F j The cosine similarity is , if , then establish an associated link; link weight w ijThe similarity linear mapping is The cross-switch matrix connects the corresponding nodes according to the instructions; the link weight is converted by the digital-to-analog converter into an analog voltage signal, which drives the variable resistor to adjust the connection strength; the state machine controls the timing of the comparator array and the switch matrix according to the system clock signal to ensure low-latency response; the adjacent node cache information is preloaded to the comparator through the high-speed bus to further shorten the response time.

[0015] Preferably, in the topology adjustment stage, the algebraic sum of the connection strengths of all nodes is calculated through an accumulator. If the algebraic sum is not zero, the divider calculates the scaling factor based on the total number of nodes and adjusts the weight of each node proportionally; the scaled strength value is written into the connection weight register through the digital-to-analog converter, and at the same time, the priority queue filters the top three links and sets the remaining links to zero; the dynamic balance constraint monitors the change in connection strength in real time through a feedback loop, and triggers a renormalization process if an imbalance is detected; during the normalization process, the accumulator and the divider work together through a hardware pipeline to ensure computational efficiency and real-time performance.

[0016] Preferably, the cutoff frequency of the low-pass filter is dynamically adjusted according to the rule update frequency, and the digital signal processor calculates the rule update rate in real time and drives the voltage-controlled filter to adjust the parameters; when the user transaction frequency statistical value is higher than the preset threshold, the cutoff frequency is inversely proportional to the frequency value to suppress high-frequency noise, which is specifically achieved through an inverse proportional amplifier; the filtering timing is strictly aligned with the system master clock through a hardware synchronization circuit to ensure that the control signal is synchronized with the market rhythm; the filtered signal is decomposed into three independent channels by an analog-to-digital converter, which act on the feature correction of value intensity, risk correlation, and time decay rate respectively, and each channel is equipped with an independent gain adjustment circuit to adapt to calibration requirements of different dimensions.

[0017] Preferably, the policy persistence operation is implemented through the page write mechanism of the non-volatile memory, and the characteristic vector and execution mark of the effective policy are written in a page-aligned format; before writing, the memory controller checks the erase status of the target page, and triggers the sector erase instruction if it is not erased; during the writing process, the temperature sensor data adjusts the signal reference value in real time through the compensation circuit to ensure data reliability; after the writing is completed, the verification circuit compares the header check code and the tail redundant check bit, and triggers the rewrite process if they are inconsistent; the memory address mapping table records the storage location of each policy to support fast retrieval and update.

[0018] (3) Beneficial effects The present invention provides a method and system for building an inclusive financial business system based on artificial intelligence. It has the following beneficial effects: 1. The present invention maps financial rule data into multi-dimensional modulated signals through a three-dimensional feature conversion module, and uses logarithmic, inverse proportional, and inverse tangent functions to achieve nonlinear feature expression, significantly enhancing the ability to characterize complex market behaviors; the multi-stage state machine compresses market behavior capture, rule positioning, topology adjustment and other operations to millisecond-level response through hardware-level event triggering and parallel computing architecture, greatly improving real-time performance; the dynamic constraint feedback loop combines control signals with feature correction mechanisms to ensure that the system can quickly calibrate strategies when regulatory rules change and adapt to high-frequency fluctuations in the financial market; the physical signal cross-validation module uses timing consistency detection and logical completeness review to doubly ensure the credibility of strategy generation, reduce the need for manual intervention, and overall achieve automated and efficient operation of the entire process from data analysis to strategy execution.

[0019] 2. The dynamic balance constraint rule of the present invention forces the algebraic sum of connection strengths to be zero through a normalization algorithm, and combines with a priority link screening mechanism to effectively prevent policy deviations caused by topology imbalance; the temperature compensation circuit and the second-order filtering algorithm work together to eliminate the signal reference drift caused by ambient temperature changes in real time, ensuring the stability of feature vector generation; the amplitude limiting mechanism and the dynamic range compression strategy are automatically activated when rules conflict or the market fluctuates violently, suppressing the risk of signal overload and avoiding hardware damage; the page writing and redundancy check mechanism of the non-volatile memory ensures that the policy data is complete and persistent in the event of abnormal power outages or system failures; the phase-locked loop synchronization and timing monitoring circuit of the reference clock strictly align the operating rhythm of each module, reduce the accumulation of timing deviations, and ultimately achieve high robustness and long-term reliable operation of the system in complex environments. DETAILED DESCRIPTION

[0020] The following is a clear and complete description of the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present invention.

