Financial risk early warning system and method based on big data

Abstract

The invention relates to the technical field of financial risks, in particular to a financial risk early warning system and method based on big data, and the system comprises an information collection module which is used for collecting financial data content information, modification information, access authority information and modification user information; the risk identification module determines risk financial data according to the sensitive function modification mode or the modification position node of the financial data; the risk judgment module is used for calculating the risk type tendency value of the risk financial data according to the modification concealment index of the risk financial data and the data pollution diffusion entropy, and further determining the risk tendency type of the risk financial data; and the risk early warning module determines to send out direct early warning or indirect early warning to the risk financial data according to the risk tendency type of the risk financial data and / or the risk transition coefficient of the risk financial data, and through the method, the data abnormal risk identification accuracy of the financial product early warning system is improved, and then accurate grading early warning is realized.

Description

Technical Field

[0001] The present invention relates to the field of financial risk technology, and in particular to a financial risk early warning system and method based on big data. Background Art

[0002] The financial industry is facing increasingly complex and diverse risk challenges in the process of digital development. With the continuous expansion of financial business and the explosive growth of data volume, traditional risk warning methods are unable to effectively deal with the various risks hidden in massive financial data. The development of big data technology has provided new means and ideas for financial risk warning. By utilizing big data technology, financial data can be collected and analyzed comprehensively and in real time, and potential risk characteristics and patterns in the data can be mined. At the same time, combined with advanced algorithms and models, it can more accurately identify, judge and warn of financial risks, improve the accuracy and timeliness of risk warnings, help financial institutions better prevent and respond to various financial risks, and ensure the stable operation of the financial system.

[0003] For example, Chinese patent application publication number CN119130660A discloses a financial risk early warning system and method based on big data, comprising: a processing module for obtaining the current profitability, asset quality, and operating status values of a financial institution based on all profit parameters, asset parameters, and operating parameters of the financial institution at the current moment; a second judgment module for obtaining a second risk judgment result of the financial institution at the current moment based on the current profitability, asset quality, and operating status values of the financial institution; and an implicit early warning module for obtaining a financial risk early warning result of the financial institution at the current moment based on the second risk judgment result of the financial institution at the current moment. This invention enables a holistic assessment of the financial health of a financial institution at the current moment and provides timely financial risk early warnings for the financial institution.

[0004] However, existing technologies have the problem that the financial product early warning system has poor accuracy in identifying data anomaly risks, resulting in the inability to accurately grade early warnings. Summary of the Invention

[0005] To this end, the present invention provides a financial risk early warning system and method based on big data to overcome the problem that the financial product early warning system in the prior art has poor accuracy in identifying data abnormality risks, resulting in the inability to accurately grade early warnings.

[0006] To achieve the above objectives, the present invention provides a financial risk early warning system and method based on big data, comprising:

[0007] An information collection module, which is used to collect financial data content information, financial data modification information, financial data access permission information, and financial data modification user information;

[0008] a risk identification module connected to the information collection module, configured to determine risky financial data based on whether a sensitive function modification pattern of the financial data is a normal modification pattern or whether a modification position node of the financial data has an abnormal correlation;

[0009] a risk judgment module, connected to the risk identification module, configured to calculate a risk type tendency value of the risk financial data based on a modification concealment index and a data contamination diffusion entropy of the risk financial data, and determine a risk tendency type of the risk financial data based on a comparison result of the risk type tendency value with a preset risk type tendency value;

[0010] a risk warning module, connected to the risk judgment module, for determining whether to issue a direct warning or an indirect warning for the risk financial data based on the risk tendency type of the risk financial data and / or the risk transition coefficient of the risk financial data;

[0011] An adjustment module is respectively connected to the risk identification module, the risk judgment module and the risk warning module, and is used to determine the adjustment of the risk identification judgment parameters and / or the preset risk type tendency value based on the risk adaptive balance coefficient and adjustment sensitivity within a preset period.

[0012] Furthermore, the risk identification module determines risky financial data based on whether the sensitive function modification mode of the financial data is a normal modification mode or whether there is an abnormal correlation between the modification position nodes of the financial data; wherein,

[0013] If the sensitive function modification mode of the financial data is an abnormal modification mode or there is an abnormal correlation between the modification position nodes of the financial data, the financial data is determined to be risky financial data.

[0014] Furthermore, whether the sensitive function modification mode of the financial data is a normal modification mode is determined based on the matching degree of the modification behavior of the operating subject, and whether there is an abnormal correlation between the modification position node of the financial data is determined based on whether there is a modification trend with the same abnormal characteristics between the modification position node and its associated position node.

