Transaction abnormality alarm method and device

Abstract

The present invention provides a transaction anomaly alarm method and device, particularly relating to the financial field. The method comprises: obtaining future data of the target asset before a change at multiple future time points based on an asset data association impact model corresponding to both the target asset and its associated assets, the target asset's current target data, the associated assets' current associated data, and the current interest rate; obtaining future data of the target asset after a change at multiple future time points based on the interest rate fluctuation range, the asset data association impact model, the current target data, the current associated data, and the current interest rate; and determining whether a transaction anomaly exists for the target asset based on the future data before and after the change corresponding to the future time points, and if so, issuing a transaction anomaly alarm. The present invention can improve the accuracy of transaction anomaly alarms, thereby increasing the efficiency of transaction anomaly alarms.

Description

Technical Field

[0001] The present invention relates to the field of transaction processing technology, in particular to the financial field, and more particularly to a transaction abnormality alarm method and device. Background Art

[0002] Asset data often fluctuates due to various financial and social factors. Excessive fluctuations in asset data can hinder the coordination and overall planning of related transactions, hindering the smooth progress of transactions and potentially leading to losses for both parties. Consequently, large fluctuations in asset data can lead to corresponding transaction anomalies. Activating these anomaly alerts often requires predicting fluctuations in the target asset data, thus providing a basis for generating alerts based on these fluctuations.

[0003] In existing technology, transaction anomaly alerts primarily rely on staff analyzing target assets to predict their fluctuations, which in turn triggers anomaly alerts. However, this approach relies on staff experience, and the information used for analysis may not be strongly correlated with asset data fluctuations. This often results in false alerts when no anomalies exist, or no alerts when anomalies do exist. Consequently, the accuracy of transaction anomaly alerts is low.

[0004] In summary, the prior art has the problem that the accuracy of transaction anomaly alarms is low, thereby causing the efficiency of transaction anomaly alarms to be low. Summary of the Invention

[0005] One object of the present invention is to provide a transaction anomaly alarm method to address the low accuracy and, consequently, low efficiency of transaction anomaly alarms in the prior art. Another object of the present invention is to provide a transaction anomaly alarm device. A further object of the present invention is to provide a computer device. Yet another object of the present invention is to provide a readable medium.

[0006] In order to achieve the above objectives, one aspect of the present invention discloses a transaction abnormality alarm method, the method comprising:

[0007] Obtaining future data of the target asset before changes at multiple future time points based on an asset data association impact model corresponding to a target asset and associated assets corresponding to the target asset, current target data of the target asset, current associated data of the associated assets, and a current interest rate;

[0008] Obtaining the target asset's future data after changes at the plurality of future time points based on the expected interest rate change range, the asset data association impact model, the current target data, the current association data, and the current interest rate;

[0009] Based on the future data before the change and the future data after the change corresponding to the multiple future time points, it is determined whether there is a transaction abnormality in the target asset. If so, a transaction abnormality alarm is issued.

[0010] Optionally, further including:

[0011] Before obtaining the pre-change future data of the target asset at multiple future time points based on the asset data association impact model corresponding to the target asset and the associated assets corresponding to the target asset, the current target data of the target asset, the current association data of the associated assets, and the current interest rate, an initial target model of the target asset and initial association models of the associated assets corresponding to the target asset are constructed based on a preset data dynamic function;

[0012] Obtaining a target model in the asset data association impact model based on the target historical data information of the target asset, the associated historical data information of the associated assets, the corresponding historical interest rates, and the initial target model;

[0013] Based on the target historical data information, the associated historical data information, the historical interest rate and the initial association model, an association model in the asset data association impact model is obtained.

[0014] Optionally, constructing an initial target model of the target asset and an initial association model of associated assets corresponding to the target asset based on a preset data dynamic function includes:

[0015] Based on the plurality of data dynamic functions, respectively constructing a first data item corresponding to the impact of the target asset itself on the target asset data, a second data item corresponding to the impact of the associated asset on the target asset data, a third data item corresponding to the impact of the associated asset and the target asset on the target asset data, a fourth data item corresponding to the impact of the associated asset itself on the associated asset data, a fifth data item corresponding to the impact of the target asset on the associated asset data, and a sixth data item corresponding to the impact of the associated asset and the target asset on the associated asset data;

[0016] Constructing the initial target model based on the preset first data change amount data item, the time difference data item, the first data item, the second data item and the third data item;

[0017] The initial association model is constructed based on the preset second data variation data item, the time difference data item, the fourth data item, the fifth data item and the sixth data item.

[0018] Optionally, obtaining the target model in the asset data association impact model based on the target historical data information of the target asset, the associated historical data information of the associated assets, the corresponding historical interest rates, and the initial target model includes:

[0019] Based on the historical target data of multiple historical time points in the target historical data information, obtaining the historical target data changes corresponding to the historical time points;

[0020] Substituting historical target data at multiple historical time points in the target historical data information, historical target data changes, historical associated data at multiple historical time points in the associated historical data information, the historical interest rate, and historical time differences between corresponding adjacent historical time points into the initial target model to obtain multiple undetermined target models;

[0021] Solve the multiple undetermined target models to determine the corresponding first coefficient of the first data item, the second coefficient of the second data item and the third coefficient of the third data item, and obtain the target model based on the first coefficient, the second coefficient, the third coefficient and the initial target model.

[0022] Optionally, obtaining the association model in the asset data association impact model based on the target historical data information, the associated historical data information, the historical interest rate, and the initial association model includes:

[0023] Based on the historical associated data of multiple historical time points in the associated historical data information, obtaining the historical associated data changes corresponding to the historical time points;

[0024] Substituting historical target data at multiple historical time points in the target historical data information, historical associated data at multiple historical time points in the associated historical data information, changes in historical associated data, the historical interest rate, and historical time differences corresponding to adjacent historical time points into the initial association model to obtain multiple undetermined association models;

[0025] Solve the multiple undetermined association models to determine the corresponding fourth coefficient of the fourth data item, the fifth coefficient of the fifth data item, and the sixth coefficient of the sixth data item, and obtain the association model based on the fourth coefficient, the fifth coefficient, the sixth coefficient, and the initial association model.

[0026] Optionally, further including:

[0027] Before obtaining the pre-change future data of the target asset at multiple future time points based on the asset data association impact model corresponding to the target asset and the associated assets corresponding to the target asset, the current target data of the target asset, the current associated data of the associated assets, and the current interest rate,

[0028] Based on the verification target data of multiple verification time points of the target asset, the verification association data of multiple verification time points of the associated assets, the corresponding verification interest rates and the verification time differences of the corresponding adjacent verification time points, it is determined whether the asset data association impact model is applicable. If not, a model applicability alarm is issued.

[0029] Optionally, the determining whether the asset data association impact model is applicable based on the verification target data of the target asset at multiple verification time points, the verification association data of the associated assets at multiple verification time points, the corresponding verification interest rates, and the verification time differences of the corresponding adjacent verification time points includes:

[0030] Input the corresponding verification time difference, verification rate, verification target data corresponding to the earliest verification time point, and verification association data into the target model and association model in the asset data association impact model for cross-iteration calculation to obtain test output data of multiple verification time points of the target asset;

[0031] Determining error rates corresponding to a plurality of the verification time points based on the corresponding verification target data and test output data;

[0032] Determine whether there is no verification time point whose corresponding error rate is greater than or equal to the preset error rate threshold; if not, issue a model applicability alarm.

[0033] Optionally, obtaining the pre-change future data of the target asset at multiple future time points based on the asset data association impact model corresponding to the target asset and the associated assets corresponding to the target asset, the current target data of the target asset, the current association data of the associated assets, and the current interest rate includes:

[0034] The preset predicted time difference, the current target data, the current associated data and the current interest rate are input into the target model and the associated model in the asset data associated impact model for cross-iterative calculation to obtain the future data of the target asset before the change at multiple future time points.

[0035] Optionally, obtaining the changed future data of the target asset at multiple future time points based on the expected interest rate change range, the asset data association impact model, the current target data, the current association data, and the current interest rate includes:

[0036] Based on the interest rate change range and the current interest rate, obtain the expected interest rate after the change;

[0037] The predicted time difference, the current target data, the current associated data and the expected interest rate after the change are input into the target model and the associated model in the asset data association impact model for cross-iteration calculation to obtain the future data of the target asset after the change at multiple future time points.

[0038] Optionally, the determining whether there is any transaction anomaly in the target asset based on the pre-change future data and the post-change future data corresponding to the plurality of future time points includes:

[0039] Obtaining a target data change range corresponding to the future time point based on the corresponding future data before and after the change;

[0040] Determine whether there is a future time point where the corresponding target data change amplitude is greater than or equal to a preset data change amplitude threshold. If so, issue a transaction abnormality alarm.

[0041] Optionally, the determining whether there is any transaction anomaly in the target asset based on the pre-change future data and the post-change future data corresponding to the plurality of future time points includes:

[0042] Obtaining a target data change range corresponding to the future time point based on the corresponding future data before and after the change;

[0043] Obtaining a target data change rate corresponding to the future time point based on the target data change amplitude and the future data before the change or the future data after the change;

[0044] Determine whether there is a future time point where the corresponding target data change rate is greater than or equal to a preset data change rate threshold. If so, issue a transaction abnormality alarm.

[0045] In order to achieve the above objectives, another aspect of the present invention discloses a transaction abnormality alarm device, the device comprising:

[0046] A first prediction module is configured to obtain pre-change future data of the target asset at multiple future time points based on an asset data association impact model corresponding to a target asset and associated assets corresponding to the target asset, current target data of the target asset, current associated data of the associated assets, and a current interest rate;

[0047] A second prediction module is configured to obtain the target asset's future data after changes at a plurality of future time points based on the expected interest rate change range, the asset data association impact model, the current target data, the current association data, and the current interest rate;

[0048] The abnormality alarm module is used to determine whether there is a transaction abnormality in the target asset based on the future data before the change and the future data after the change corresponding to the multiple future time points, and if so, issue a transaction abnormality alarm.

[0049] The present invention also discloses a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the method described above is implemented when the processor executes the program.

[0050] The present invention also discloses a computer-readable medium on which a computer program is stored. When the program is executed by a processor, the method described above is implemented.