[0021] The embodiment of the present invention provides a method and system for building an inclusive financial business system based on artificial intelligence. Specifically, when the system starts, the three-dimensional feature conversion module receives original text data from the financial business rule database, parses the text information through the natural language processing model, and extracts economic value parameters, risk transmission parameters, and time impact parameters, which correspond to the three dimensions of value intensity, risk correlation, and time decay rate respectively; the economic value parameters are processed by the logarithmic conversion unit and input into the logarithmic amplifier circuit to generate an amplitude modulation signal, whose dynamic range is dynamically adjusted by the volatility of the financial market. When the variance of the price series increases, the signal amplitude gradient is compressed to suppress excessive fluctuations. At the same time, the analog-to-digital converter converts the amplitude signal into a digital signal. The signal is written into the designated address of the non-volatile memory; the risk conduction parameter is converted into a frequency modulation signal through a voltage-controlled oscillator, and the frequency value is inversely proportional to the risk conduction speed. The deviation is calibrated in real time by a frequency counter to ensure that the signal frequency strictly matches the risk parameter; the time-influence parameter is generated by a phase modulator to generate a phase offset signal, and its phase angle is nonlinearly mapped to the proportion of the remaining validity period. The digital-to-analog converter writes the phase parameter into the memory and stores historical data to support retrospective analysis; the three-dimensional modulated signals are fused into a regular feature vector through an analog superposition circuit. At the same time, the temperature sensor monitors the chip temperature change and the signal reference value is corrected in real time through a second-order filter compensation circuit to eliminate the drift caused by the ambient temperature. The multi-stage state machine module adopts a hardware-level event trigger mechanism. In the market behavior capture stage, user transaction behavior data is collected at a fixed sampling frequency. After the analog-to-digital converter converts the transaction behavior into a digital signal, the transaction frequency per unit time is counted in real time through a circular buffer and a sliding window algorithm. Each sampling updates the tail data of the buffer and triggers the accumulator to calculate the mean and variance of the data in the window. The statistical results are transmitted to the rule space positioning module through the bus; in the rule space positioning stage, a hardware accelerator is used to optimize the nearest neighbor search algorithm to calculate the Euclidean distance between the current market state and all rule nodes in the three-dimensional feature space. The error tolerance is set to a preset threshold to reduce the computational complexity. , the neighboring node information is cached in the cache and a fast access channel is established; in the association rule activation stage, the feature similarity of adjacent nodes is compared through a parallel comparator array. When the similarity exceeds the activation threshold, the crossbar switch matrix establishes an association link. The link weight is generated by the digital-to-analog converter according to the similarity value. The analog voltage signal drives the variable resistor to adjust the connection strength; in the topology adjustment stage, the algebraic sum of the connection strengths of all nodes is calculated through the accumulator. If the algebraic sum is not zero, the weight of each node is scaled proportionally, and the top three links are retained through the priority queue. The strength of the remaining links is set to zero. The updated topology is written to the register and fed back to the dynamic balance constraint module in real time; After receiving regulatory text data, the dynamic constraint feedback module parses the constraint elements and encodes them into continuous control signals through a finite state machine. The pulse width is proportional to the constraint strength. The control signal is filtered through a programmable low-pass filter to eliminate high-frequency noise. The filter cutoff frequency is dynamically adjusted according to the rule update frequency. When the user's transaction frequency statistics are higher than the preset threshold, the cutoff frequency is inversely proportional to the frequency value to suppress high-frequency interference. The processed control signal is decomposed into three independent adjustment channels. The value gain calibration channel superimposes the control signal on the reference voltage of the amplitude modulation signal through an adder. The risk attenuation compensation channel adjusts the duty cycle of the frequency modulation signal through a subtractor. The time synchronization correction channel aligns the phase modulation signal with the system clock through a phase comparator. The closed-loop adjustment mechanism monitors the deviation of the corrected eigenvector in real time. If the deviation exceeds the threshold, the control signal regeneration process is triggered. At the same time, the feedback loop ensures that the adjustment signal is strictly synchronized with the market rhythm through a phase-locked loop circuit. The physical signal cross-validation module detects the phase difference of the feature vector through the main verification channel. If the absolute value of the phase difference exceeds the threshold, an alarm signal is generated and the policy execution is interrupted. The auxiliary verification channel uses a counter to count the duty cycle distribution of the pulse signal. If the ratio of the high-level duration to the period does not meet the preset logical conditions, the policy is marked as invalid. The double verification result is judged by the logic AND gate. Only when the main channel does not alarm and the auxiliary channel passes the review, the policy execution port is activated and a valid mark is written to the non-volatile memory. During the verification process, the reference clock synchronizes the timing of the main and auxiliary channels through the phase-locked loop circuit, and the clock deviation monitoring circuit detects the clock signals of each module in real time. If the deviation exceeds the tolerance, the calibration process is triggered and data writing is suspended. The timing synchronization module generates a master clock signal using a high-precision crystal oscillator, and the phase-locked loop circuit distributes the clock to each module. The clock division coefficient of the three-dimensional feature conversion module is set to an integer multiple of the data sampling rate to ensure synchronization between analog-to-digital conversion and signal modulation. The clock division coefficient of the multi-stage state machine matches the sliding window update period, ensuring a seamless connection between the market behavior capture and rule positioning stages. The timing monitoring circuit detects clock deviations in each module in real time and triggers the phase-locked loop to resynchronize if the deviation exceeds the preset tolerance. During the calibration process, the non-volatile memory suspends write operations and resumes after the clock stabilizes. The temperature sensor monitors environmental changes and adjusts the clock signal reference value through a compensation circuit, further eliminating temperature drift in combination with a second-order filtering algorithm. After receiving the verified policy signal, the policy execution and storage module executes the financial business operation through the policy activation port, and the policy parameters and execution results are written to the non-volatile memory in a page-aligned format; before writing, the memory controller checks the erase status of the target page, and triggers the sector erase instruction if it is not erased; during the writing process, the temperature sensor data adjusts the signal reference value in real time through the compensation circuit to ensure data reliability; after the writing is completed, the verification circuit compares the header check code and the tail redundant check bit, and triggers the rewrite process if they are inconsistent; the memory address mapping table records the storage location of each policy to support fast retrieval and update; when the system is running, the dynamic balance constraint rule forces the algebraic sum of the connection strength to be zero through the normalization algorithm, the priority link screening mechanism retains key connections, and the amplitude limiting mechanism and dynamic range compression strategy are automatically activated when there is a rule conflict or market fluctuation, suppressing the risk of signal overload, and ultimately achieving full-process automated, efficient operation and high-robust operation.