[0015] Furthermore, the risk judgment module calculates the risk type tendency value of the risk financial data based on the modification concealment index and data pollution diffusion entropy of the risk financial data, and determines the risk tendency type of the risk financial data based on the risk type tendency value; wherein,

[0016] If the risk type tendency value is greater than the preset risk type tendency value, the risk type is determined to be a strong risk type;

[0017] If the risk type tendency value is less than or equal to the preset risk type tendency value, the risk type is determined to be a weak risk type.

[0018] Furthermore, the preset risk type propensity value is determined based on the historical average risk type propensity value of the same type of financial data.

[0019] Furthermore, the risk warning module determines whether to issue a direct warning or an indirect warning for the risk financial data based on the risk tendency type of the risk financial data and / or the risk transition coefficient of the risk financial data; wherein,

[0020] If the risk tendency type of the risky financial data is a strong risk type or the risk transition coefficient of the risky financial data is greater than a preset transition coefficient, determining to issue a direct warning for the risky financial data;

[0021] If the risk tendency type of the risky financial data is a weak risk type and the risk transition coefficient of the risky financial data is less than or equal to a preset transition coefficient, it is determined to issue an indirect warning for the risky financial data.

[0022] Furthermore, the risk transition coefficient of the risk financial data is determined based on the probability that the risk financial data in the historical data changes from one risk state to another risk state under different conditions, the preset transition coefficient is determined based on the average value of the risk transition coefficients of the risk financial data in the historical data, and the preset risk type tendency value is determined based on the historical average risk type tendency value of the same type of financial data.

[0023] Furthermore, the adjustment module determines to adjust the preset matching degree or the preset risk type tendency value based on the risk adaptive balance coefficient and the adjustment sensitivity within the preset period; wherein,

[0024] If the risk adaptive balance coefficient within the preset period is greater than the preset balance coefficient and the adjustment sensitivity is less than the preset sensitivity, it is determined to adjust the preset matching degree;

[0025] If the risk adaptive balance coefficient within the preset period is less than or equal to the preset balance coefficient and the adjustment sensitivity is greater than or equal to the preset sensitivity, it is determined to adjust the preset risk type propensity value;

[0026] If the risk adaptive balance coefficient within the preset period is greater than the preset balance coefficient and the adjustment sensitivity is greater than or equal to the preset sensitivity, it is determined to adjust the preset matching degree and the preset risk type propensity value.

[0027] Furthermore, the risk adaptive balance coefficient is determined based on the risk false alarm rate, the preset period is determined based on the historical average value of the same type of financial data in a stable market environment, the sensitivity is determined based on the adjustment amplitude during several adjustments within the preset period, and the preset sensitivity is determined based on the average value of the historical sensitivity of the same type of financial data.

[0028] A financial risk early warning method applied to a financial risk early warning system based on big data, comprising:

[0029] Step S1, obtaining financial data content information, financial data modification information, financial data access permission information, and financial data modification user information;

[0030] Step S2: determining whether the financial data is risky financial data based on whether the sensitive function modification mode of the financial data is a normal modification mode or whether there is an abnormal correlation between the modification position nodes of the financial data;

[0031] Step S3, calculating a risk type tendency value based on the modification concealment index and data contamination diffusion entropy of the risky financial data, and comparing the risk type tendency value with a preset risk type tendency value to determine the risk tendency type of the risky financial data;

[0032] Step S4, determining whether to issue a direct warning or an indirect warning for the risky financial data based on the risk tendency type and / or risk transition coefficient of the risky financial data;

[0033] Step S5: Dynamically adjust the preset matching degree and the preset risk type tendency value based on the risk adaptive balance coefficient and the adjustment sensitivity.

[0034] Compared with the prior art, the beneficial effect of the present invention is that the present invention determines risky financial data by whether the sensitive function modification mode of the financial data is a normal modification mode or whether there is an abnormal correlation between the modification position nodes of the financial data. According to the fact that the sensitive function modification mode of the financial data is an abnormal modification mode or there is an abnormal correlation between the modification position nodes of the financial data, it indicates that there may be abnormal changes in the financial data, potentially indicating financial risk phenomena, and determines risky financial data. According to the fact that the sensitive function modification mode of the financial data is a normal modification mode or there is no abnormal correlation between the modification position nodes of the financial data, it indicates that the financial data has no risk and is accurately determined to be not financial data. The above method improves the accuracy of the financial product early warning system in identifying abnormal data risks, thereby achieving accurate graded early warnings.

[0035] Furthermore, the present invention calculates the risk type tendency value of the risk financial data through the modification concealment index of the risk financial data and the data pollution diffusion entropy, and determines the risk tendency type of the risk financial data based on the risk type tendency value. According to the risk type tendency value being greater than the preset risk type tendency value, it indicates that the financial data has a higher risk of abnormal changes, which may have a greater impact on financial business, and the risk type is accurately determined to be a strong risk type; according to the risk type tendency value being less than or equal to the preset risk type tendency value, it indicates that the risk of abnormal changes in the financial data is relatively low, and the impact on financial business is small, and the risk type is accurately determined to be a weak risk type. The above method improves the accuracy of the financial product early warning system in identifying data abnormality risks, thereby achieving accurate graded early warning.

