Financial risk identification system driven by big data

Abstract

The invention relates to the technical field of risk identification, in particular to a big data driven financial risk identification system, which comprises a time data synchronization module, a multi-dimensional feature mapping module, a financial behavior analysis module and a risk dynamic identification module. According to the method, the financial data flow is synchronized in real time, multi-dimensional scale transformation is applied, and capturing and processing of financial data are optimized, so that more efficient data flow analysis is realized, subtle changes and key forms of time sequence data can be accurately captured in a complex financial environment, and the sensitivity and accuracy of risk identification are greatly improved; refined similarity calculation and behavior recognition help the system to find potential risks earlier and ensure that financial institutions can respond quickly, so that risk events are effectively prevented and reduced, the market with dynamic changes is dealt with, and the rationality of risk management strategy formulation and resource allocation can be remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of risk identification, and particularly to a big data-driven financial risk identification system. Background Art

[0002] The technical field of risk identification is an important branch of financial technology, which involves using technical means to predict and identify potential financial risks, widely applying algorithms, big data analysis and machine learning technologies, and identifying risk factors by analyzing historical data and real-time data. The purpose is to enhance the risk management ability of financial institutions, optimize resource allocation, and improve decision-making efficiency. Technical applications include but are not limited to credit scoring, market risk analysis, fraud detection, and operational risk management.

[0003] Among them, a financial risk identification system refers to an integrated technical system designed to identify and evaluate potential risks in the financial environment in real time, so as to take timely measures for risk mitigation. Usually based on advanced data analysis and prediction models, it can process a large amount of data and extract risk signals from it, and has a wide range of uses. For example, it helps banks and financial institutions evaluate the risks of loan applications, monitor the market risks of investment portfolios, or provide risk predictions for insurance companies to optimize their insurance products and pricing strategies.

[0004] Although the existing technologies adopt various data analysis and machine learning technologies for risk management, they often have deficiencies in data synchronization and real-time processing, frequently encountering problems such as delays and insufficient accuracy when processing synchronization in a big data environment, especially in cases where real-time analysis and rapid response are required, and often unable to provide sufficient support. In addition, the existing systems are usually limited in their ability to identify subtle changes in data, which restricts the effect of predicting and identifying risks in a rapidly changing market environment. For example, when the market fluctuates rapidly, traditional systems may not be able to capture key signals in time, resulting in untimely or inaccurate risk assessments. The time lag and insensitivity may cause financial institutions to face losses from sudden risks. Summary of the Invention

[0005] The purpose of the present invention is to solve the deficiencies existing in the prior art, and to propose a big data-driven financial risk identification system.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions: A big data-driven financial risk identification system includes:

[0007] A time data synchronization module collects real-time financial data streams, performs time tag synchronization, calculates the time difference of each data point, executes dynamic time adjustment according to the calculation results, performs frequency correction and time mark calibration on the data, and outputs synchronized time series data;

[0008] The multidimensional feature mapping module converts the time data into a spatial feature vector through multidimensional scaling based on the synchronized time series data, performs dimension mapping on the data at each time point, extracts key spatial features, and generates a mapping feature set;

[0009] The financial behavior analysis module uses feature analysis to extract sequence patterns based on the mapping feature set, identifies key morphological changes in the time series, monitors and compares historical trend data in real time based on the identification results, calculates the similarity between data points, reveals abnormal behavior of financial data, and outputs behavior identification results;

[0010] Based on the behavior identification results, the risk dynamic identification module evaluates the risk level of abnormal behavior by comparing the risk thresholds at different time points, determines potential risk threats based on the evaluation results, classifies the risk level and marks sensitive events, and outputs financial risk warning records.

[0011] As a further solution of the present invention, the step of calculating the time difference is:

[0012] Perform data collection initialization processing, establish a secure connection with the financial market data interface, batch pull real-time transaction data of different financial products, and generate original real-time financial data streams;

[0013] Based on the original real-time financial data stream, add an exact timestamp of the server receiving each data, synchronize the time label, and obtain a real-time financial data stream with a time label;

[0014] Using the real-time financial data stream with time tags, the formula is adopted:

[0015] Δt i =t i+1 -t i

[0016] Calculate the time difference Δt between adjacent data points i , generate time difference data, where t i Represents the timestamp of the i-th data point, t i+1 Represents the timestamp of the i+1th data point.