[0022] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for building an inclusive financial business system based on artificial intelligence, characterized in that: The steps include: S1 receives financial business rule data and generates a rule feature vector containing value intensity, risk correlation, and time decay rate through a three-dimensional feature conversion module. The rule feature vector expresses the dynamic change relationship between the three dimensions of value intensity, risk correlation, and time decay rate in the form of an analog modulation signal; S2 uses a multi-stage state machine to process the feature vector, sequentially executing market behavior capture, rule space positioning, association rule activation, topology structure adjustment, business policy generation, and policy persistence operations. Each stage triggers state transitions through hardware-level events. Market behavior capture includes collecting user transaction behavior features. S3 receives regulatory text data and builds a dynamic constraint feedback loop, converts the requirements in the regulatory text data into continuous control signals and injects the control signals into the feature vector generation logic of the three-dimensional feature conversion module, forming a closed-loop regulation mechanism for value gain calibration, risk attenuation compensation, and time synchronization correction; S4 performs dual verification on the generated strategy through the physical signal cross-verification module, and activates the strategy execution when the waveform characteristics of the primary verification channel and the timing characteristics of the secondary verification channel meet the preset coupling conditions; S5 maintains system timing synchronization and coordinates the operating rhythm of each module through a reference clock, so that the timing deviation of feature conversion, state transfer, and constraint feedback is lower than the preset tolerance; the reference clock shares the clock source with the physical signal cross-validation module and synchronizes the operating rhythm of each module through a phase-locked loop circuit.