[0051] The transaction abnormality alarm method and device provided by the present invention obtain the target asset's future data before the change at multiple future time points based on the asset data association impact model corresponding to the target asset and the associated assets corresponding to the target asset, the current target data of the target asset, the current associated data of the associated assets and the current interest rate. It can use a model related to the corresponding data impact characteristics between the target asset and the associated assets, and use the corresponding actual current data and interest rate as input for calculation and processing to accurately and quickly predict the future data of the target asset under the current interest rate, thereby improving the accuracy of the overall transaction abnormality alarm; by using the expected interest rate change range, the asset data association impact model, the current target data, the current associated data and the current interest rate Interest rate, obtain the future data of the target asset after the change at multiple future time points, and use a model related to the corresponding data impact characteristics between the target asset and the related assets, with the corresponding actual current data, interest rate and predicted interest rate change range as input for calculation and processing, accurately and quickly predict the future data of the target asset at the same time after the interest rate change, thereby improving the accuracy of the overall transaction abnormality alarm; by judging whether there is a transaction abnormality of the target asset based on the future data before the change and the future data after the change corresponding to multiple future time points, if so, a transaction abnormality alarm is issued, which can comprehensively reflect the future data before the change and the future data after the change at the same time that reflect the relative volatility of the target asset, thereby improving the accuracy of the transaction abnormality alarm.

[0052] The transaction abnormality alarm method and device provided by the present invention can, on the one hand, realize automated execution in the form of programs, functions, algorithms, software or applications, thereby greatly reducing the dependence on the work experience of staff. On the other hand, it can fully consider the impact of associated assets and corresponding interest rates related to the target asset on the target asset data, so that the basis information of the analysis has a strong correlation with the fluctuation of the asset data. Specifically, the change in interest rate is closely related to the fluctuation of target asset data, and the fluctuation of target asset data is further affected by the situation of its corresponding associated assets. The determined future data before the change is the future data of the target asset predicted under the assumption that the interest rate remains unchanged, and the future data before the change is closely related to the target asset's own situation, the situation of the associated assets, and the interest rate; the determined future data after the change is the future data of the target asset predicted under the assumption that the interest rate changes (consistent with the actual situation, in which the interest rate often changes), and the future data after the change is also closely related to the target asset's own situation, the situation of the associated assets, and the interest rate. Therefore, the combined future data before the change and the future data after the change can intuitively and accurately represent the impact of interest rate changes on the fluctuation of target asset data. Moreover, since the corresponding associated asset information is also used as a basis, the combined future data before the change and the future data after the change can also intuitively and accurately represent the impact of the situation of the associated assets on the fluctuation of target asset data, which is in line with relevant economic laws. Therefore, the transaction abnormality alarm method and device provided by the present invention can greatly improve the accuracy of predicting the fluctuation of target asset data, thereby greatly improving the accuracy of transaction abnormality alarm based on the fluctuation of target asset data.

[0053] In summary, the transaction abnormality alarm method and device provided by the present invention can improve the accuracy of transaction abnormality alarms, thereby improving the efficiency of transaction abnormality alarms. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0055] Figure 1 A schematic diagram showing a process of a transaction abnormality alarm method according to an embodiment of the present invention;

[0056] Figure 2 A schematic diagram showing an optional step of obtaining a target model according to an embodiment of the present invention is shown;

[0057] Figure 3A schematic diagram showing an optional step of obtaining an association model according to an embodiment of the present invention is shown;

[0058] Figure 4 A schematic diagram showing an optional step of determining whether a target asset has transaction anomalies according to an embodiment of the present invention is shown;

[0059] Figure 5 A schematic diagram showing another optional step of determining whether there is a transaction anomaly in a target asset according to an embodiment of the present invention is shown;

[0060] Figure 6 A module diagram of a transaction abnormality alarm device according to an embodiment of the present invention is shown;

[0061] Figure 7 A schematic structural diagram of a computer device suitable for implementing an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0062] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0063] The terms “first,” “second,” etc. used herein do not particularly refer to an order or sequence, nor are they used to limit the present invention. They are only used to distinguish elements or operations described with the same technical terms.

[0064] The words “include,” “including,” “have,” “contain,” etc. used in this document are open-ended terms, meaning including but not limited to.

[0065] As used herein, "and / or" includes any and all combinations of the items mentioned.

[0066] It should be noted that the acquisition, storage, use, and processing of data in the technical solution of the present invention are in compliance with the relevant provisions of national laws and regulations.

[0067] It should be noted that the transaction abnormality alarm method and device disclosed in this application can be used in the field of transaction processing technology, and can also be used in any field other than the field of transaction processing technology. The application field of the transaction abnormality alarm method and device disclosed in this application is not limited.

[0068] The embodiment of the present invention discloses a transaction abnormality alarm method, such as Figure 1 As shown, the method specifically includes the following steps:

[0069] S101: Based on the asset data association impact model corresponding to the target asset and the associated assets corresponding to the target asset, the current target data of the target asset, the current associated data of the associated assets and the current interest rate, obtain the future data of the target asset before the change at multiple future time points.

[0070] S102: Based on the expected interest rate change range, the asset data association impact model, the current target data, the current association data and the current interest rate, obtain the future data of the target asset after the change at the multiple future time points.

[0071] S103: Based on the future data before the change and the future data after the change corresponding to the multiple future time points, determine whether there is any transaction abnormality in the target asset, and if so, issue a transaction abnormality alarm.

[0072] Exemplarily, after determining that the target asset has a transaction anomaly, a corresponding transaction anomaly prompt message may be sent to the asset holder corresponding to the target asset (e.g., the terminal of the asset holder corresponding to the target asset), and the asset holder may be advised to replace the corresponding target asset before re-participating in the transaction (e.g., the corresponding asset replacement suggestion message is sent at the same time). Alternatively, after determining that the target asset has a transaction anomaly, based on the pre-change future data and post-change future data corresponding to multiple future time points, a data difference value for each future time point may be obtained (e.g., the absolute value of the difference between the pre-change future data and the post-change future data of the future time point is used as the data difference value), and a future time point with a data difference value less than a preset data difference value threshold may be used as a pending transaction time point, and transaction processing may be performed based on the target asset within the transaction time period between the pending transaction time point and the next future time point and / or the transaction time period between the pending transaction time point and the previous future time point. Preferably, after determining that the target asset has a transaction anomaly, based on the future data before the change and the future data after the change corresponding to multiple future time points, a data difference value for each future time point is obtained (for example, the absolute value of the difference between the future data before the change and the future data after the change at the future time point is used as the data difference value), and the data difference value is divided by the corresponding future data before the change or the future data after the change to obtain a data difference coefficient, and then the future time point with a data difference coefficient less than a preset data difference coefficient threshold is used as the pending transaction time point, and the transaction processing can be performed based on the target asset during the transaction time period between the pending transaction time point and the next future time point and / or the transaction time period between the pending transaction time point and the previous future time point. The specific processing method of the relevant data and information involved in the transaction processing can be determined by those skilled in the art according to actual circumstances, and the embodiment of the present invention does not limit this. Further, after determining that the target asset has a transaction anomaly, the relevant target asset information of the target asset and the relevant transaction object asset information of the transaction object asset that is the transaction object of the target asset can be sent to the corresponding staff member who specializes in handling abnormal transactions (specifically, it can be sent to the staff terminal of the staff member who specializes in handling abnormal transactions) so that the staff member can perform the transaction processing. It should be noted that the specific content and type of relevant measures after determining that there is an abnormal transaction in the target asset can be determined by those skilled in the art based on actual circumstances. The above description is only an example and does not constitute a limitation.

[0073] For example, the associated assets may be, but are not limited to, assets that have a clear financial relationship of mutual influence, interaction, or interaction with the target asset. For example, if the target asset is a mixed fund A, and when the mixed fund A is established, it is determined that a portion of its income needs to be used to invest in bond B, then there is at least a mutual influence, interaction, and interaction relationship between the mixed fund A and bond B, and bond B is the associated asset of the mixed fund A. For another example, if the target asset is mobile phone C, and mobile phone C requires a certain silicon chip D to be manufactured, then there is at least a mutual influence and interaction relationship between mobile phone C and silicon chip D, and therefore, silicon chip D is the associated asset of mobile phone C. It should be noted that the nature and determination of associated assets can be determined by those skilled in the art based on actual circumstances. The above description is only an example and does not constitute a limitation.

[0074] For example, the data in the embodiments of the present invention may be, but is not limited to, relevant value, the number of transactions within a preset time period for all assets of the same type, or the relative popularity coefficient of the same type of assets as a whole, preferably value. Specifically, it may be, but is not limited to, the unit value or total value of the asset. The value may also be further taken as the corresponding unit price or total price. It should be noted that the specific nature of the data can be determined by those skilled in the art based on actual circumstances, and the above description is merely an example and does not constitute a limitation.

[0075] For example, the transaction anomaly alert may include, but is not limited to, sending a warning message such as "The data for asset xxx fluctuates significantly, making it unsuitable for trading based on it. A transaction anomaly exists. Please verify and address it promptly" to the corresponding transaction parties, managers, or staff. It should be noted that the specific implementation of the transaction anomaly alert can be determined by those skilled in the art based on actual circumstances, and the above description is merely an example and does not constitute a limitation.

[0076] Exemplarily, an asset data association impact model corresponds to a target asset and an associated asset, a target asset corresponds to a current target data, an associated asset corresponds to a current associated data, and a target asset corresponds to a future data before the change and a future data after the change at a future time point. It should be noted that the relevant corresponding relationships can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation. The transaction abnormality alarm method and device provided by the present invention obtains the future data of the target asset before the change at multiple future time points based on the asset data association impact model corresponding to the target asset and the associated assets corresponding to the target asset, the current target data of the target asset, the current associated data of the associated asset and the current interest rate. It can use a model related to the corresponding data impact characteristics between the target asset and the associated asset, and perform calculations with the corresponding actual current data and interest rate as input to accurately and quickly predict the future data of the target asset at the current interest rate, thereby improving the accuracy of the overall transaction abnormality alarm; by based on the expected interest rate change range, the asset data association impact model, the current target data, the current associated data and the current interest rate. The method and apparatus for detecting an abnormality in the transaction are used to determine whether the target asset has an abnormal transaction. The method and apparatus can be used to detect the abnormal transaction in the target asset at a plurality of future time points ...Specifically, the change in interest rate is closely related to the fluctuation of target asset data, and the fluctuation of target asset data is further affected by the situation of its corresponding associated assets. The determined future data before the change is the future data of the target asset predicted under the assumption that the interest rate remains unchanged, and the future data before the change is closely related to the target asset's own situation, the situation of the associated assets, and the interest rate; the determined future data after the change is the future data of the target asset predicted under the assumption that the interest rate changes (consistent with the actual situation, in which the interest rate often changes), and the future data after the change is also closely related to the target asset's own situation, the situation of the associated assets, and the interest rate. Therefore, the combined future data before the change and the future data after the change can intuitively and accurately represent the impact of interest rate changes on the fluctuation of target asset data. Moreover, since the corresponding associated asset information is also used as a basis, the combined future data before the change and the future data after the change can also intuitively and accurately represent the impact of the situation of the associated assets on the fluctuation of target asset data, which is in line with relevant economic laws. Therefore, the transaction abnormality alarm method and device provided by the present invention can greatly improve the accuracy of predicting the fluctuation of target asset data, thereby greatly improving the accuracy of transaction abnormality alarm based on the fluctuation of target asset data.