[0017] As a further solution of the present invention, the step of acquiring the synchronous time series data is:

[0018] The time difference data is used to perform linear interpolation using the formula:

[0019]

[0020] Calculate a new timestamp t for each data point new,i , get the adjusted timestamp, where t iis the original timestamp, Δt i is the time difference value, g is the normalization constant, and n is the total number of data points;

[0021] Use the adjusted timestamp to perform frequency correction on the data stream, ensure that the data acquisition frequency matches the preset standard, analyze the original data frequency through Fourier transform and apply a filter to adjust it to the target frequency, and generate a frequency-corrected data stream;

[0022] Perform timestamp calibration based on the frequency-corrected data stream, apply time series analysis to relabel the timestamp of each data point, and output synchronized time series data.

[0023] As a further solution of the present invention, the steps for obtaining the mapping feature set are as follows:

[0024] Based on the synchronized time series data, perform multidimensional scaling transformation, preliminarily locate the multidimensional space position of each time data point, and generate a primary space feature vector set;

[0025] Using the primary space feature vector set, adopt the formula:

[0026]

[0027] Calculate the Euclidean distance d between vector i and vector j ij , and generate an optimized space feature vector set, where v ik and v jk are the coordinate values of vectors i and j in the kth dimension, and m represents the number of coordinate dimensions;

[0028] According to the optimized space feature vector set, synthesize the spatial attributes and relationships of the vectors, reveal the key spatial characteristics of the original data, and generate mapping features.

[0029] As a further solution of the present invention, the steps for identifying the key morphological changes are as follows:

[0030] According to the mapping feature set, extract key time-intensity feature data, and use a sliding time window to generate windowed feature data;

[0031] Perform principal component analysis on the windowed feature data, compress the data dimension, extract key information, and generate dimension-reduced feature data;

[0032] Input the dimension-reduced feature data into a deep learning model to identify and classify the key morphological changes of the time series, and adopt the formula:

[0033]

[0034] Calculate the pattern probability p(t) of the time series, and generate the identification result of the key morphological changes, where αh is the influence coefficient for the h-th mode, γ h represents the attenuation parameter corresponding to the corresponding mode, μ h is the central time point of the corresponding mode, and M represents the number of modes.

[0035] As a further aspect of the present invention, the step of obtaining the behavior recognition result is as follows:

[0036] Utilize the key morphological change recognition result to continuously monitor new time series data, compare it with historical trend data, perform difference analysis, and generate a real-time monitoring and comparison result;

[0037] Based on the real-time monitoring and comparison result, extract data points and use the formula

[0038]

[0039] Calculate the similarity coefficient cs, perform similarity evaluation, and generate a data point similarity evaluation result, where A and B respectively represent the feature vectors of the current data point and the historical data point, and ‖A‖ and ‖B‖ respectively represent the norms;

[0040] Based on the data point similarity evaluation result, perform threshold determination to analyze the abnormality degree of the data points, identify the data points that significantly deviate from the normal mode, reveal potential abnormal financial data behaviors, and generate a behavior recognition result.

[0041] As a further aspect of the present invention, the step of evaluating the risk level is as follows:

[0042] Utilize the behavior recognition result to extract the key time points in the behavior data, compare them with the set benchmark risk threshold, and generate a behavior feature comparison result for each time point;

[0043] Analyze the behavior feature comparison result for each time point and use the formula:

[0044]

[0045] Calculate the risk index R(t) for each time point to obtain a preliminary risk assessment result, where ΔB(t) represents the deviation of the behavior data from the benchmark threshold, and B base is the benchmark behavior threshold;

[0046] According to the preliminary risk assessment result, classify the risk levels for each time point that deviates from the benchmark threshold, determine the risk level for each time point, and generate a risk level evaluation result.

[0047] As a further aspect of the present invention, the step of obtaining the financial risk warning record is as follows:

[0048] Analyze the risk level assessment results, identify abnormal behavior events, and generate a list of potential risk threats;

[0049] Based on the list of potential risk threats, classify them according to the degree of risk, mark the abnormal events as sensitive events, and use the formula:

[0050]

[0051] Calculate the percentage S of sensitive events in the total events to obtain the risk degree classification result, where R high is the number of abnormal events, and R total is the total number of events;

[0052] According to the risk degree classification result, based on the characteristics and historical data analysis of sensitive events, output a financial risk warning record.