[0036] Furthermore, the present invention determines whether to issue a direct warning or an indirect warning to the risk financial data based on the risk tendency type of the risk financial data and / or the risk transition coefficient of the risk financial data. If the risk tendency type of the risk financial data is a strong risk type and the risk transition coefficient of the risk financial data is greater than a preset transition coefficient, it means that the financial data not only has a high risk tendency itself, but also has a high possibility of its risk state changing. The direct warning to the risk financial data is accurately determined. If the risk tendency type of the risk financial data is a weak risk type and the risk transition coefficient of the risk financial data is less than or equal to the preset transition coefficient, it means that the risk tendency of the financial data is low, and the possibility of its risk state changing is small, and the potential impact on financial business is relatively small. The risk financial data is accurately determined. An indirect warning is issued. According to the fact that the risk tendency type of the risk financial data is a strong risk type and the risk transition coefficient of the risk financial data is less than or equal to the preset transition coefficient, it means that although the financial data itself has a high risk tendency, the possibility of its risk status changing is relatively small at present, and the impact on financial business in the short term may be limited. It is determined not to issue a warning. According to the fact that the risk tendency type of the risk financial data is a weak risk type and the risk transition coefficient of the risk financial data is greater than the preset transition coefficient, it means that the risk tendency of the financial data itself is low, but the possibility of risk status changing is high. However, due to the low risk tendency itself, it is comprehensively judged that the current impact on financial business is not significant, and it is accurately determined not to issue a warning. The above method improves the accuracy of the financial product early warning system in identifying abnormal data risks, thereby achieving accurate graded early warnings.

[0037] Furthermore, the present invention determines to adjust the preset matching degree or the preset risk type tendency value through the risk adaptive balance coefficient and / or adjustment sensitivity within the preset period. According to the risk adaptive balance coefficient within the preset period being greater than the preset balance coefficient and the adjustment sensitivity being less than the preset sensitivity, it indicates that the risk false alarm rate of the current system is relatively high, and the system is not sensitive enough to the risk change. It accurately determines to adjust the preset matching degree. According to the risk adaptive balance coefficient within the preset period being less than or equal to the preset balance coefficient and the adjustment sensitivity being greater than or equal to the preset sensitivity, it indicates that the risk false alarm rate of the previous system is relatively low, but the system is too sensitive to the risk change, which may lead to Frequent adjustments are made to accurately determine the adjustment of the preset risk type tendency value. According to the risk adaptive balance coefficient within the preset period being greater than the preset balance coefficient and the adjustment sensitivity being greater than or equal to the preset sensitivity, it indicates that the current system has a high risk false alarm rate. At the same time, the system's response to risk changes is too sensitive. Accurately determine the adjustment of the normal mode and the preset risk type tendency value. According to the risk adaptive balance coefficient within the preset period being less than or equal to the preset balance coefficient and the adjustment sensitivity being less than the preset sensitivity, it indicates that the system is running stably and accurately determine that no adjustment is required. The above method improves the accuracy of the financial product early warning system in identifying abnormal data risks, thereby achieving accurate graded early warnings. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 This is a schematic diagram of the structure of a financial risk early warning system based on big data according to an embodiment of the present invention;

[0039] Figure 2 This is a workflow diagram of the risk identification module of the financial risk early warning system based on big data according to an embodiment of the present invention;

[0040] Figure 3 This is a workflow diagram of the risk judgment module of the financial risk early warning system based on big data according to an embodiment of the present invention;

[0041] Figure 4 This is a workflow diagram of the risk warning module of the financial risk warning system based on big data in an embodiment of the present invention. DETAILED DESCRIPTION

[0042] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention.

[0043] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0044] Furthermore, it should be noted that, in the description of the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0045] See also Figure 1-Figure 4 As shown, Figure 1 This is a schematic diagram of the structure of a financial risk early warning system based on big data according to an embodiment of the present invention; Figure 2 This is a workflow diagram of the risk identification module of the financial risk early warning system based on big data according to an embodiment of the present invention; Figure 3 This is a workflow diagram of the risk judgment module of the financial risk early warning system based on big data according to an embodiment of the present invention; Figure 4 This is a workflow diagram of the risk warning module of the financial risk warning system based on big data in an embodiment of the present invention.