[0077] In summary, the transaction abnormality alarm method and device provided by the present invention can improve the accuracy of transaction abnormality alarms, thereby improving the efficiency of transaction abnormality alarms.

[0078] In an optional embodiment, further comprising:

[0079] Before obtaining the pre-change future data of the target asset at multiple future time points based on the asset data association impact model corresponding to the target asset and the associated assets corresponding to the target asset, the current target data of the target asset, the current association data of the associated assets, and the current interest rate, an initial target model of the target asset and initial association models of the associated assets corresponding to the target asset are constructed based on a preset data dynamic function;

[0080] Obtaining a target model in the asset data association impact model based on the target historical data information of the target asset, the associated historical data information of the associated assets, the corresponding historical interest rates, and the initial target model;

[0081] Based on the target historical data information, the associated historical data information, the historical interest rate and the initial association model, an association model in the asset data association impact model is obtained.

[0082] For example, one target asset corresponds to one initial target model and one target model, and one associated asset corresponds to one initial associated model and one associated model. It should be noted that the corresponding relationships can be determined by those skilled in the art based on actual circumstances, and the above description is merely an example and does not constitute a limitation.

[0083] For example, the initial target model, target model, initial association model, and association model can be implemented by, but not limited to, corresponding equations or can be a neural network model, etc., preferably implemented by corresponding equations. It should be noted that the specific forms of the initial target model, target model, initial association model, and association model can be determined by those skilled in the art based on actual circumstances, and the above description is merely illustrative and does not constitute a limitation thereto.

[0084] Through the above steps, based on the target historical data of the actual target asset, the associated historical data of the associated assets and the corresponding historical interest rates, a target model representing the impact characteristics of the associated assets and the corresponding interest rates on the target asset data and an association model representing the impact characteristics of the target assets and the corresponding interest rates on the associated asset data can be constructed respectively. In this way, the asset data association impact model is more consistent with the corresponding data impact characteristics between the target asset and the associated assets and is more comprehensive, thereby preparing for the subsequent accurate determination of the future data before and after the change of the target asset, and improving the accuracy of the overall transaction abnormality alarm.

[0085] In an optional embodiment, the constructing of the initial target model of the target asset and the initial association model of the associated assets corresponding to the target asset based on the preset data dynamic function includes:

[0086] Based on the plurality of data dynamic functions, respectively constructing a first data item corresponding to the impact of the target asset itself on the target asset data, a second data item corresponding to the impact of the associated asset on the target asset data, a third data item corresponding to the impact of the associated asset and the target asset on the target asset data, a fourth data item corresponding to the impact of the associated asset itself on the associated asset data, a fifth data item corresponding to the impact of the target asset on the associated asset data, and a sixth data item corresponding to the impact of the associated asset and the target asset on the associated asset data;

[0087] Constructing the initial target model based on the preset first data change amount data item, the time difference data item, the first data item, the second data item and the third data item;

[0088] The initial association model is constructed based on the preset second data variation data item, the time difference data item, the fourth data item, the fifth data item and the sixth data item.

[0089] For example, when corresponding target models and associated models are implemented via corresponding equations, the corresponding data items can be, but are not limited to, equation terms. The properties of the target models and associated models can be understood as, but are not limited to, submodels of the corresponding asset data association impact model, while the relevant data items can be understood as components of the relevant execution logic of the corresponding submodels. Within the model, the corresponding execution logic can naturally be embodied as corresponding functions or equations. It should be noted that the specific structure and properties of the model can be determined by those skilled in the art based on actual circumstances, and the above description is merely illustrative and does not constitute a limitation.

[0090] For example, the data dynamic function can be determined by those skilled in the art according to actual conditions, and the present invention does not limit this. For example, the data dynamic function can include but is not limited to f(x)=x 2 +x, etc., such as f(x) = e 0.01x Exponential functions such as f(x)=ln(x), logarithmic functions such as f(x)=sin(x), and trigonometric functions such as f(x)=2 0.1x sin(x)+x 2 Preferably, the data dynamic function can be taken as f(x)=x.

[0091] Exemplarily, based on the multiple data dynamic functions, a first data item corresponding to the impact of the target asset itself on the target asset data, a second data item corresponding to the impact of the associated asset on the target asset data, a third data item corresponding to the joint impact of the associated asset and the target asset on the target asset data, a fourth data item corresponding to the impact of the associated asset itself on the associated asset data, a fifth data item corresponding to the impact of the target asset on the associated asset data, and a sixth data item corresponding to the joint impact of the associated asset and the target asset on the associated asset data are respectively constructed. This can be, but is not limited to, multiplying the functional formula of the corresponding data dynamic function (usually the formula on the right side of the function equal sign) with the corresponding item coefficient variable to obtain the corresponding first data item, second data item, third data item, fourth data item, fifth data item and sixth data item, respectively, wherein the variable of the functional formula of the first data item is the target asset data X, the variable of the functional formula of the second data item is the associated asset data Y, the functional formula of the third data item is the target asset data X and the associated asset data Y, the functional formula of the fourth data item is the associated asset data Y, the functional formula of the fifth data item is the target asset data X, and the functional formula of the sixth data item is the target asset data X and the associated asset data Y.

[0092] For example, the first data item can be expressed as but not limited to arf1(X), where r represents the interest rate variable, a represents the term coefficient of the first data item, and f1(X) represents the functional expression of the first data item; the second data item can be expressed as but not limited to bg1(Y), where b represents the term coefficient of the second data item, and g1(Y) represents the functional expression of the second data item; the third data item can be expressed as but not limited to αh1(XY), where α represents the term coefficient of the third data item, and h1(XY) represents the function of the third data item. Formula; the fourth data item can be expressed as but not limited to crf2(Y), where r represents the interest rate variable, c represents the item coefficient of the fourth data item, and f2(Y) represents the functional formula of the fourth data item; the fifth data item can be expressed as but not limited to dg2(X), where d represents the item coefficient of the fifth data item, and g2(X) represents the functional formula of the fifth data item; the sixth data item can be expressed as but not limited to βh2(XY), where β represents the item coefficient of the sixth data item, and h2(XY) represents the functional formula of the sixth data item.

[0093] Among them, the specific logic of the functional formulas of the first data item, the second data item, the third data item, the fourth data item, the fifth data item and the sixth data item can be the same (for example, their logic is f(x)=x) or different.

[0094] It should be noted that, for the specific implementation method of constructing, based on multiple data dynamic functions, the first data item corresponding to the impact of the target asset itself on the target asset data, the second data item corresponding to the impact of the associated asset on the target asset data, the third data item corresponding to the joint impact of the associated asset and the target asset on the target asset data, the fourth data item corresponding to the impact of the associated asset itself on the associated asset data, the fifth data item corresponding to the impact of the target asset on the associated asset data, and the sixth data item corresponding to the joint impact of the associated asset and the target asset on the associated asset data, can be determined by those skilled in the art according to actual circumstances. The above description is only an example and does not constitute a limitation to this.

[0095] Exemplarily, the time difference data item may be, but is not limited to, a corresponding time difference variable, the time dimension of which needs to be the same as the time dimension of the interest rate. For example, if the time difference variable takes a value of 1 day, the interest rate needs to be the corresponding daily interest rate; if the time difference variable takes a value of 1 year, the interest rate needs to be an annual interest rate. Preferably, the time difference variable takes a value of 1 day, and the corresponding interest rate is a daily interest rate. It should be noted that the specific configuration and nature of the time difference data item can be determined by those skilled in the art based on actual circumstances, and the above description is merely an example and does not constitute a limitation thereto.

[0096] Exemplarily, the first data delta data item and the second data delta data item may be, but are not limited to, corresponding data delta variables. It should be noted that the specific configuration and nature of the first data delta data item and the second data delta data item can be determined by those skilled in the art based on actual circumstances, and the above description is merely an example and does not constitute a limitation thereto.

[0097] Exemplarily, the initial target model is constructed based on the preset first data variation data item, time difference data item, the first data item, the second data item, and the third data item. This can be, but is not limited to, performing operations such as addition, subtraction, multiplication, and division on the first data item, the second data item, and the third data item to obtain first sub-logical information, then performing operations such as multiplication and division on the first sub-logical information and the time difference data item to form first main logical information, and finally, setting the first main logical information on one side of the equal sign and the first data variation data item on the other side of the equal sign to construct the initial target model. The following shows an exemplary expression of the execution logic of the initial target model:

[0098] dX=(arf1(X)-bg1(Y)+αh1(XY))dt

[0099] Among them, dX represents the first data change data item, dt represents the time difference data item, arf1(X) represents the first data item, bg1(Y) represents the second data item, αh1(XY) represents the third data item, (arf1(X)-bg1(Y)+αh1(XY)) represents the first sub-logical information, and (arf1(X)-bg1(Y)+αh1(XY))dt represents the first main logical information.