[0053] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0054] In the present invention, by real-time synchronizing financial data streams and applying multidimensional scaling transformation, the capture and processing of financial data are optimized, more efficient data stream analysis is achieved, and the subtle changes and key patterns of time series data can be accurately captured in a complex financial environment, greatly improving the sensitivity and accuracy of risk identification. The refined similarity calculation and behavior recognition help the system detect potential risks earlier, ensure that financial institutions can respond quickly, thus effectively preventing and mitigating risk events, and then coping with the dynamically changing market, and can significantly improve the formulation of risk management strategies and the rationality of resource allocation. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 is the system flow chart of the present invention;

[0056] Figure 2 is the calculation flow chart of the time difference of the present invention;

[0057] Figure 3 is the acquisition flow chart of the synchronous time series data of the present invention;

[0058] Figure 4 is the acquisition flow chart of the mapping feature set of the present invention;

[0059] Figure 5 is the identification flow chart of the key pattern changes of the present invention;

[0060] Figure 6 is the acquisition flow chart of the behavior recognition result of the present invention;

[0061] Figure 7 is the assessment flow chart of the risk level of the present invention;

[0062] Figure 8 This is the flowchart for obtaining the financial risk warning records of the present invention. Detailed implementation manners

[0063] In order to make the objectives, technical solutions and advantages of the present invention more clear and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0064] In the description of the present invention, it should be understood that the orientation or positional relationships indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. are based on the orientation or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation of the present invention. In addition, in the description of the present invention, "a plurality of" means two or more, unless otherwise specifically defined.

[0065] Please refer to Figure 1 , a financial risk identification system driven by big data includes:

[0066] The time data synchronization module collects real-time financial data streams, performs time tag synchronization, calculates the time differences of each data point, performs dynamic time adjustment according to the calculation results, corrects the data frequency and calibrates the time stamps, and outputs synchronized time series data;

[0067] The multi-dimensional feature mapping module converts the time data into spatial feature vectors through multi-dimensional scaling transformation based on the synchronized time series data, performs dimensional mapping on the data of each time point, extracts key spatial features, and generates a mapping feature set;

[0068] The financial behavior analysis module extracts sequence patterns by feature analysis according to the mapping feature set, identifies key morphological changes in the time series, monitors and compares historical trend data in real time according to the identification results, calculates the similarity between data points, reveals abnormal financial data behaviors, and outputs behavior identification results;

[0069] The risk dynamic identification module evaluates the risk levels of abnormal behaviors by comparing the risk thresholds of different time points based on the behavior identification results, determines potential risk threats according to the evaluation results, classifies the risk levels and marks sensitive events, and outputs financial risk warning records.

[0070] Synchronized time series data includes data alignment metric results, calibrated frequency values, and marked synchronization points. The mapped feature set includes a feature vector sequence, dimension identifiers, and key feature marker records. The behavior recognition results include morphological variation metrics, sequence similarity scores, and an anomaly metric set. The financial risk warning record specifically refers to risk level classification, threat recognition identifiers, and event sensitivity marking results.

[0071] Please refer to Figure 2 , and the calculation steps for time difference are as follows:

[0072] Perform data acquisition initialization processing, establish a secure connection with the financial market data interface, batch pull real-time transaction data of different financial products, and generate an original real-time financial data stream;

[0073] When initializing the data acquisition module, it is necessary to ensure the security and reliability of the connection with the financial market data interface. To ensure the security and integrity of data transmission, first, through data encryption and authentication protocols, ensure that data is not stolen or tampered with during transmission. Subsequently, use a real-time data monitoring system to monitor the quality of the incoming data to filter out any abnormal or damaged data packets, thereby ensuring that the obtained data stream is complete. Through specific operations, ensure the accurate collection of real-time transaction data of financial products and generate an original real-time financial data stream.

[0074] Based on the original real-time financial data stream, add the exact timestamp of server reception and synchronize the time tag to each data to obtain a time-tagged real-time financial data stream;

[0075] During the execution of time synchronization, the operation of adding timestamps to each data needs to be accurate to milliseconds to ensure the time accuracy of subsequent data processing. First, calibrate the server's clock to ensure its consistency with international standard time. Subsequently, at the moment when the data arrives at the server, generate a timestamp through a high-precision clock and attach this timestamp to each data. The process involves the capture and marking of time, ensuring that each data can reflect its true reception time. The processed data stream then carries a time tag, which can provide an accurate time reference for subsequent time difference calculation and data analysis.

[0076] Using the time-tagged real-time financial data stream, adopt the formula:

[0077] Δt i =t i+1 -t i

[0078] Calculate the time difference value Δt between adjacent data points i , and generate time difference data. Among them, t i represents the timestamp of the i-th data point, and t i+1Represents the timestamp of the (i + 1)-th data point;

[0079] Collect t i = 1500000000.123 seconds and t i+1 = 1500000000.456 seconds, the calculated time difference is Δt i = 1500000000.456 - 1500000000.123 = 0.333 seconds. The result shows that the time interval between these two data points is 0.333 seconds, and the time difference will be used for subsequent analysis to identify possible time deviations or for further time correction.