[0046] The financial risk early warning system and method based on big data in the embodiment of the present invention include:

[0047] An information collection module, which is used to collect financial data content information, financial data modification information, financial data access permission information, and financial data modification user information;

[0048] a risk identification module connected to the information collection module, configured to determine risky financial data based on whether a sensitive function modification pattern of the financial data is a normal modification pattern or whether a modification position node of the financial data has an abnormal correlation;

[0049] a risk judgment module, connected to the risk identification module, configured to calculate a risk type tendency value of the risk financial data based on a modification concealment index and a data contamination diffusion entropy of the risk financial data, and determine a risk tendency type of the risk financial data based on a comparison result of the risk type tendency value with a preset risk type tendency value;

[0050] a risk warning module, connected to the risk judgment module, for determining whether to issue a direct warning or an indirect warning for the risk financial data based on the risk tendency type of the risk financial data and / or the risk transition coefficient of the risk financial data;

[0051] An adjustment module is respectively connected to the risk identification module, the risk judgment module and the risk warning module, and is used to determine the adjustment of the risk identification judgment parameters and / or the preset risk type tendency value based on the risk adaptive balance coefficient and adjustment sensitivity within a preset period.

[0052] In the embodiment of the present invention, the financial data content information includes but is not limited to "balance sheet data, transaction contract terms and investment portfolio details", the financial data modification information includes but is not limited to "modification timestamp, comparison of data before and after modification and modification operation type", the financial data access permission information includes but is not limited to "user permission level, accessible data range and permission effective time", and the financial data modification user information includes but is not limited to "user identity ID, department and operating device IP address".

[0053] Specifically, the risk identification module determines the risky financial data based on whether the sensitive function modification mode of the financial data is a normal modification mode or whether there is an abnormal correlation between the modification position nodes of the financial data under the condition of determining the risky financial data;

[0054] If the sensitive function modification mode of the financial data is an abnormal modification mode or there is an abnormal correlation between the modification position nodes of the financial data, the risk identification module determines that the financial data is risky financial data;

[0055] If the sensitive function modification mode of the financial data is a normal modification mode or there is no abnormal correlation between the modification position nodes of the financial data, the risk identification module determines that the financial data is normal financial data.

[0056] In the embodiment of the present invention, whether the sensitive function modification mode of the financial data is a normal modification mode is determined based on the comparison result of the modification behavior matching degree of the operating subject and the preset matching degree (when the modification behavior matching degree is less than the preset matching degree, the sensitive function modification mode of the financial data is determined to be an abnormal modification mode). For example, the average number of modifications per day in the history of a user is 5 times, and the number of fields involved in a single modification does not exceed 3, but the number of modifications on the current operation day is 9 times, and the number matching degree is 9 minus 5 divided by 5 multiplied by the weight of 0.5 to obtain 0.4, and 6 sensitive fields are modified, and the field matching degree is 6 minus 3 divided by 3 multiplied by the weight of 0.5 to obtain 0.5, and the modification behavior matching degree is 0.4 multiplied by 0.5 to obtain 0.25, which is lower than the preset matching degree and is determined to be abnormal. The preset matching degree range is set to 0.8-0.9, and the preferred value is 0.85. The reason for selecting this preferred value is that setting the preset matching degree too high may cause normal business fluctuations to be misjudged as risks, thereby increasing the false alarm rate. Setting the preset matching degree too low may make the system insensitive to covert attacks, thereby reducing Low warning accuracy. Whether the modified location node of the financial data has an abnormal correlation is determined based on whether there is a modification trend of the same abnormal feature between the modified location node and its associated location nodes. The modification trend of the same feature is determined based on the logical correlation between the contents of multiple modification operations and conforms to the known modification pattern. The presence of a modification trend of the same abnormal feature between the modified location node and its associated location nodes determines that the modified location node of the financial data has an abnormal correlation. For example, in a certain bank system, the anti-fraud department monitored the following operations: within 10 minutes, 20 different customer accounts (all bound to the same IP address) successively performed the chain operations of "modifying reserved mobile phone numbers, increasing transfer limits, and adding unfamiliar receiving accounts", and the newly added receiving accounts were all the same remote accounts. This indicates that the presence of a modification trend of the same abnormal feature between the modified location node and its associated location nodes determines that the modified location node of the financial data has an abnormal correlation. However, the above value is not limited to this, and those skilled in the art can also adjust the value according to actual needs.

[0057] The present invention determines risky financial data by whether the sensitive function modification mode of the financial data is a normal modification mode or whether there is an abnormal correlation between the modification position nodes of the financial data. According to the fact that the sensitive function modification mode of the financial data is an abnormal modification mode or there is an abnormal correlation between the modification position nodes of the financial data, it indicates that there may be abnormal changes in the financial data, potentially indicating financial risk phenomena, and determines risky financial data. According to the fact that the sensitive function modification mode of the financial data is a normal modification mode or there is no abnormal correlation between the modification position nodes of the financial data, it indicates that the financial data has no risk and is accurately determined to be not financial data. The above method improves the accuracy of the financial product early warning system in identifying abnormal data risks, thereby achieving accurate graded early warnings.