[0100] It should be noted that the specific implementation method of constructing the initial target model based on the preset first data change data item, time difference data item, the first data item, the second data item and the third data item can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0101] Exemplarily, the initial association model is constructed based on the preset second data variation data item, the time difference data item, the fourth data item, the fifth data item, and the sixth data item. This may be, but is not limited to, performing operations such as addition, subtraction, multiplication, and division on the fourth data item, the fifth data item, and the sixth data item to obtain second sub-logical information, then performing operations such as multiplication and division on the second sub-logical information and the time difference data item to form second main logical information, and finally, setting the second main logical information on one side of the equal sign and the second data variation data item on the other side of the equal sign to construct the initial association model. The following shows an exemplary expression of the execution logic of the initial association model:

[0102] dY=(crf2(Y)-dg2(X)+βh2(XY))dt

[0103] Among them, dY represents the second data change data item, dt represents the time difference data item, crf2(Y) represents the fourth data item, dg2(X) represents the fifth data item, βh2(XY) represents the sixth data item, (crf2(Y)-dg2(X)+βh2(XY)) represents the second sub-logic information, and (crf2(Y)-dg2(X)+βh2(XY))dt represents the second main logic information.

[0104] It should be noted that the specific implementation method of constructing the initial association model based on the preset second data change data item, the time difference data item, the fourth data item, the fifth data item and the sixth data item can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0105] Through the above steps, the initial target model can be further correlated with the impact of the target asset's own situation on its own data, the impact of associated assets on the target asset's data, and the combined impact of the target and associated assets on the target asset's data, as well as the interplay and collaborative characteristics between these impacts. This allows the initial target model to more accurately and comprehensively align with the impact of data fluctuations on the target asset. The initial correlation model can also be further correlated with the impact of associated assets' own situation on their own data, the impact of the target asset on the associated asset's data, and the combined impact of the target and associated assets on the associated asset's data, as well as the interplay and collaborative characteristics between these impacts. This allows the initial correlation model to more accurately and comprehensively align with the impact of data fluctuations on the associated assets. Target asset data fluctuations are influenced by associated asset data, and associated asset data fluctuations affect associated asset data. Therefore, the above steps can further ensure that the subsequently determined asset data correlation impact model is more closely aligned with the corresponding data impact characteristics between the target asset and its associated assets, thereby improving the accuracy of overall transaction anomaly alerts.

[0106] In an optional embodiment, if Figure 2 As shown, the target model in the asset data association impact model is obtained based on the target historical data information of the target asset, the associated historical data information of the associated assets, the corresponding historical interest rate and the initial target model, including the following steps:

[0107] S201: Based on the historical target data of multiple historical time points in the target historical data information, obtain the historical target data changes corresponding to the historical time points.

[0108] S202: Substitute the historical target data of multiple historical time points in the target historical data information, the historical target data changes, the historical associated data of multiple historical time points in the associated historical data information, the historical interest rate and the historical time difference between the corresponding adjacent historical time points into the initial target model to obtain multiple undetermined target models.

[0109] S203: Solve the multiple undetermined target models to determine the corresponding first coefficient of the first data item, the second coefficient of the second data item, and the third coefficient of the third data item, and obtain the target model based on the first coefficient, the second coefficient, the third coefficient, and the initial target model.

[0110] For example, the time difference between adjacent historical time points can be fixed (i.e., the historical time points can be expressed as an arithmetic progression after being arranged in sequence. In layman's terms, the time difference corresponding to different groups of adjacent historical time points is the same), or it can be variable. If the time difference between adjacent historical time points is fixed, then there is an example as follows: if there are historical time points A, B, C, and D, then the time difference between B and A, the time difference between C and B, and the time difference between D and C are all the same.

[0111] Exemplarily, the multiple historical time points corresponding to the target historical data information are preferably the same as the multiple historical time points corresponding to the associated historical data information. For example, the multiple historical time points corresponding to the target historical data information include time point A, time point B, and time point C, and the multiple historical time points corresponding to the associated historical data information also include time point A, time point B, and time point C.

[0112] Exemplarily, step S201 may be, but is not limited to, for each historical time point, subtracting the historical target data of the current time point from the historical target data of the next historical time point of the current historical time point to obtain the historical target data change corresponding to the current historical time point. For example, if there is a historical time point A (the corresponding historical target data is x1), a historical time point B (the corresponding historical target data is x2), and a historical time point C (the corresponding historical target data is x3), then the historical target data change corresponding to historical time point A is x2-x1, and the historical target data change corresponding to historical time point B is x3-x2. If there is no next historical time point after historical time point C, then there is no need to make historical time point C correspond to a historical target data change (and the historical target data corresponding to historical time point C does not need to be involved in the corresponding solution operation subsequently). It should be noted that the specific implementation method of step S201 can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation thereto.

[0113] For example, a pending target model corresponds to a historical target data item at a time point, a historical associated data item at a time point, a change in the historical target data item at a time point, and a historical time difference corresponding to a time point (i.e., the time difference between the next historical time point after the current historical time point and the current historical time point). In summary, a pending target model corresponds to a historical time point. It should be noted that the relevant correspondence can be determined by those skilled in the art based on actual circumstances, and the above description is merely an example and does not constitute a limitation.

[0114] Exemplarily, step S202 may include, but is not limited to, substituting the historical target data and historical interest rate into the first data item in the initial target model, substituting the historical target data change into the first data change data item in the initial target model, substituting the historical correlation data into the second data item in the initial target model, substituting the historical correlation data and historical target data into the third data item in the initial target model, and substituting the historical time difference into the time difference data item in the initial target model to obtain an undetermined target model (at this time, the value of the corresponding item coefficient variable is still uncertain, that is, the corresponding item coefficient variable is the variable that needs to be solved). It should be noted that the specific implementation of step S202 can be determined by those skilled in the art based on actual circumstances, and the above description is only an example and does not constitute a limitation thereto.

[0115] Exemplarily, the step S203 may be, but is not limited to, first grouping multiple undetermined target models to obtain multiple undetermined target model groups, wherein the number of undetermined target models in one undetermined target model group is the number of variables (term coefficient variables) that need to be solved in one of the undetermined target models (generally 3 in the embodiment of the present invention), and there may or may not be intersections between the multiple undetermined target model groups. Then, for each of the target model groups to be determined (when the relevant logic of the model can be embodied as an equation), an equation-solving operation is performed using methods such as the Euler method, the Runge-Kutta method, and other equation-solving methods to obtain the first coefficient to be processed, the second coefficient to be processed, and the third coefficient to be processed corresponding to each of the target model groups to be determined (one target model group to be determined corresponds to one first coefficient to be processed, one second coefficient to be processed, and one third coefficient to be processed). After that, the average value of the plurality of first coefficients to be processed is used as the first coefficient, the average value of the plurality of second coefficients to be processed is used as the second coefficient, and the average value of the plurality of third coefficients to be processed is used as the third coefficient. Finally, the first coefficient, the second coefficient, and the third coefficient are substituted into the initial target model to obtain the target model. It should be noted that the specific implementation of step S203 can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation thereto.

[0116] Preferably, if historical target data or historical associated data is missing for some of the multiple historical time points, data repair and data cleaning methods such as spline interpolation can be used to repair the missing historical target data or historical associated data for the historical time points, so that the historical time points that originally had missing corresponding data can now correspond to one historical target data and one historical associated data. Further preferably, if the time difference between the multiple historical time points is fixed, the missing historical target data or historical associated data for the corresponding historical time points can be determined using the following formula:

[0117]

[0118] Where i represents the number of the previous historical time point with no missing historical target data and no missing historical associated data relative to the current historical time point (the historical time point corresponding to the missing data), k represents the number of the next historical time point with no missing historical target data and no missing historical associated data relative to the current historical time point (the historical time point corresponding to the missing data), j represents the difference between the number of the current historical time point and the number of the previous historical time point with no missing historical target data and no missing historical associated data, and x i+j Indicates the historical target data / historical related data at the current historical time point, x i Represents the historical target data / historical associated data at the previous historical time point with no missing historical target data and no missing historical associated data relative to the current historical time point (the historical time point corresponding to the missing data), x k Represents the historical target data / historical associated data at the next historical time point relative to the current historical time point (the historical time point corresponding to the missing data), at which neither the historical target data nor the historical associated data is missing. When determining the missing historical target data at the corresponding historical time point, the other data substituted is the other historical target data; when determining the missing historical associated data at the historical time point, the other data substituted is the other historical associated data.

[0119] It should be noted that the method for repairing missing data at historical time points can be determined by those skilled in the art based on actual conditions. The above description is only an example and does not constitute a limitation thereto.

[0120] Through the above steps, the actual historical information of multiple historical time points can be used as a parameter basis to process the initial target model to obtain multiple undetermined target models, so that the corresponding first coefficient, second coefficient, and third coefficient can be accurately determined by comprehensively solving the multiple undetermined target models, thereby improving the accuracy of the determined target model, thereby improving the calculation accuracy of the asset data association impact model, and thus improving the accuracy of the overall transaction abnormality alarm.

[0121] In an optional embodiment, if Figure 3 As shown, obtaining the association model in the asset data association impact model based on the target historical data information, the associated historical data information, the historical interest rate and the initial association model includes the following steps:

[0122] S301: Based on the historical associated data at multiple historical time points in the associated historical data information, obtain the historical associated data changes corresponding to the historical time points.

[0123] S302: Substitute the historical target data of multiple historical time points in the target historical data information, the historical associated data of multiple historical time points in the associated historical data information, the change in historical associated data, the historical interest rate and the historical time difference between the corresponding adjacent historical time points into the initial association model to obtain multiple undetermined association models.

[0124] S303: Solve the multiple undetermined association models to determine the corresponding fourth coefficient of the fourth data item, the fifth coefficient of the fifth data item, and the sixth coefficient of the sixth data item, and obtain the association model based on the fourth coefficient, the fifth coefficient, the sixth coefficient, and the initial association model.

[0125] Exemplarily, step S301 may be, but is not limited to, for each historical time point, subtracting the historical associated data of the current time point from the historical associated data of the next historical time point of the current historical time point to obtain the historical associated data change corresponding to the current historical time point. For example, if there is historical time point A (corresponding historical associated data is y1), historical time point B (corresponding historical associated data is y2), and historical time point C (corresponding historical associated data is y3), then the historical associated data change corresponding to historical time point A is y2-y1, and the historical associated data change corresponding to historical time point B is y3-y2. If there is no next historical time point after historical time point C, then historical time point C does not need to be assigned a historical associated data change (and the historical associated data corresponding to historical time point C does not need to be subsequently included in the corresponding solution operation). It should be noted that the specific implementation of step S301 can be determined by those skilled in the art based on actual conditions, and the above description is only an example and does not constitute a limitation thereto.