[0080] Please refer to Figure 3 , the steps for obtaining synchronized time series data are as follows:

[0081] Using the time difference data, perform linear interpolation, using the formula:

[0082]

[0083] Calculate the new timestamp t of each data point new,i , to obtain the adjusted timestamp, where t i is the original timestamp, Δt i is the time difference value, g is the normalization constant, and n is the total number of data points;

[0084] In a certain dataset, t i = 1500000000.050, Δt i = 0.300, n = 100, g = 0.1, then the adjusted timestamp t new,i is calculated as follows: First, calculate the square root of each time difference divided by the sum of the time differences, that is Then multiply by the constant g, the result is 0.018257×0.1≈0.001826, and finally, after calculation, the adjusted timestamp is 1500000000.050 + 0.001826 = 1500000000.051826. This shows that the adjustment process has indeed reduced the time difference and improved the data synchronization accuracy.

[0085] Use the adjusted timestamp to perform frequency correction on the data stream to ensure that the data acquisition frequency matches the preset standard. Analyze the original data frequency through Fourier transform and apply a filter to adjust it to the target frequency to generate a frequency-corrected data stream;

[0086] In the detailed process of data frequency correction, first, the Fourier transform is applied to analyze the frequency characteristics of the original data stream in detail, converting the time series data into a frequency domain representation to identify the main frequency components in the data and any abnormal frequency noise. Based on the continuous fast Fourier transform algorithm, it is ensured that the frequency of each data point is accurately measured and recorded. Once the frequency analysis is completed, a customized digital filter is used to process the data stream. The filter is designed to cut or enhance specific frequency ranges, so that the frequency of the data stream is strictly consistent with the preset operating frequency. Through these steps, the data stream is adjusted to a unified frequency standard, improving the consistency and accuracy of data processing. Finally, the adjusted data is reconverted into a time series through the inverse Fourier transform to ensure that all operations maintain the integrity and validity of the data. The correction process is particularly important for subsequent data analysis because it directly affects the quality and effect of data analysis.

[0087] According to the data stream after frequency correction, time stamp calibration is performed. Time series analysis is applied to relabel the time stamps of each data point, and synchronized time series data is output.

[0088] During the execution of time stamp calibration, first, the data stream after frequency correction is utilized. Advanced time series analysis methods are applied to accurately calibrate the time stamps of each data point. The process involves complex time series modeling techniques, such as the ARIMA model or seasonal decomposition. Through the model, the time structure and pattern of the data can be analyzed in detail to ensure that the time mark of each data point accurately reflects its original acquisition time. Especially when dealing with large-scale or high-frequency data, the calibration process is particularly crucial because even a tiny time error can lead to significant deviations in the analysis results. In addition, time stamp calibration also includes the adjustment of the data time axis to reflect the order and time interval of actual events. This is usually achieved through interpolation and data alignment techniques to ensure that each time point in the dataset is in the correct time series position, so that the finally output time series data is not only synchronized in time but also more stable and reliable, providing a solid foundation for further data analysis.

[0089] Please refer to Figure 4 , and the steps for obtaining the mapping feature set are as follows:

[0090] Based on the synchronized time series data, multidimensional scaling transformation is performed to preliminarily locate the multidimensional space position of each time data point, generating a primary set of spatial feature vectors.

[0091] Starting from synchronous time series data, the multi-dimensional scaling (MDS) method is used to transform the data into a set of primary spatial feature vectors. The key lies in how to process and map the multi-dimensional data of the time series. The specific steps include first collecting and synchronizing all relevant time series data, which may come from different sensors or measurement devices. Ensuring the time consistency and integrity of the data is the top priority in this step; next, select a suitable multi-dimensional scaling algorithm, which usually depends on the dimension of the data and the expected complexity of the results; then, execute the algorithm to map the data at each time point into a multi-dimensional space. The position of each data point in the multi-dimensional space is represented by a set of coordinates, and this set of coordinates constitutes the primary spatial feature vector. In this process, the relative positions and distances between data points are preserved for further analysis and processing in subsequent steps. In addition, it is necessary to evaluate the quality of the preliminary mapping results, including checking whether the distribution of data points in the new space is reasonable, whether there are outliers or clustering phenomena, which is crucial for ensuring the accuracy of the next data processing. The generated set of primary spatial feature vectors will provide the basic data and reference framework for the next step of optimization and feature extraction.