2. The method for building an inclusive financial business system based on artificial intelligence according to claim 1, characterized in that: The S1 step includes: S11 parses the text information corresponding to the input financial business rule data based on the financial business rule data received by S1, and extracts the value intensity, the risk correlation, and the time decay rate into quantifiable semantic elements of economic value parameters, risk transmission parameters, and time impact parameters, respectively, through a natural language processing model; S12 applies a logarithmic transformation relationship to the value parameter to generate an amplitude modulation signal, whose signal strength is nonlinearly positively correlated with the rule-based economic contribution; S13 applies an inverse proportional conversion relationship to the risk parameter to generate a frequency modulation signal, where the signal change rate is inversely proportional to the risk transmission speed; S14 applies an inverse tangent conversion relationship to the aging parameter to generate a phase modulation signal, wherein the phase angle and the remaining validity period ratio are nonlinearly mapped, and the signal is written into a nonvolatile memory through the three-dimensional feature conversion module.

3. The method for building an AI-based inclusive financial business system according to claim 2, characterized in that: The dynamic adjustment of the logarithmic conversion relationship in S12 includes: S121 automatically compresses the signal dynamic range based on the variance value of the price sequence in the user's trading behavior characteristics, and the signal amplitude change gradient is negatively correlated with the volatility of the financial market; S122 activates the amplitude limiting mechanism when a rule conflict is detected, constraining the maximum output signal to not exceed the system safety threshold; S123: The three-dimensional feature conversion module has a built-in chip temperature sensor and a signal compensation circuit. The temperature sensor data is used to compensate the signal reference value in real time through the temperature compensation circuit, and the second-order filtering algorithm in the temperature compensation circuit is used to eliminate the signal reference value drift caused by the external ambient temperature change. When the amplitude limiting mechanism described in S124 is started, the dynamic range compression mechanism is automatically paused.

4. The method for building an AI-based inclusive financial business system according to claim 1, characterized in that: The control logic of the multi-stage state machine in S2 includes: In the S21 market behavior capture phase, user transaction behavior characteristics are collected at a fixed sampling frequency and converted into digital signals. At the same time, the unit time frequency statistics of the user transaction behavior characteristics are calculated in real time using a sliding window algorithm; The S22 rule space positioning stage uses the nearest neighbor search algorithm to determine the current market state coordinates in the three-dimensional feature space; In the S23 association rule activation phase, feature similarities of adjacent rule nodes are compared in parallel, and an association link is established when the similarity exceeds the activation threshold; In the S24 topology adjustment phase, the connection strength between rule nodes is dynamically adjusted according to market changes. The connection strength adjustment must meet the dynamic balance constraint rule, and the algebraic sum of the connection strength changes of all nodes is made zero through the normalization algorithm, and the links with the top three connection strength values are retained during execution.

5. The method for building an inclusive financial business system based on artificial intelligence according to claim 4, characterized in that: The execution process of the nearest neighbor search algorithm includes: S221 calculates the multidimensional space distance between the current market state and all rule nodes, wherein the multidimensional space distance is calculated using the three-dimensional coordinates of value intensity, risk correlation, and time decay rate described in S1; S222 uses the nearest neighbor search algorithm to accelerate the distance sorting process, and the error tolerance is set to a preset threshold to reduce the computational complexity while ensuring accuracy; S223 caches the neighboring node information including the connection strength ranking result, so as to support the operation of retaining the strength link and establish a fast access channel to shorten the response time of the subsequent association rule activation phase.

6. The method for building an inclusive financial business system based on artificial intelligence according to claim 1, characterized in that: The construction process of the dynamic constraint feedback loop in S3 includes: S31 parses the regulatory text and extracts constraint elements, encoding them into a control signal with time continuity; S32 eliminates high-frequency noise components in the control signal through filtering algorithms, retaining the effective regulation components that are synchronized with the market rhythm; S33 decomposes the processed control signal into three independent adjustment channels, which act on the feature correction of the three dimensions of value intensity, risk correlation and time decay rate respectively.