[0126] For example, a potential correlation model corresponds to a historical target data item at a time point, a historical correlation data item at a time point, a change in the historical correlation data item at a time point, and a historical time difference corresponding to a time point (i.e., the time difference between the next historical time point after the current historical time point and the current historical time point). In summary, a potential correlation model corresponds to a historical time point. It should be noted that the relevant correspondence can be determined by those skilled in the art based on actual circumstances, and the above description is merely an example and does not constitute a limitation.

[0127] Exemplarily, step S302 may include, but is not limited to, substituting the historical correlation data and historical interest rate into the fourth data item in the initial correlation model, substituting the historical correlation data change into the second data change data item in the initial correlation model, substituting the historical target data into the fifth data item in the initial correlation model, substituting the historical correlation data and historical target data into the third data item in the initial correlation model, and substituting the historical time difference into the time difference data item in the initial correlation model, to obtain an undetermined correlation model (at this point, the value of the corresponding item coefficient variable is still uncertain, i.e., the corresponding item coefficient variable is the variable that needs to be solved). It should be noted that the specific implementation of step S302 can be determined by those skilled in the art based on actual circumstances, and the above description is merely an example and does not constitute a limitation thereto.

[0128] Exemplarily, step S303 may be, but is not limited to, first grouping multiple undetermined association models to obtain multiple undetermined association model groups, wherein the number of undetermined association models in one undetermined association model group is the number of variables (term coefficient variables) that need to be solved in one of the undetermined association models (generally 3 in the embodiment of the present invention), and there may or may not be intersections between the multiple undetermined association model groups. Then, for each of the undetermined association model groups (when the relevant logic of the model can be embodied as an equation), an equation-solving operation is performed using methods such as the Euler method, the Runge-Kutta method, and other equation-solving methods to obtain the corresponding fourth coefficient to be processed, the fifth coefficient to be processed, and the sixth coefficient to be processed for each of the undetermined association model groups (one undetermined association model group corresponds to one fourth coefficient to be processed, one fifth coefficient to be processed, and one sixth coefficient to be processed). After that, the average value of the plurality of fourth coefficients to be processed is used as the fourth coefficient, the average value of the plurality of fifth coefficients to be processed is used as the fifth coefficient, and the average value of the plurality of sixth coefficients to be processed is used as the sixth coefficient. Finally, the fourth coefficient, the fifth coefficient, and the sixth coefficient are substituted into the initial association model to obtain the association model. It should be noted that the specific implementation of step S303 can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation thereto.

[0129] Through the above steps, the actual historical information of multiple historical time points can be used as a parameter basis to process the initial correlation model to obtain multiple undetermined correlation models, so that the corresponding fourth coefficient, fifth coefficient, and sixth coefficient can be accurately determined by comprehensively solving the multiple undetermined correlation models, thereby improving the accuracy of the determined relevant models, thereby improving the calculation accuracy of the asset data correlation impact model, and thus improving the accuracy of the overall transaction abnormality alarm.

[0130] In an optional embodiment, further comprising:

[0131] Before obtaining the pre-change future data of the target asset at multiple future time points based on the asset data association impact model corresponding to the target asset and the associated assets corresponding to the target asset, the current target data of the target asset, the current associated data of the associated assets, and the current interest rate,

[0132] Based on the verification target data of multiple verification time points of the target asset, the verification association data of multiple verification time points of the associated assets, the corresponding verification interest rates and the verification time differences of the corresponding adjacent verification time points, it is determined whether the asset data association impact model is applicable. If not, a model applicability alarm is issued.

[0133] Exemplarily, the verification time point is similar to the historical time point, both being time points in the past relative to the current time. However, the verification time point is generally different from the historical time point.

[0134] Exemplarily, the time differences between adjacent verification time points are fixed (i.e., the verification time points can be expressed as an arithmetic progression after being arranged in sequence. In layman's terms, the time differences corresponding to different groups of adjacent verification time points are all the same). For example, if there are verification time point A, verification time point B, verification time point C, and verification time point D, then the time difference between verification time point B and verification time point A, the time difference between verification time point C and verification time point B, and the time difference between verification time point D and verification time point C are all the same.

[0135] For example, one verification time point corresponds to one verification target data and one verification related data. It should be noted that the corresponding relationship can be determined by those skilled in the art according to actual conditions, and the above description is only an example and does not constitute a limitation thereto.

[0136] Exemplarily, the verification time difference of the verification time points may be, but is not limited to, a time difference between a verification time point next to the current verification time point and the current verification time point.

[0137] For example, historical target data, verification target data, historical correlation data, and verification correlation data can all be obtained or parsed from relevant asset information. It should be noted that the source of relevant historical data can be determined by those skilled in the art based on actual circumstances, and the above description is merely an example and does not constitute a limitation.

[0138] For example, the model suitability alert may include, but is not limited to, sending a warning message to relevant staff, such as "The accuracy of the target model and / or associated models in this model does not meet the corresponding business requirements. Please reset." It should be noted that the specific implementation of the model suitability alert can be determined by those skilled in the art based on actual circumstances. The above description is merely an example and does not constitute a limitation.

[0139] Preferably, if the asset data correlation impact model is determined to be inapplicable, a new data dynamics function can be selected to construct the corresponding initial target model and initial correlation model, thereby performing operations such as solving calculations to re-derive the target model and correlation model, thereby remodeling the asset data correlation impact model. This helps further improve the computational accuracy of the asset data correlation impact model, thereby further improving the accuracy of overall transaction anomaly alerts.

[0140] Through the above steps, the accuracy of the asset data association impact model can be tested using actual relevant data, which is conducive to timely discovery when the model's applicability is poor and prevents it from participating in subsequent applications. This can help ensure that the accuracy of the asset data association impact model that is subsequently involved in production and application meets the corresponding requirements and is applicable, thereby indirectly improving the accuracy of the overall transaction abnormality alarm.

[0141] In an optional embodiment, the determining whether the asset data association impact model is applicable based on the verification target data of the target asset at multiple verification time points, the verification association data of the associated assets at multiple verification time points, the corresponding verification interest rates, and the verification time differences of the corresponding adjacent verification time points includes:

[0142] Input the corresponding verification time difference, verification rate, verification target data corresponding to the earliest verification time point, and verification association data into the target model and association model in the asset data association impact model for cross-iteration calculation to obtain test output data of multiple verification time points of the target asset;

[0143] Determining error rates corresponding to a plurality of the verification time points based on the corresponding verification target data and test output data;

[0144] Determine whether there is no verification time point whose corresponding error rate is greater than or equal to the preset error rate threshold; if not, issue a model applicability alarm.

[0145] Exemplarily, the corresponding verification time difference, verification rate, verification target data corresponding to the earliest verification time point, and verification association data are input into the target model and association model in the asset data association influence model for cross-iteration calculation to obtain test output data for multiple verification time points of the target asset, which may be, but is not limited to:

[0146] The verification target data corresponding to the earliest verification time point is used as the current target data to be verified, and the verification associated data corresponding to the earliest verification time point is used as the current associated data to be verified;

[0147] The value of the time difference data item in the target model and the value of the time difference data item in the associated model are both set as the verification time difference, and the value of the interest rate variable in the first data item in the target model and the value of the interest rate variable in the fourth data item in the associated model are both set as the verification interest rate;

[0148] The earliest verification time point is used as the current verification time point;

[0149] Repeating the first cross-iteration step until the number of times the first cross-iteration step is performed reaches a value obtained by subtracting 1 from the number of verification time points; wherein the first cross-iteration step includes:

[0150] Input the current target data to be verified into the first data item and the third data item of the target model, and input the current associated data to be verified into the second data item and the third data item of the target model, perform operations, and output the target data to be verified at the next verification time point corresponding to the current verification time point;

[0151] Input the current associated data to be verified into the fourth data item and the sixth data item of the association model, and input the current target data to be verified into the fifth data item and the sixth data item of the association model, perform operations, and output the associated data to be verified at a verification time point next to the current verification time point;

[0152] The target data to be verified at the next verification time point corresponding to the current verification time point is used as the current target data to be verified, and the associated data to be verified at the next verification time point corresponding to the current verification time point is used as the current associated data to be verified;

[0153] The next checkpoint after the current checkpoint is used as the updated current checkpoint.

[0154] After repeatedly executing the first cross-iteration step, the target data to be verified at multiple verification time points is used as the test output data.

[0155] Furthermore, the corresponding verification time difference, verification rate, verification target data corresponding to the earliest verification time point, and verification association data are input into the target model and association model in the asset data association impact model for cross-iteration calculation to obtain test output data of multiple verification time points of the target asset. An example is as follows:

[0156] It is known that the verification time points include verification time point A, verification time point B, verification time point C and verification time point D, and verification time point A is the earliest verification time point, the verification target data corresponding to verification time point A is a (which is also the target data to be verified corresponding to verification time point A), and the verification associated data is b (which is also the associated data to be verified corresponding to verification time point A); at this time, the value of the time difference data item in the target model and the value of the time difference data item in the associated model have been determined to be the verification time difference q corresponding to verification time point A, verification time point B, verification time point C and verification time point D, and the value of the interest rate variable in the first data item in the target model and the value of the interest rate variable in the fourth data item in the associated model are both set to the corresponding verification interest rate p.

[0157] Substitute the target data a to be verified and the associated data b to be verified corresponding to verification time point A into the target model for calculation, and obtain the target data c to be verified corresponding to verification time point B. Substitute the target data a to be verified and the associated data b to be verified corresponding to verification time point A into the association model for calculation, and obtain the associated data d to be verified corresponding to verification time point B.

[0158] Substitute the target data c to be verified and the associated data d to be verified corresponding to verification time point B into the target model for calculation, and obtain the target data e to be verified corresponding to verification time point C. Substitute the target data c to be verified and the associated data d to be verified corresponding to verification time point B into the association model for calculation, and obtain the associated data f to be verified corresponding to verification time point C.

[0159] Substitute the target data e to be verified and the associated data f to be verified corresponding to the verification time point C into the target model for calculation, and obtain the target data g to be verified corresponding to the verification time point D. Substitute the target data e to be verified and the associated data f to be verified corresponding to the verification time point C into the association model for calculation, and obtain the associated data h to be verified corresponding to the verification time point D.

[0160] In this way, the test output data includes the target data a to be verified corresponding to the verification time point A, the target data c to be verified corresponding to the verification time point B, the target data e to be verified corresponding to the verification time point C, and the target data g to be verified corresponding to the verification time point D.