[0092] Using the set of primary spatial feature vectors, the formula:

[0093]

[0094] Calculate the Euclidean distance d between vector i and vector j ij , to generate an optimized set of spatial feature vectors, where, v ik and v jk are the coordinate values of vectors i and j in the k-th dimension, and m represents the dimension of the coordinates;

[0095] In a three-dimensional space, the coordinates of vector i and vector j are (1, 2, 3) and (4, 5, 6) respectively, then:

[0096]

[0097] The result shows that the distance between vector i and vector j in the three-dimensional space is approximately 5.20 units. The value can help understand and evaluate the similarity or difference in attributes between two data points.

[0098] Based on the optimized set of spatial feature vectors, integrating the spatial attributes and relationships of the vectors, revealing the key spatial characteristics of the original data, and generating mapping features;

[0099] After obtaining the optimized spatial feature vector set, further synthesize the spatial attributes and relationships of the vectors and perform the extraction of key spatial features. The process details the contribution degree of each vector and its relative importance in the entire dataset, and uses statistical analysis and pattern recognition techniques to identify those representative feature vectors. First, normalize the optimized feature vectors to eliminate the influence of different dimensions; then, use the clustering analysis method to group the vectors according to the distance between them, and each group represents a potential pattern or trend in the data; then, select the most representative vector from each group, and the vector plays a key role in defining the structure and distribution characteristics of the dataset; finally, combine the selected vectors to form the final mapping feature set. The mapping feature set not only reflects the core structure of the original data, but also provides an important information source for subsequent data analysis and applications. In this way, the mapping feature set can effectively capture and represent the key spatial characteristics of the data, providing a basis for data interpretation and further data-driven decision-making.

[0100] Please refer to Figure 5 , the identification steps of key morphological changes are as follows:

[0101] According to the mapping feature set, extract the key time-intensity feature data, and use a sliding time window to generate windowed feature data;

[0102] The process of extracting key time-intensity feature data from the mapping feature set is achieved by screening and analyzing a large amount of time series data. First, according to the change characteristics of the time series, select representative data points that can reflect the key trends and fluctuations of the overall time series; next, use the sliding time window technology to divide the time series into multiple continuous window segments according to the preset time window length. Each window segment contains a fixed number of data points, which can capture the local characteristics of the data within each window while maintaining the continuity of the time series; perform statistical analysis on the data within each time window, calculate statistical indicators such as its average value, standard deviation, and extreme value, which helps to extract more accurate and useful feature information from the original time series data, and finally generate windowed feature data.

[0103] Perform principal component analysis on the windowed feature data, compress the data dimension, extract key information, and generate dimensionality-reduced feature data;

[0104] Perform principal component analysis on the windowed feature data, with the aim of reducing redundant information in the data through mathematical transformation, improving the efficiency and accuracy of data analysis. In the specific operation process, first calculate the covariance matrix between the features to understand the correlation between different features; next, based on the covariance matrix, extract the eigenvalues and corresponding eigenvectors, where the eigenvectors represent the main change directions of the data; then, according to the magnitudes of the eigenvalues, select the principal components with high contribution degrees, which can retain the most important information while reducing the data dimension; in this way, the original high-dimensional data is transformed into low-dimensional but key information-containing feature data, and this dimension-reduced data is more suitable for complex pattern analysis and classification processing, and finally generates dimension-reduced feature data to provide input data for the next deep learning model.

[0105] Input the dimension-reduced feature data into the deep learning model to identify and classify the key morphological changes of the time series, using the formula:

[0106]

[0107] Calculate the pattern probability p(t) of the time series to generate the identification result of key morphological changes, where α h is the influence coefficient of the h-th pattern, γ h represents the decay parameter corresponding to the pattern, μ h is the central time point of the corresponding pattern, and M represents the number of patterns;

[0108] Assume M = 3, α1 = 0.5, α2 = 0.3, α3 = 0.2, γ1 = 1.0, γ2 = 0.8, γ3 = 1.2, μ1 = 10, μ2 = 20, μ3 = 15, and calculate the value of p(12) when t = 12: Calculate to get p(12) = 0.5×e (-4) +0.3×e (-51.2) +0.2×e (-10.8) ≈0.5×0.0183 + 0.3×1.938×10 -23 +0.2×0.00002 ≈ 0.00915 + 0 + 0 = 0.00915. The result shows that at the moment t = 12, the pattern probability of the time series is 0.00915, indicating that the key morphological changes at this moment are relatively low and it is less likely to have significant events or features.