7. The method for building an AI-based inclusive financial business system according to claim 6, characterized in that: The parameter adjustment logic of the filtering algorithm includes: S321 dynamically sets the filter cutoff frequency according to the rule update frequency, so that the filter cutoff frequency value is linearly positively correlated with the error tolerance threshold of the nearest neighbor search algorithm described in S222; S322 automatically adjusts the filtering strength based on the unit time frequency statistics of the user transaction behavior characteristics by adjusting the cutoff frequency. When the unit time frequency statistics are higher than a preset threshold, the cutoff frequency is inversely proportional to the unit time frequency statistics. The preset threshold is dynamically adjusted based on the real-time frequency statistics output by the sliding window algorithm in S21. The S323 uses a hardware-level signal synchronization mechanism to strictly align the filtering timing with the system master clock.

8. The method for building an inclusive financial business system based on artificial intelligence according to claim 1, characterized in that: The execution logic of the physical signal cross-validation in S4 includes: The S41 main verification channel analyzes the phase relationship of the feature vectors, detects the timing consistency during the strategy generation process, and generates an alarm signal when the absolute value of the phase difference exceeds the threshold; The S42 auxiliary verification channel measures the duty cycle distribution of the pulse signal to verify the logical completeness of the strategy decision; S43: When and only when the primary verification channel does not generate an alarm signal and the review process of the secondary verification channel passes, activate the policy execution port and write a valid mark to the non-volatile memory.

9. The method for building an artificial intelligence-based inclusive financial business system and the system for building an artificial intelligence-based inclusive financial business system according to claim 1 are characterized in that: Includes the following modules: The three-dimensional feature conversion module is used to receive financial business rule data and generate rule feature vectors containing value intensity, risk correlation, and time decay rate, dynamically expressing the three-dimensional feature relationship through analog modulation signals; Its built-in logarithmic conversion unit, inverse proportional conversion unit and inverse tangent conversion unit respectively convert economic value parameters, risk transmission parameters and time-effect impact parameters into amplitude, frequency and phase modulation signals and write them into non-volatile memory; The module integrates a temperature sensor and a second-order filter compensation circuit to eliminate signal drift caused by ambient temperature; The multi-stage state machine module uses multi-stage control logic triggered by hardware-level events to sequentially execute market behavior capture, rule space positioning, association rule activation, topology structure adjustment, business policy generation, and persistence operations. It includes a sliding window algorithm unit to count user transaction frequency in real time, a nearest neighbor search unit to locate market state coordinates, a parallel comparison unit to activate associated links, and a normalization adjustment unit to dynamically balance connection strength and retain the top three links. The dynamic constraint feedback module is used to parse regulatory text data and generate continuous control signals. After eliminating high-frequency noise through a filtering algorithm, it is decomposed into three adjustment channels: value gain calibration, risk attenuation compensation, and time synchronization correction. Its filter cutoff frequency is linearly related to the nearest neighbor search error tolerance, and is dynamically adjusted based on the frequency of user transaction behavior. It is aligned with the system master clock through a hardware-level synchronization mechanism. The physical signal cross-verification module includes a primary verification channel and a secondary verification channel. The primary channel verifies strategy timing consistency through phase difference detection, while the secondary channel verifies logical integrity through pulse duty cycle. After dual-channel signal coupling, the strategy execution port is activated and a valid flag is written to non-volatile memory. The module shares a clock source with the system reference clock and maintains timing synchronization through a phase-locked loop circuit. The timing synchronization module coordinates the operation rhythm of each module through a reference clock to ensure that the timing deviation of feature conversion, state transfer, and constraint feedback is below the preset tolerance. It integrates a phase-locked loop circuit and a clock source distribution unit to achieve hardware-level synchronization of the operation rhythm between modules. The policy execution and storage module is used to receive verified policy signals and execute financial business operations, while persistently storing policy parameters and execution results in non-volatile memory. The module includes a policy activation port, a security threshold limiting unit, and a fast access channel to support low-latency response in high-concurrency business scenarios.