[0161] Furthermore, the current target data to be verified is input into the first data item and the third data item of the target model, and the current associated data to be verified is input into the second data item and the third data item of the target model, and operations are performed to obtain the target data to be verified at the next verification time point corresponding to the current verification time point. This can be done by inputting the current target data to be verified into the first data item and the third data item of the target model, and inputting the current associated data to be verified into the second data item and the third data item of the target model to obtain target verification fluctuation amplitude, and superimposing the target verification amplitude on the current target data to be verified to obtain the associated data to be verified at the next verification time point corresponding to the current verification time point.

[0162] Correspondingly, the current associated data to be verified is input into the fourth data item and the sixth data item of the associated model, and the current target data to be verified is input into the fifth data item and the sixth data item of the associated model, and calculation is performed to obtain the associated data to be verified at the next verification time point corresponding to the current verification time point. It can be that the current associated data to be verified is input into the fourth data item and the sixth data item of the associated model, and the current target data to be verified is input into the fifth data item and the sixth data item of the associated model to obtain the associated verification fluctuation amplitude, and the associated verification wave amplitude is superimposed on the current associated data to be verified to obtain the associated data to be verified at the next verification time point corresponding to the current verification time point.

[0163] It should be noted that the specific implementation method of inputting the corresponding verification time difference, verification interest rate, the verification target data corresponding to the earliest verification time point, and verification association data into the target model and the association model in the asset data association influence model for cross-iteration operation to obtain the test output data of multiple verification time points of the target asset can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0164] Exemplarily, the error rates corresponding to the plurality of verification time points are determined based on the corresponding verification target data and test output data. This may be, but is not limited to, dividing the absolute value of the difference between the verification target data and the test output data corresponding to each verification time point by the corresponding verification target data or test output data to obtain the corresponding error rate. One verification time point corresponds to one error rate. It should be noted that the specific implementation of determining the error rates corresponding to the plurality of verification time points based on the corresponding verification target data and test output data can be determined by those skilled in the art based on actual circumstances, and the above description is for illustrative purposes only and does not constitute a limitation thereto.

[0165] Exemplarily, the preset error rate threshold may be determined by those skilled in the art according to actual conditions, and the embodiment of the present invention does not limit this. For example, the preset error rate threshold may be, but is not limited to, 5%.

[0166] Through the above steps, the target model and the associated model in the asset data association model can be subjected to a more in-depth and detailed accuracy test, which greatly improves the accuracy of the accuracy test. This is more conducive to ensuring that the accuracy of the asset data association impact model involved in subsequent production and use meets the corresponding requirements and is applicable, and can further indirectly improve the accuracy of the overall transaction abnormality alarm.

[0167] In an optional embodiment, obtaining the pre-change future data of the target asset at multiple future time points based on the asset data association impact model corresponding to the target asset and the associated assets corresponding to the target asset, the current target data of the target asset, the current association data of the associated assets, and the current interest rate includes:

[0168] The preset predicted time difference, the current target data, the current associated data and the current interest rate are input into the target model and the associated model in the asset data associated impact model for cross-iterative calculation to obtain the future data of the target asset before the change at multiple future time points.

[0169] Preferably, the target model can be replaced by a trained target dynamic neural network model, and the associated model can be replaced by a trained associated dynamic neural network model. Correspondingly, before formal use, the historical target data, historical associated data, and corresponding historical interest rates at multiple corresponding historical time points can be used as input samples of the target dynamic neural network model, and the historical target data changes corresponding to the input samples of the target dynamic neural network model can be used as input samples of the target dynamic neural network model, and the input samples and output samples of the target dynamic neural network model can be used to train the target dynamic neural network model. Correspondingly, the historical target data, historical associated data, and corresponding historical interest rates at multiple corresponding historical time points can be used as input samples of the associated dynamic neural network model, and the historical associated data changes corresponding to the input samples of the associated dynamic neural network model can be used as output samples of the associated dynamic neural network model, and the input samples and output samples of the associated dynamic neural network model can be used to train the associated dynamic neural network model.

[0170] Exemplarily, one future time point corresponds to one future data before the change, and one target asset corresponds to multiple future data before the change at multiple future time points.

[0171] For example, the predicted time difference can be determined by those skilled in the art based on actual conditions, and the present invention does not impose any limitation thereto. For example, the predicted time difference can be, but is not limited to, 1 day. The dimension of the predicted time difference must correspond to the dimension of the current interest rate. For example, if the predicted time difference is 1 day, the current interest rate must be the current daily interest rate.

[0172] For example, the number of the multiple future time points can be determined by those skilled in the art based on actual conditions, and the embodiments of the present invention do not limit this. For example, the number of the multiple future time points can be, but is not limited to, 365, 366, 10, or 30.

[0173] For example, the time differences between adjacent future time points can be fixed (i.e., the future time points can be represented as an arithmetic progression when arranged in sequence. In layman's terms, the time differences corresponding to different groups of adjacent future time points are the same). For example, if there are future time points A, B, C, and D, then the time difference between future time points B and A, the time difference between future time points C and B, and the time difference between future time points D and C are all the same.

[0174] Exemplarily, the inputting of the preset predicted time difference, the current target data, the current associated data, and the current interest rate into the target model and the associated model in the asset data associated impact model to perform cross-iterative operations to obtain the future data of the target asset before the change at multiple future time points may be, but is not limited to:

[0175] Setting the value of the time difference data item in the target model and the value of the time difference data item in the associated model to the predicted time difference, and setting the value of the interest rate variable in the first data item in the target model and the value of the interest rate variable in the fourth data item in the associated model to the current interest rate;

[0176] Superimposing the predicted time difference on the current time to obtain the earliest future time point;

[0177] Input the current target data into the first data item and the third data item of the target model, and input the current associated data into the second data item and the third data item of the target model, perform calculations to obtain a target predicted fluctuation range, superimpose the target predicted fluctuation range on the current target data to obtain pre-change future data corresponding to the earliest future time point, and use the pre-change future data corresponding to the earliest future time point as the current pre-change future data;

[0178] Inputting the current correlation data into the fourth data item and the sixth data item of the correlation model, and inputting the current target data into the fifth data item and the sixth data item of the correlation model, performing calculations to obtain a correlation prediction fluctuation amplitude, superimposing the current correlation data with the correlation prediction fluctuation amplitude to obtain pre-change future correlation data corresponding to the earliest future time point, and using the pre-change future correlation data corresponding to the earliest future time point as the current pre-change future correlation data;

[0179] Taking the earliest of said future time points as the present future time point;

[0180] Repeating the second cross-iteration step until the number of times the second cross-iteration step is performed reaches a value obtained by subtracting 1 from the preset number of future time points; wherein the second cross-iteration step includes:

[0181] Input the current future data before the change into the first data item and the third data item of the target model, and input the current future associated data before the change into the second data item and the third data item of the target model, perform calculations to obtain the current target predicted fluctuation range, and superimpose the current future data before the change on the current target predicted fluctuation range to obtain output future data before the change corresponding to the next future time point of the current future time point;

[0182] Inputting the current pre-change future correlation data into the fourth data item and the sixth data item of the correlation model, and inputting the current pre-change future correlation data into the fifth data item and the sixth data item of the correlation model, performing calculations to obtain the current correlation predicted fluctuation amplitude, superimposing the current pre-change future correlation data on the current correlation predicted fluctuation amplitude, and outputting the pre-change future correlation data corresponding to the next future time point of the current future time point; wherein the next future time point of the current future time point is obtained by superimposing the predicted time difference on the current future time point;

[0183] The outputted future data before the change at the next future time point corresponding to the current future time point is used as the current future data before the change, and the outputted future associated data before the change at the next future time point corresponding to the current future time point is used as the current future associated data before the change;

[0184] Taking the next future time point after the current future time point as the updated current future time point;

[0185] After repeatedly executing the second cross-iteration step, the pre-change future data of the target asset at multiple future time points are obtained.

[0186] Furthermore, the preset predicted time difference, the current target data, the current associated data and the current interest rate are input into the target model and the associated model in the asset data associated impact model to perform cross-iterative operations to obtain the future data of the target asset before the change at multiple future time points. The following examples are provided:

[0187] The known future time points include future time point A, future time point B, future time point C, and future time point D, with future time point A being the earliest future time point (which can be determined by adding the predicted time difference to the current time). At this time, the values ​​of the time difference data items in the target model and the values ​​of the time difference data items in the associated model are both determined to be the predicted time differences q corresponding to future time points A, B, C, and D. The value of the interest rate variable in the first data item in the target model and the value of the interest rate variable in the fourth data item in the associated model are both set to the corresponding current interest rate p.

[0188] The pre-change future data (predicted data corresponding to the target asset) corresponding to future time point A has been initially calculated to be a, and the pre-change future associated data (predicted data corresponding to the associated asset) is b;

[0189] Substitute the pre-change future data a and pre-change future associated data b corresponding to future time point A into the target model for calculation, and obtain the pre-change future data c corresponding to future time point B. Substitute the pre-change future data a and pre-change future associated data b corresponding to future time point A into the association model for calculation, and obtain the pre-change future associated data d corresponding to future time point B.

[0190] Substitute the pre-change future data c and pre-change future associated data d corresponding to future time point B into the target model for calculation, and obtain the pre-change future data e corresponding to future time point C. Substitute the pre-change future data c and pre-change future associated data d corresponding to future time point B into the association model for calculation, and obtain the pre-change future associated data f corresponding to future time point C.

[0191] Substitute the pre-change future data e and pre-change future associated data f corresponding to future time point C into the target model for calculation, and obtain the pre-change future data g corresponding to future time point D. Substitute the pre-change future data e and pre-change future associated data f corresponding to future time point C into the association model for calculation, and obtain the pre-change future associated data h corresponding to future time point D.

[0192] In the above example, the description of the superimposed corresponding predicted fluctuation range is omitted.

[0193] In this way, the future data before the change of multiple future time points include the future data before the change a corresponding to future time point A, the future data before the change c corresponding to future time point B, the future data before the change e corresponding to future time point C, and the future data before the change g corresponding to future time point D.

[0194] It should be noted that the specific implementation method of inputting the preset prediction time difference, the current target data, the current associated data and the current interest rate into the target model and the associated model in the asset data association influence model for cross-iteration calculation to obtain the future data of the target asset before the change at multiple future time points can be determined by those skilled in the art based on actual conditions. The above description is only an example and does not constitute a limitation to this.