[0109] The deep learning model is specifically a long short-term memory network (LSTM): LSTM is a special recurrent neural network (RNN), suitable for processing time series data, capable of capturing long-term dependencies, and can be used to learn the key patterns of time series and predict the probability of future key state changes based on the input sequence.

[0110] Use a large amount of time series data and its annotation information (such as the occurrence of critical moments) to train the model so that it learns the mapping relationship from the original time series data to the critical morphological changes.

[0111] Use probability for classification:

[0112] The trained model can input new time series data, calculate the probability p(t) of critical state changes, and then determine whether the time point belongs to critical morphological changes according to a preset threshold, thereby completing the judgment process.

[0113] Please refer to Figure 6 , and the steps for obtaining the behavior recognition result are:

[0114] Utilize the critical morphological change recognition result to monitor new time series data in real time, compare it with historical trend data, conduct a difference analysis, and generate a real-time monitoring and comparison result;

[0115] Utilize the recognized critical morphological changes to monitor new time series data in real time and compare it with the archived historical trend data. The process involves extracting relevant time series data from the historical data storage system. The data is usually arranged in chronological order and contains past performance data points and possible trend changes. Then, use data mining techniques, especially time series analysis techniques, to conduct a detailed comparison and analysis of the new data and historical data. The difference analysis mainly calculates the statistical differences between the new and old data points, such as changes in statistical indicators such as the mean and standard deviation, so as to determine the similarity or difference between the data points. The finally generated real-time monitoring report will describe these differences in detail, providing a basis for subsequent data processing and decision-making.

[0116] Based on the real-time monitoring and comparison result, extract data points and use the formula,

[0117]

[0118] Calculate the similarity coefficient cs, conduct a similarity evaluation, and generate a data point similarity evaluation result, where A and B respectively represent the feature vectors of the current data point and the historical data point, and ‖A‖ and ‖B‖ respectively represent the norms;

[0119] Suppose A = (1, 2, 3) and B = (2, 3, 4).

[0120] Then And A·B = 1×2 + 2×3 + 3×4 = 20. Therefore, the similarity coefficient is calculated as: This result indicates that the current data point and the historical data point have a high similarity in the direction of the feature vector (0.993 is very close to 1), indicating that they may belong to similar data patterns or trends.

[0121] Based on the evaluation results of data point similarity, perform threshold determination to analyze the abnormality degree of data points, identify data points that significantly deviate from the normal pattern, reveal potential abnormal behaviors in financial data, and generate behavior recognition results;

[0122] Based on the evaluation results of data point similarity, apply the threshold determination method to analyze the abnormality degree of data points. It depends on the comparison between the similarity results and the preset threshold. By setting a threshold, for example, setting the threshold to 0.5, compare the similarity scores of each data point. If the similarity score is lower than 0.5, it is considered that the difference between this data point and the historical data is significant, indicating an abnormality or a new trend. This method is simple but very effective. It can quickly screen out the data points that need further analysis, identify the data points that significantly deviate from the normal pattern, further reveal potential abnormal behaviors in financial data, and finally generate behavior recognition results, providing strong data support for decision-makers.

[0123] Please refer to Figure 7 , the evaluation steps for the risk level are as follows:

[0124] Utilize the behavior recognition results to extract the key time points in the behavior data, compare them with the set benchmark risk threshold, and generate the comparison results of the behavior characteristics for each time point;

[0125] Through the behavior recognition results, extract the key time points in the behavior data, compare them with the set benchmark risk threshold, analyze the differences between each time point and the benchmark risk threshold, and evaluate the comparison results of the behavior characteristics. The key parameters in the process include the time point characteristic data and the benchmark risk threshold. The time point characteristic data is automatically extracted by the behavior recognition system and compared with the preset risk threshold in the database. The comparison algorithm uses a simple difference calculation, that is, the difference between the behavior characteristic value of each data point and the risk threshold, which can quickly identify the behaviors that deviate from the normal range, thereby initially identifying potential risk points and generating the comparison results of the behavior characteristics for each time point.

[0126] Analyze the comparison results of the behavior characteristics for each time point, using the formula:

[0127]

[0128] Calculate the risk index R(t) for each time point to obtain the preliminary risk assessment result. Among them, ΔB(t) represents the deviation between the behavior data and the benchmark threshold, and B base is the benchmark behavior threshold;

[0129] At time point t, the behavior data value is 120, while the benchmark threshold is 100, then ΔB(t) is 20. If the benchmark threshold B base is 100, then the risk index R(t) is The result shows that the risk index at time point t is 0.2, indicating the degree of deviation between the behavioral data at this time point and the benchmark threshold. Based on the degree of deviation, a further risk level assessment of the behavior at time point t can be carried out.