[0058] Specifically, the risk judgment module calculates the risk type tendency value of the risk financial data according to the modification concealment index and data pollution diffusion entropy of the risk financial data under the condition of determining the risk tendency type of the risk financial data, and determines the risk tendency type of the risk financial data based on the risk type tendency value;

[0059] If the risk type tendency value is greater than a preset risk type tendency value, the risk judgment module determines that the risk type of the financial data is a strong risk type;

[0060] If the risk type tendency value is less than or equal to a preset risk type tendency value, the risk judgment module determines that the risk type of the financial data is a weak risk type.

[0061] In the embodiments of the present invention, the risk type propensity value of the risky financial data is determined based on the modification concealment index and data contamination diffusion entropy of the risky financial data. The risk type propensity value of the risky financial data is calculated as: modification concealment index × data contamination diffusion entropy × 2. The modification concealment index is obtained by dividing the standard deviation between the current modification time and the historical operation time by the total range of the historical operation time. The data contamination diffusion entropy is determined by multiplying the number of affected data by the duration of the impact, divided by 10. For example, a set of risky financial data has a modification concealment index of 0.8 and a data contamination diffusion entropy of 0.6. The risk type propensity value is set as: modification concealment index × data contamination diffusion entropy × 2, resulting in a risk type propensity value of 0.8 × 0.6 × 2 = 0.96. The preset risk type propensity value is determined based on the historical average risk type propensity value of financial data of the same type. For example, if the historical averages of two sets of financial data of the same type are 0.8 and 0.9, respectively, the preset risk type propensity value is determined to be 0.85. However, the above values are not limited to this, and those skilled in the art may also adjust the values according to actual needs.

[0062] The present invention calculates the risk type tendency value of the risk financial data through the modification concealment index of the risk financial data and the data pollution diffusion entropy, and determines the risk tendency type of the risk financial data based on the risk type tendency value. According to the risk type tendency value being greater than the preset risk type tendency value, it indicates that the financial data has a higher risk of abnormal changes, which may have a greater impact on financial business, and the risk type is accurately determined to be a strong risk type; according to the risk type tendency value being less than or equal to the preset risk type tendency value, it indicates that the risk of abnormal changes in the financial data is relatively low, and the impact on financial business is small, and the risk type is accurately determined to be a weak risk type. The above method improves the accuracy of the financial product early warning system in identifying data abnormality risks, thereby achieving accurate graded early warning.

[0063] Specifically, the risk warning module determines whether to issue a direct warning or an indirect warning for the risky financial data based on the risk tendency type of the risky financial data and / or the risk transition coefficient of the risky financial data under the condition of determining whether to issue a direct warning or an indirect warning for the risky financial data;

[0064] If the risk tendency type of the risky financial data is a strong risk type and the risk transition coefficient of the risky financial data is greater than a preset transition coefficient, the risk warning module determines to issue a direct warning for the risky financial data;

[0065] If the risk tendency type of the risky financial data is a weak risk type and the risk transition coefficient of the risky financial data is less than or equal to a preset transition coefficient, the risk warning module determines to issue an indirect warning for the risky financial data;

[0066] If the risk tendency type of the risk financial data is a strong risk type and the risk transition coefficient of the risk financial data is less than or equal to the preset transition coefficient, the risk warning module determines not to issue a warning;

[0067] If the risk tendency type of the risky financial data is a weak risk type and the risk transition coefficient of the risky financial data is greater than a preset transition coefficient, the risk warning module determines not to issue a warning.

[0068] The direct warning described in the embodiment of the present invention provides immediate warnings to the risk responsible party through a multi-channel collaborative mechanism, including but not limited to at least two of system pop-ups, SMS notifications, voice calls, and physical alarms; the indirect warning prompts the risk monitoring personnel through the risk monitoring system, including but not limited to at least one of risk dashboard marking, automatic work order generation, and regular report push.

[0069] The range of the risk transition coefficient of the risk financial data in the embodiment of the present invention is set to 0.3-0.5, and the preferred value is 0.4. The reason for selecting this preferred value is that the transition coefficient threshold is too low (such as 0.3), resulting in normal fluctuations being warned; the threshold is too high (such as 0.5), resulting in no warning before the risk outbreak. Historical data show that a threshold of 0.5 will miss 23% of major risks. For example, the preset transition coefficient is determined based on the average value of the risk transition coefficient of the risk financial data in the historical data. For example, the risk transition coefficients of two groups of risk financial data of the same type are 0.4 and 0.3 respectively, and the preset transition coefficient is determined to be 0.35. However, the above value is not limited to this. Those skilled in the art can also adjust the value according to actual needs.