[0195] Through the above steps, the preset prediction time difference, the current target data, the current related data and the current interest rate can be fully used as inputs, and the asset data association impact model can be used for more detailed and in-depth calculation and processing. The calculation process can be better combined with the influence of related assets in a cross-iteration manner, thereby greatly improving the accuracy of future data before the prediction change, and further greatly improving the accuracy of the overall transaction abnormality alarm.

[0196] In an optional embodiment, obtaining the changed future data of the target asset at the plurality of future time points based on the expected interest rate change range, the asset data association impact model, the current target data, the current association data, and the current interest rate includes:

[0197] Based on the interest rate change range and the current interest rate, obtain the expected interest rate after the change;

[0198] The predicted time difference, the current target data, the current associated data and the expected interest rate after the change are input into the target model and the associated model in the asset data association impact model for cross-iteration calculation to obtain the future data of the target asset after the change at multiple future time points.

[0199] Exemplarily, the method for obtaining the expected interest rate after the change based on the interest rate fluctuation range and the current interest rate may be, but is not limited to, adding the interest rate fluctuation range to the current interest rate to obtain the expected interest rate after the change. The interest rate fluctuation range may be, but is not limited to, 0.0001, 0.0002, -0.0001 or -0.0002, etc. The specific range may be determined by those skilled in the art based on actual conditions. It should be noted that the specific implementation method for obtaining the expected interest rate after the change based on the interest rate fluctuation range and the current interest rate may be, but is not limited to, adding the interest rate fluctuation range to the current interest rate to obtain the expected interest rate after the change, and the specific value of the interest rate fluctuation range may be determined by those skilled in the art based on actual conditions. The above description is for example only and does not constitute a limitation thereto.

[0200] Exemplarily, the predicted time difference, the current target data, the current associated data, and the expected interest rate after change are input into the target model and the associated model in the asset data associated impact model to perform cross-iterative operations to obtain the future data after change of the target asset at multiple future time points, which may be, but is not limited to:

[0201] Setting the value of the time difference data item in the target model and the value of the time difference data item in the associated model to the predicted time difference, and setting the value of the interest rate variable in the first data item in the target model and the value of the interest rate variable in the fourth data item in the associated model to the changed interest rate;

[0202] Superimposing the predicted time difference on the current time to obtain the earliest future time point;

[0203] Input the current target data into the first data item and the third data item of the target model, and input the current associated data into the second data item and the third data item of the target model, perform calculations to obtain the predicted fluctuation range of the target after the change, superimpose the predicted fluctuation range of the target after the change on the current target data, and obtain the changed future data corresponding to the earliest future time point, and use the changed future data corresponding to the earliest future time point as the current changed future data;

[0204] Inputting the current association data into the fourth data item and the sixth data item of the association model, and inputting the current target data into the fifth data item and the sixth data item of the association model, respectively, performing calculations to obtain a predicted fluctuation range of the association after the change; superimposing the predicted fluctuation range of the association after the change on the current association data to obtain the future association data after the change corresponding to the earliest future time point; and using the future association data after the change corresponding to the earliest future time point as the current future association data after the change;

[0205] Taking the earliest of said future time points as the present future time point;

[0206] Repeating the third cross-iteration step until the number of times the third cross-iteration step is performed reaches a value obtained by subtracting 1 from the preset number of future time points; wherein the third cross-iteration step includes:

[0207] Input the current changed future data into the first data item and the third data item of the target model, and input the current changed future associated data into the second data item and the third data item of the target model, perform calculations to obtain the current changed target predicted fluctuation range, superimpose the current changed future data on the current changed target predicted fluctuation range, and obtain the output changed future data corresponding to the next future time point of the current future time point;

[0208] Inputting the current changed future associated data into the fourth data item and the sixth data item of the association model, and inputting the current changed future data into the fifth data item and the sixth data item of the association model, performing calculations to obtain the current changed future associated predicted fluctuation amplitude, superimposing the current changed future associated data on the current changed future associated fluctuation amplitude, and outputting the changed future associated data corresponding to the next future time point of the current future time point; wherein the next future time point of the current future time point is obtained by superimposing the predicted time difference on the current future time point;

[0209] The outputted changed future data at the next future time point corresponding to the current future time point is used as the current changed future data, and the outputted changed future associated data at the next future time point corresponding to the current future time point is used as the current changed future associated data;

[0210] Taking the next future time point after the current future time point as the updated current future time point;

[0211] After repeatedly executing the third cross-iteration step, the changed future data of the target asset at multiple future time points are obtained.

[0212] Furthermore, the predicted time difference, the current target data, the current associated data, and the expected interest rate after change are input into the target model and the associated model in the asset data associated impact model for cross-iteration calculation to obtain the target asset's future data after change at multiple future time points. Examples are as follows:

[0213] The known future time points include future time point A, future time point B, future time point C, and future time point D, with future time point A being the earliest future time point (which can be determined by adding the predicted time difference to the current time). At this time, the values ​​of the time difference data items in the target model and the values ​​of the time difference data items in the associated model are both determined to be the predicted time differences q corresponding to future time points A, B, C, and D. The value of the interest rate variable in the first data item in the target model and the value of the interest rate variable in the fourth data item in the associated model are both set to the corresponding changed interest rate p.

[0214] The changed future data (the predicted data corresponding to the target asset) corresponding to the future time point A has been initially calculated to be a, and the changed future associated data (the predicted data corresponding to the associated asset) is b;

[0215] Substitute the changed future data a and the changed future associated data b corresponding to future time point A into the target model for calculation, and obtain the changed future data c corresponding to future time point B. Substitute the changed future data a and the changed future associated data b corresponding to future time point A into the association model for calculation, and obtain the changed future associated data d corresponding to future time point B.

[0216] Substitute the changed future data c and the changed future associated data d corresponding to future time point B into the target model for calculation, and obtain the changed future data e corresponding to future time point C. Substitute the changed future data c and the changed future associated data d corresponding to future time point B into the association model for calculation, and obtain the changed future associated data f corresponding to future time point C.

[0217] Substitute the changed future data e and the changed future associated data f corresponding to future time point C into the target model for calculation, and obtain the changed future data g corresponding to future time point D. Substitute the changed future data e and the changed future associated data f corresponding to future time point C into the association model for calculation, and obtain the changed future associated data h corresponding to future time point D.

[0218] In the above example, the description of the superimposed corresponding predicted fluctuation range is omitted.

[0219] In this way, the changed future data of multiple future time points include the changed future data a corresponding to future time point A, the changed future data c corresponding to future time point B, the changed future data e corresponding to future time point C, and the changed future data g corresponding to future time point D.

[0220] It should be noted that the specific implementation method for inputting the predicted time difference, the current target data, the current associated data and the expected interest rate after the change into the target model and the associated model in the asset data association impact model for cross-iteration operation to obtain the future data of the target asset after the change at multiple future time points can be determined by those skilled in the art based on actual conditions. The above description is only an example and does not constitute a limitation to this.

[0221] Through the above steps, the preset prediction time difference, the current target data, the current related data and the interest rate after the change can be fully used as input, and the asset data association impact model can be used for more detailed and in-depth calculation and processing. The calculation process can be better combined with the influence of related assets by means of cross-iteration, thereby greatly improving the accuracy of the predicted future data after the change, and further greatly improving the accuracy of the overall transaction abnormality alarm.

[0222] In an optional embodiment, if Figure 4 As shown, the process of determining whether there is a transaction anomaly in the target asset based on the pre-change future data and the post-change future data corresponding to the multiple future time points includes the following steps:

[0223] S401: Obtaining a target data change range corresponding to the future time point based on the corresponding future data before the change and the future data after the change.

[0224] S402: Determine whether there is a future time point where the corresponding target data has a change range greater than or equal to a preset data change range threshold. If so, issue a transaction abnormality alarm.

[0225] Exemplarily, step S401 may include, but is not limited to, using the absolute value of the difference between the future data after the change and the future data before the change as the corresponding target data change range. Here, one future time point corresponds to one future data before the change, one future data after the change, and one target data change range. It should be noted that the specific implementation of step S401 can be determined by those skilled in the art based on actual circumstances, and the above description is for illustrative purposes only and does not constitute a limitation thereto.

[0226] For example, the data change amplitude threshold can be determined by those skilled in the art according to actual conditions, and the embodiments of the present invention do not limit this. For example, the data change amplitude threshold can be, but is not limited to, 100, 500, 1000, 5000, or 10000, etc.

[0227] Preferably, the process of repeatedly obtaining a new changed interest rate based on the changed interest rate and the expected interest rate change range can be repeated, thereby using the new changed interest rate to determine new changed future data for multiple future time points, and judging whether there is any transaction anomaly in the target asset based on the new changed future data and the corresponding pre-change future data or the last changed future data, so as to further improve the comprehensiveness of the transaction anomaly alert.

[0228] Through the above steps, it is possible to more finely and accurately determine the fluctuations of target asset data due to interest rate changes and related assets at different future times based on the corresponding data changes, thereby improving the accuracy of transaction abnormality alerts.

[0229] In an optional embodiment, if Figure 5 As shown, the process of determining whether there is a transaction anomaly in the target asset based on the pre-change future data and the post-change future data corresponding to the multiple future time points includes the following steps:

[0230] S501: Based on the corresponding future data before the change and the future data after the change, obtain the target data change range corresponding to the future time point.

[0231] S502: Obtaining a target data change rate corresponding to the future time point based on the target data change amplitude and the future data before the change or the future data after the change.

[0232] S503: Determine whether there is a future time point where the corresponding target data change rate is greater than or equal to a preset data change rate threshold. If so, issue a transaction abnormality alarm.

[0233] For example, the specific implementation of step S501 can refer to the description of step S401 in the embodiment of the present invention, which will not be repeated here.

[0234] Exemplarily, step S502 may be, but is not limited to, dividing the target data change amplitude by the future data before or after the change to obtain the target data change rate corresponding to the future time point. Each future time point corresponds to one target data change rate. It should be noted that the specific implementation of step S502 can be determined by those skilled in the art based on actual circumstances, and the above description is for illustrative purposes only and does not constitute a limitation thereto.