[0130] According to the preliminary risk assessment results, classify the risk levels for each time point that deviates from the benchmark threshold, determine the risk level of each time point, and generate the risk level assessment results.

[0131] According to the preliminary risk assessment results, identify high-risk behaviors and mark them as abnormal. In this process, the key steps include comparing the risk level of each time point with the set risk level threshold. The risk level classification uses a simple classification model, and the model parameters include the risk level threshold, which is set according to historical data and the guidance of the risk management department. The risk index of each time point is compared with these thresholds. If the risk index exceeds the threshold, it is marked as abnormal, thereby determining the risk level of each time point and generating the final risk level assessment results.

[0132] Please refer to Figure 8 , and the steps to obtain the financial risk early warning record are as follows:

[0133] Analyze the risk level assessment results, identify abnormal behavior events, and generate a list of potential risk threats.

[0134] By analyzing the risk level assessment of the previous key item, identify the behavior events of high-risk indicators. This process involves in-depth analysis of the risk assessment results. First, according to the preset risk classification model, screen out the behaviors with high-risk indicators. For example, if the behavioral data shows that the behavioral indicators at a certain time point far exceed those at other time points, this behavior will be classified as a high-risk behavior. This classification is based on data analysis and predefined risk thresholds. Then, further screen these behaviors to determine which behaviors may lead to serious negative consequences, such as financial losses or safety accidents. Each behavior event will be detailedly recorded and the risk level will be determined. According to their risk levels, these behaviors will be added to the list of potential risk threats, providing a basis for subsequent risk management measures.

[0135] Based on the list of potential risk threats, classify according to the risk degree, mark the abnormal events as sensitive events, and use the formula:

[0136]

[0137] Calculate the percentage S of sensitive events in the total events to obtain the risk degree classification result, where R high is the number of abnormal events, and R total is the total number of events;

[0138] During a certain period of time, a total of 1000 events were identified, among which 100 were high-risk events. The percentage of sensitive events is calculated as follows: This result indicates that 10% of the events are marked as high-risk, which is an important indicator of the current security management status. It reflects the proportion of sensitive events among all identified events, helps decision-makers understand the prevalence of high-risk events, and accordingly adjust risk control strategies.

[0139] According to the classification results of the risk level, based on the characteristics and historical data analysis of sensitive events, output financial risk warning records;

[0140] According to the classification results of the risk level, mark and record sensitive events, which involves marking and recording the classified events. First, the system will classify the events according to their risk levels, and specifically mark the events classified as high-risk or sensitive. The mark reflects the potential threat level that the event may pose to the organization. Each marked event will be detailedly recorded in the financial risk warning system. The system outputs corresponding financial risk warning records based on the characteristics of these events and the comparative analysis with historical data, providing necessary information for management to formulate effective response measures and prevention strategies to ensure that the organization's risks are under control.

[0141] The above is only the preferred embodiment of the present invention and does not limit the present invention in other forms. Any person skilled in the art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A big data driven financial risk identification system, characterized in that: The system comprises: The time data synchronization module collects real-time financial data streams, synchronizes time tags, calculates the time difference of each data point, performs dynamic time adjustment based on the calculation results, performs frequency correction and time tag calibration on the data, and outputs synchronized time series data; The multidimensional feature mapping module converts the time data into a spatial feature vector through multidimensional scaling based on the synchronized time series data, performs dimension mapping on the data at each time point, extracts key spatial features, and generates a mapping feature set; The financial behavior analysis module uses feature analysis to extract sequence patterns based on the mapping feature set, identifies key morphological changes in the time series, monitors and compares historical trend data in real time based on the recognition results, calculates the similarity between data points, reveals abnormal behavior of financial data, and outputs behavior recognition results; Based on the behavior identification results, the risk dynamic identification module evaluates the risk level of abnormal behavior by comparing the risk thresholds at different time points, determines potential risk threats based on the evaluation results, classifies the risk levels and marks sensitive events, and outputs financial risk warning records.