[0070] The present invention determines whether to issue a direct warning or an indirect warning to the risky financial data based on the risk tendency type of the risky financial data and / or the risk transition coefficient of the risky financial data. According to the fact that the risk tendency type of the risky financial data is a strong risk type and the risk transition coefficient of the risky financial data is greater than a preset transition coefficient, it means that the financial data not only has a high risk tendency itself, but also has a high possibility of its risk state changing. The direct warning to the risky financial data is accurately determined. According to the fact that the risk tendency type of the risky financial data is a weak risk type and the risk transition coefficient of the risky financial data is less than or equal to the preset transition coefficient, it means that the risk tendency of the financial data is low, and the possibility of its risk state changing is small, and the potential impact on financial business is relatively small, the indirect warning to the risky financial data is accurately determined. After receiving the warning, according to the risk tendency type of the risk financial data being a strong risk type and the risk transition coefficient of the risk financial data being less than or equal to the preset transition coefficient, it means that although the financial data itself has a high risk tendency, the possibility of its risk status changing is relatively small at present, and the impact on financial business in the short term may be limited. It is determined that no warning will be issued. According to the risk tendency type of the risk financial data being a weak risk type and the risk transition coefficient of the risk financial data being greater than the preset transition coefficient, it means that the risk tendency of the financial data itself is low, but the possibility of risk status changing is high. However, due to its low risk tendency, it is comprehensively judged that the current impact on financial business is not significant, and it is accurately determined that no warning will be issued. The above method improves the accuracy of the financial product early warning system in identifying abnormal data risks, thereby achieving accurate graded early warnings.

[0071] Specifically, the adjustment module determines to adjust the preset matching degree and / or the preset risk type tendency value according to the risk adaptive balance coefficient and adjustment sensitivity within a preset period under the condition that the preset matching degree or the preset risk type tendency value is adjusted;

[0072] If the risk adaptive balance coefficient within the preset period is greater than the preset balance coefficient and the adjustment sensitivity is less than the preset sensitivity, it is determined to adjust the preset matching degree;

[0073] If the risk adaptive balance coefficient within the preset period is less than or equal to the preset balance coefficient and the adjustment sensitivity is greater than or equal to the preset sensitivity, it is determined to adjust the preset risk type propensity value;

[0074] If the risk adaptive balance coefficient within the preset period is greater than the preset balance coefficient and the adjustment sensitivity is greater than or equal to the preset sensitivity, it is determined to adjust the preset matching degree and the preset risk type propensity value;

[0075] If the risk adaptive balance coefficient within the preset period is less than or equal to the preset balance coefficient and the adjustment sensitivity is less than the preset sensitivity, it is determined that no adjustment is required.

[0076] In the embodiment of the present invention, the adjustment amount of the preset matching degree is related to the difference between the risk adaptive balance coefficient and the preset balance coefficient. The larger the difference, the larger the adjustment range. Assume that the risk adaptive balance coefficient is α, the preset balance coefficient is β, and the adjustment coefficient factor is k (k is a constant determined according to historical data and business experience, for example, k=0.1), then the adjustment coefficient A=k×(α-β). For example, the risk adaptive balance coefficient α=0.4, the preset balance coefficient β=0.3, the adjustment range factor k=0.1, the current preset matching degree M=0.85, then the adjustment coefficient A=0.1×(0.4-0.3)=0.01, the new preset matching degree M′=0.85+0.01=0.86, which is still within a reasonable range. The adjustment amount of the preset risk type propensity value is The difference between the adjustment sensitivity and the preset sensitivity is related. The larger the difference, the larger the adjustment range. Suppose the adjustment sensitivity is γ, the preset sensitivity is δ, and the adjustment range factor is m (m is a constant determined based on historical data and business experience, for example, m = 0.05), then the adjustment coefficient B = m×(γ-δ). For example, the adjustment sensitivity γ = 0.2, the preset sensitivity δ = 0.15, the adjustment range factor m = 0.05, the current preset risk type tendency value N = 0.8, then the adjustment range B = 0.05×(0.2-0.15) = 0.0025, the new preset risk type tendency value N′ = 0.8+0.0025 = 0.8025, which is within a reasonable range, but the above values are not limited to this. Those skilled in the art can also adjust the value according to actual needs.

[0077] In the embodiment of the present invention, the risk adaptive balance coefficient is determined according to the risk false alarm rate. For example, the false alarm rate of some financial data is 0.3, and the risk adaptive balance coefficient is determined to be 0.3. The preset balance coefficient range is set to 0.2-0.4, and the preferred value is 0.3. The reason for selecting this preferred value is that through long-term observation and analysis of the same type of financial data in a stable market environment, it is found that when the preset balance coefficient is 0.3, the overall performance of the system is best. The preset period range is set to 8h-12h, and the preferred value is 10h. The reason for selecting this preferred value is that if the period is too short, the system will occupy a large amount of computing resources; and if the period is too long, it will affect the timeliness of risk warning. The period of 10h can ensure timely detection of risks while reasonably allocating system resources. The sensitivity is determined according to the adjustment amplitude of several adjustments within the preset period. For example, within a preset period, the system adjusts a certain type of financial data twice. Each time an adjustment is made, we record the adjustment range. During the first adjustment, a parameter in the normal modification mode is adjusted from 10 times to 8 times, with an adjustment range of 20%; during the second adjustment, the preset risk type propensity value is adjusted from 0.6 to 0.55, with an adjustment range of 8.33%, and the sensitivity is determined to be 0.14. The preset sensitivity is determined based on the average historical sensitivity of the same type of financial data. For example, the historical sensitivities of two groups of the same type of financial data are 0.1 and 0.2 respectively, and the preset sensitivity is 0.15. However, the above values are not limited to this, and those skilled in the art can also adjust the values according to actual needs.

[0078] In the embodiment of the present invention, the adjustment amount of the preset matching degree is positively correlated with the difference between the risk adaptive balance coefficient and the preset balance coefficient, and the adjustment amount of the preset risk type tendency value is positively correlated with the difference between the adjusted sensitivity and the preset sensitivity.

[0079] The present invention adjusts the preset matching degree or the preset risk type tendency value by determining the risk adaptive balance coefficient and / or adjustment sensitivity within a preset period. According to the risk adaptive balance coefficient within the preset period being greater than the preset balance coefficient and the adjustment sensitivity being less than the preset sensitivity, it indicates that the risk false alarm rate of the current system is relatively high, and the system is not sensitive enough to the risk change. The preset matching degree is accurately determined to be adjusted. According to the risk adaptive balance coefficient within the preset period being less than or equal to the preset balance coefficient and the adjustment sensitivity being greater than or equal to the preset sensitivity, it indicates that the risk false alarm rate of the previous system is relatively low, but the system is too sensitive to the risk change, which may lead to frequent adjustments. Adjustment, accurately determine the adjustment of the preset risk type tendency value, according to the risk adaptive balance coefficient within the preset period is greater than the preset balance coefficient and the adjustment sensitivity is greater than or equal to the preset sensitivity, it means that the current system has a high risk false alarm rate, and the system is too sensitive to risk changes. Accurately determine the adjustment of the normal mode and the preset risk type tendency value, according to the risk adaptive balance coefficient within the preset period is less than or equal to the preset balance coefficient and the adjustment sensitivity is less than the preset sensitivity, it means that the system is running stably, accurately determine that no adjustment is needed, and the above method improves the accuracy of the financial product early warning system in identifying abnormal data risks, thereby achieving accurate graded early warning.

[0080] Specifically, a financial risk early warning system applied to a financial risk early warning system and method includes:

[0081] Step S1, obtaining financial data content information, financial data modification information, financial data access permission information, and financial data modification user information;

[0082] Step S2: determining whether the financial data is risky financial data based on whether the sensitive function modification mode of the financial data is a normal modification mode or whether there is an abnormal correlation between the modification position nodes of the financial data;

[0083] Step S3, calculating a risk type tendency value based on the modification concealment index and data contamination diffusion entropy of the risky financial data, and comparing the risk type tendency value with a preset risk type tendency value to determine the risk tendency type of the risky financial data;

[0084] Step S4, determining whether to issue a direct warning or an indirect warning for the risky financial data based on the risk tendency type and / or risk transition coefficient of the risky financial data;

[0085] Step S5: Dynamically adjust the preset matching degree and the preset risk type tendency value based on the risk adaptive balance coefficient and the adjustment sensitivity.

[0086] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

Claims

1. A financial risk early warning system based on big data, characterized by: include: An information collection module, which is used to collect financial data content information, financial data modification information, financial data access permission information, and financial data modification user information; a risk identification module connected to the information collection module, configured to determine risky financial data based on whether a sensitive function modification pattern of the financial data is a normal modification pattern or whether a modification position node of the financial data has an abnormal correlation; a risk judgment module, connected to the risk identification module, configured to calculate a risk type tendency value of the risk financial data based on a modification concealment index and a data contamination diffusion entropy of the risk financial data, and determine a risk tendency type of the risk financial data based on a comparison result of the risk type tendency value with a preset risk type tendency value; a risk warning module, connected to the risk judgment module, for determining whether to issue a direct warning or an indirect warning for the risk financial data based on the risk tendency type of the risk financial data and / or the risk transition coefficient of the risk financial data; An adjustment module is respectively connected to the risk identification module, the risk judgment module and the risk warning module, and is used to determine the adjustment of the risk identification judgment parameters and / or the preset risk type tendency value based on the risk adaptive balance coefficient and adjustment sensitivity within a preset period.