2. The big data driven financial risk identification system according to claim 1, characterized in that: The steps for calculating the time difference are: Perform data collection initialization processing, establish a secure connection with the financial market data interface, batch pull real-time transaction data of different financial products, and generate original real-time financial data streams; Based on the original real-time financial data stream, add an exact timestamp of the server receiving each data, synchronize the time label, and obtain a real-time financial data stream with a time label; Using the real-time financial data stream with time tags, the formula is adopted: Δt i =t i+1 ―t i Calculate the time difference Δt between adjacent data points i , generate time difference data, where t i Represents the timestamp of the i-th data point, t i+1 Represents the timestamp of the i+1th data point.

3. The big data driven financial risk identification system according to claim 2, characterized in that: The steps for acquiring the synchronized time series data are as follows: The time difference data is used to perform linear interpolation using the formula: Calculate a new timestamp t for each data point new,i , get the adjusted timestamp, where t i is the original timestamp, Δt i is the time difference value, g is the normalization constant, and n is the total number of data points; Performing frequency correction on the data stream using the adjusted timestamp to ensure that the data acquisition frequency matches a preset standard, analyzing the original data frequency by Fourier transform and applying a filter to adjust to a target frequency, and generating a frequency-corrected data stream; A time stamp calibration is performed based on the frequency-corrected data stream, and a time stamp of each data point is re-marked by applying time series analysis to output synchronized time series data.

4. The big data driven financial risk identification system according to claim 3 is characterized in that: The steps of obtaining the mapping feature set are: Based on the synchronized time series data, a multidimensional scaling transformation is performed to preliminarily locate the multidimensional spatial position of each time data point, and a primary spatial feature vector set is generated; Using the primary spatial feature vector set, the formula is adopted: Calculate the Euclidean distance d between vector i and vector j ij , generate the optimized spatial feature vector set, where v ik and v jk are the coordinate values ​​of vectors i and j in the kth dimension, and m represents the coordinate dimension; According to the optimized spatial feature vector set, the spatial attributes and relationships of the vectors are integrated to reveal the key spatial characteristics of the original data and generate mapping features.

5. The big data driven financial risk identification system according to claim 4, characterized in that: The steps for identifying the key morphological changes are: Extracting key time-intensity feature data according to the mapping feature set, and generating windowed feature data using a sliding time window; Performing principal component analysis on the windowed feature data, compressing data dimensions, extracting key information, and generating reduced-dimensional feature data; The reduced dimension feature data is input into the deep learning model to identify and classify the key morphological changes of the time series using the formula: Calculate the pattern probability p(t) of the time series and generate the key morphological change recognition results, where α h is the influence coefficient of the hth mode, γ h represents the attenuation parameter corresponding to the corresponding mode, μ h is the central time point of the corresponding mode, and M represents the number of modes.

6. The big data driven financial risk identification system according to claim 5, characterized in that: The steps for obtaining the behavior recognition result are: Using the key morphological change identification results, new time series data is monitored in real time, compared with historical trend data, and difference analysis is performed to generate real-time monitoring and comparison results; Based on the real-time monitoring and comparison results, data points are extracted and the formula is used. Calculate the similarity coefficient cs, perform similarity evaluation, and generate the data point similarity evaluation result, where A and B represent the feature vectors of the current data point and the historical data point respectively, and ‖A‖ and ‖B‖ represent the modulus length respectively; Based on the data point similarity evaluation results, a threshold determination is performed to analyze the abnormality of the data points, identify data points that significantly deviate from the normal pattern, reveal potential abnormal financial data behavior, and generate behavior recognition results.

7. The big data driven financial risk identification system according to claim 6, characterized in that: The steps for assessing the risk level are: Utilizing the behavior recognition results, extracting key time points in the behavior data, comparing them with the set baseline risk threshold, and generating behavior feature comparison results at each time point; Analyze the behavioral characteristics comparison results at each time point, using the formula: Calculate the risk index R(t) at each time point to obtain the preliminary risk assessment results, where ΔB(t) represents the deviation of the behavior data from the baseline threshold, and B base is the baseline behavioral threshold; According to the preliminary risk assessment results, each time point that deviates from the benchmark threshold is classified into a risk level, the risk level of each time point is determined, and a risk level assessment result is generated.

8. The big data driven financial risk identification system according to claim 7, characterized in that: The steps for obtaining the financial risk warning record are: Analyze the risk level assessment results, identify abnormal behavior events, and generate a list of potential risk threats; Based on the potential risk threat list, classify according to the risk level and mark abnormal events as sensitive events using the formula: Calculate the percentage S of sensitive events in total events and get the risk level classification result, where R high is the number of abnormal events, R total is the total number of events; According to the risk level classification results, based on the characteristics of sensitive events and historical data analysis, financial risk warning records are output.

Images