2. The financial risk early warning system based on big data according to claim 1 is characterized in that: The risk identification module determines risky financial data based on whether the sensitive function modification mode of the financial data is a normal modification mode or whether there is an abnormal correlation between the modification position nodes of the financial data; wherein, If the sensitive function modification mode of the financial data is an abnormal modification mode or there is an abnormal correlation between the modification position nodes of the financial data, the financial data is determined to be risky financial data.

3. The financial risk early warning system based on big data according to claim 2 is characterized in that: Whether the sensitive function modification mode of the financial data is a normal modification mode is determined based on the matching degree of the modification behavior of the operating subject, and whether there is an abnormal correlation between the modification position node of the financial data is determined based on whether there is a modification trend with the same abnormal characteristics between the modification position node and its associated position node.

4. The financial risk early warning system based on big data according to claim 3 is characterized in that: The risk judgment module calculates the risk type tendency value of the risk financial data based on the modification concealment index and data pollution diffusion entropy of the risk financial data, and determines the risk tendency type of the risk financial data based on the risk type tendency value; wherein, If the risk type tendency value is greater than the preset risk type tendency value, the risk type is determined to be a strong risk type; If the risk type tendency value is less than or equal to the preset risk type tendency value, the risk type is determined to be a weak risk type.

5. The financial risk early warning system based on big data according to claim 4 is characterized in that: The preset risk type propensity value is determined based on the historical average risk type propensity value of the same type of financial data.

6. The financial risk early warning system based on big data according to claim 5 is characterized in that: The risk warning module determines whether to issue a direct warning or an indirect warning to the risk financial data based on the risk tendency type of the risk financial data and / or the risk transition coefficient of the risk financial data; wherein, If the risk tendency type of the risky financial data is a strong risk type or the risk transition coefficient of the risky financial data is greater than a preset transition coefficient, determining to issue a direct warning for the risky financial data; If the risk tendency type of the risky financial data is a weak risk type and the risk transition coefficient of the risky financial data is less than or equal to a preset transition coefficient, it is determined to issue an indirect warning for the risky financial data.

7. The financial risk early warning system based on big data according to claim 6 is characterized in that: The risk transition coefficient of the risk financial data is determined based on the probability that the risk financial data in the historical data changes from one risk state to another under different conditions. The preset transition coefficient is determined based on the average value of the risk transition coefficients of the risk financial data in the historical data. The preset risk type tendency value is determined based on the historical average risk type tendency value of the same type of financial data.

8. The financial risk early warning system based on big data according to claim 7 is characterized in that: The adjustment module determines to adjust the preset matching degree or the preset risk type tendency value based on the risk adaptive balance coefficient and adjustment sensitivity within a preset period; wherein, If the risk adaptive balance coefficient within the preset period is greater than the preset balance coefficient and the adjustment sensitivity is less than the preset sensitivity, it is determined to adjust the preset matching degree; If the risk adaptive balance coefficient within the preset period is less than or equal to the preset balance coefficient and the adjustment sensitivity is greater than or equal to the preset sensitivity, it is determined to adjust the preset risk type propensity value; If the risk adaptive balance coefficient within the preset period is greater than the preset balance coefficient and the adjustment sensitivity is greater than or equal to the preset sensitivity, it is determined to adjust the preset matching degree and the preset risk type propensity value.

9. The financial risk early warning system based on big data according to claim 8 is characterized in that: The risk adaptive balance coefficient is determined based on the risk false alarm rate, the preset period is determined based on the historical average value of the same type of financial data in a stable market environment, the sensitivity is determined based on the adjustment range during several adjustments within the preset period, and the preset sensitivity is determined based on the average value of the historical sensitivity of the same type of financial data.

10. A financial risk early warning method applied to the financial risk early warning system based on big data according to any one of claims 1 to 9, characterized in that: include: Step S1, obtaining financial data content information, financial data modification information, financial data access permission information, and financial data modification user information; Step S2: determining whether the financial data is risky financial data based on whether the sensitive function modification mode of the financial data is a normal modification mode or whether there is an abnormal correlation between the modification position nodes of the financial data; Step S3, calculating a risk type tendency value based on the modification concealment index and data contamination diffusion entropy of the risky financial data, and comparing the risk type tendency value with a preset risk type tendency value to determine the risk tendency type of the risky financial data; Step S4, determining whether to issue a direct warning or an indirect warning for the risky financial data based on the risk tendency type and / or risk transition coefficient of the risky financial data; Step S5: Dynamically adjust the preset matching degree and the preset risk type tendency value based on the risk adaptive balance coefficient and the adjustment sensitivity.

Images