[0235] For example, the data change rate threshold can be determined by those skilled in the art according to actual conditions, and the present invention does not limit this. For example, the data change rate threshold can be, but is not limited to, 5%, 4%, or 6%.

[0236] Through the above steps, it is possible to more finely and accurately determine the fluctuations of target asset data due to interest rate changes and related assets at different future times based on the corresponding data change rate, thereby improving the accuracy of transaction abnormality alerts.

[0237] Based on the same principle, the embodiment of the present invention discloses a transaction abnormality alarm device 600, such as Figure 6 As shown, the transaction abnormality alarm device 600 includes:

[0238] A first prediction module 601 is configured to obtain the target asset's pre-change future data at multiple future time points based on an asset data association impact model corresponding to a target asset and its associated assets, the target asset's current target data, the associated assets' current associated data, and a current interest rate.

[0239] A second prediction module 602 is configured to obtain the target asset's future data after changes at multiple future time points based on the expected interest rate change range, the asset data association impact model, the current target data, the current association data, and the current interest rate;

[0240] The abnormality alarm module 603 is used to determine whether there is a transaction abnormality in the target asset based on the future data before the change and the future data after the change corresponding to the multiple future time points, and if so, issue a transaction abnormality alarm.

[0241] In an optional embodiment, the system further includes a model building module for:

[0242] Before obtaining the pre-change future data of the target asset at multiple future time points based on the asset data association impact model corresponding to the target asset and the associated assets corresponding to the target asset, the current target data of the target asset, the current association data of the associated assets, and the current interest rate, an initial target model of the target asset and initial association models of the associated assets corresponding to the target asset are constructed based on a preset data dynamic function;

[0243] Obtaining a target model in the asset data association impact model based on the target historical data information of the target asset, the associated historical data information of the associated assets, the corresponding historical interest rates, and the initial target model;

[0244] Based on the target historical data information, the associated historical data information, the historical interest rate and the initial association model, an association model in the asset data association impact model is obtained.

[0245] In an optional embodiment, the model building module is used to:

[0246] Based on the plurality of data dynamic functions, respectively constructing a first data item corresponding to the impact of the target asset itself on the target asset data, a second data item corresponding to the impact of the associated asset on the target asset data, a third data item corresponding to the impact of the associated asset and the target asset on the target asset data, a fourth data item corresponding to the impact of the associated asset itself on the associated asset data, a fifth data item corresponding to the impact of the target asset on the associated asset data, and a sixth data item corresponding to the impact of the associated asset and the target asset on the associated asset data;

[0247] Constructing the initial target model based on the preset first data change amount data item, the time difference data item, the first data item, the second data item and the third data item;

[0248] The initial association model is constructed based on the preset second data variation data item, the time difference data item, the fourth data item, the fifth data item and the sixth data item.

[0249] In an optional embodiment, the model building module is used to:

[0250] Based on the historical target data of multiple historical time points in the target historical data information, obtaining the historical target data changes corresponding to the historical time points;

[0251] Substituting historical target data at multiple historical time points in the target historical data information, historical target data changes, historical associated data at multiple historical time points in the associated historical data information, the historical interest rate, and historical time differences between corresponding adjacent historical time points into the initial target model to obtain multiple undetermined target models;

[0252] Solve the multiple undetermined target models to determine the corresponding first coefficient of the first data item, the second coefficient of the second data item and the third coefficient of the third data item, and obtain the target model based on the first coefficient, the second coefficient, the third coefficient and the initial target model.

[0253] In an optional embodiment, the model building module is used to:

[0254] Based on the historical associated data of multiple historical time points in the associated historical data information, obtaining the historical associated data changes corresponding to the historical time points;

[0255] Substituting historical target data at multiple historical time points in the target historical data information, historical associated data at multiple historical time points in the associated historical data information, changes in historical associated data, the historical interest rate, and historical time differences corresponding to adjacent historical time points into the initial association model to obtain multiple undetermined association models;

[0256] Solve the multiple undetermined association models to determine the corresponding fourth coefficient of the fourth data item, the fifth coefficient of the fifth data item, and the sixth coefficient of the sixth data item, and obtain the association model based on the fourth coefficient, the fifth coefficient, the sixth coefficient, and the initial association model.

[0257] In an optional embodiment, the system further includes a model checking module for:

[0258] Before obtaining the pre-change future data of the target asset at multiple future time points based on the asset data association impact model corresponding to the target asset and the associated assets corresponding to the target asset, the current target data of the target asset, the current associated data of the associated assets, and the current interest rate,

[0259] Based on the verification target data of multiple verification time points of the target asset, the verification association data of multiple verification time points of the associated assets, the corresponding verification interest rates and the verification time differences of the corresponding adjacent verification time points, it is determined whether the asset data association impact model is applicable. If not, a model applicability alarm is issued.

[0260] In an optional embodiment, the model checking module is used to:

[0261] Input the corresponding verification time difference, verification rate, verification target data corresponding to the earliest verification time point, and verification association data into the target model and association model in the asset data association impact model for cross-iteration calculation to obtain test output data of multiple verification time points of the target asset;

[0262] Determining error rates corresponding to a plurality of the verification time points based on the corresponding verification target data and test output data;

[0263] Determine whether there is no verification time point whose corresponding error rate is greater than or equal to the preset error rate threshold; if not, issue a model applicability alarm.

[0264] In an optional embodiment, the first prediction module 601 is configured to:

[0265] The preset predicted time difference, the current target data, the current associated data and the current interest rate are input into the target model and the associated model in the asset data associated impact model for cross-iterative calculation to obtain the future data of the target asset before the change at multiple future time points.

[0266] In an optional embodiment, the second prediction module 602 is configured to:

[0267] Based on the interest rate change range and the current interest rate, obtain the expected interest rate after the change;

[0268] The predicted time difference, the current target data, the current associated data and the expected interest rate after the change are input into the target model and the associated model in the asset data association impact model for cross-iteration calculation to obtain the future data of the target asset after the change at multiple future time points.

[0269] In an optional implementation, the abnormality alarm module 603 is configured to:

[0270] Obtaining a target data change range corresponding to the future time point based on the corresponding future data before and after the change;

[0271] Determine whether there is a future time point where the corresponding target data change amplitude is greater than or equal to a preset data change amplitude threshold. If so, issue a transaction abnormality alarm.

[0272] In an optional implementation, the abnormality alarm module 603 is configured to:

[0273] Obtaining a target data change range corresponding to the future time point based on the corresponding future data before and after the change;

[0274] Obtaining a target data change rate corresponding to the future time point based on the target data change amplitude and the future data before the change or the future data after the change;

[0275] Determine whether there is a future time point where the corresponding target data change rate is greater than or equal to a preset data change rate threshold. If so, issue a transaction abnormality alarm.

[0276] Since the principle of solving the problem by the transaction abnormality alarm device 600 is similar to that of the above method, the implementation of the transaction abnormality alarm device 600 can refer to the implementation of the above method and will not be repeated here.

[0277] The systems, devices, modules, or units described in the above embodiments may be implemented by computer chips or entities, or by products having certain functions. A typical implementation device is a computer device. Specifically, the computer device may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0278] In a typical example, a computer device specifically includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the method described above is implemented.

[0279] Reference below Figure 7 , which shows a schematic structural diagram of a computer device 700 suitable for implementing an embodiment of the present application.

[0280] like Figure 7 As shown, computer device 700 includes a central processing unit (CPU) 701, which can perform various appropriate tasks and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage portion 708 into a random access memory (RAM) 703. Various programs and data required for the operation of system 700 are also stored in RAM 703. CPU 701, ROM 702, and RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to bus 704.

[0281] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, a mouse, and the like; an output section 707 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and a speaker; a storage section 708 including devices such as a hard disk; and a communication section 709 including a network interface card such as a LAN card or a modem. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. Removable media 711, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 710 as needed, so that computer programs read therefrom can be installed in the storage section 708 as needed.

[0282] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program including program code for executing the methods illustrated in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 709 and / or installed from removable media 711.

[0283] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0284] For the convenience of description, the above devices are described as being divided into various units according to their functions. Of course, when implementing this application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0285] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0286] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0287] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0288] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "includes a..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0289] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0290] The present application may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communications network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0291] The various embodiments in this specification are described in a progressive manner. Similar parts between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the system embodiments are generally similar to the method embodiments, so the description is relatively simple. For relevant parts, refer to the description of the method embodiments.

[0292] The foregoing is merely an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A transaction abnormality alarm method, characterized in that: include: Obtaining future data of the target asset before changes at multiple future time points based on an asset data association impact model corresponding to a target asset and associated assets corresponding to the target asset, current target data of the target asset, current associated data of the associated assets, and a current interest rate; Obtaining the target asset's future data after changes at the plurality of future time points based on the expected interest rate change range, the asset data association impact model, the current target data, the current association data, and the current interest rate; Based on the future data before and after the change corresponding to the multiple future time points, determining whether there is a transaction anomaly for the target asset, and if so, issuing a transaction anomaly alarm; The transaction abnormality alarm method further includes: Before obtaining the pre-change future data of the target asset at multiple future time points based on the asset data association impact model corresponding to the target asset and the associated assets corresponding to the target asset, the current target data of the target asset, the current association data of the associated assets, and the current interest rate, an initial target model of the target asset and initial association models of the associated assets corresponding to the target asset are constructed based on a preset data dynamic function; Obtaining a target model in the asset data association impact model based on the target historical data information of the target asset, the associated historical data information of the associated assets, the corresponding historical interest rates, and the initial target model; Based on the target historical data information, the associated historical data information, the historical interest rate and the initial association model, an association model in the asset data association impact model is obtained.

2. The method according to claim 1, characterized in that The step of constructing an initial target model of the target asset and an initial association model of the associated assets corresponding to the target asset based on a preset data dynamic function includes: Based on the plurality of data dynamic functions, respectively constructing a first data item corresponding to the impact of the target asset itself on the target asset data, a second data item corresponding to the impact of the associated asset on the target asset data, a third data item corresponding to the impact of the associated asset and the target asset on the target asset data, a fourth data item corresponding to the impact of the associated asset itself on the associated asset data, a fifth data item corresponding to the impact of the target asset on the associated asset data, and a sixth data item corresponding to the impact of the associated asset and the target asset on the associated asset data; Constructing the initial target model based on the preset first data change amount data item, the time difference data item, the first data item, the second data item and the third data item; The initial association model is constructed based on the preset second data variation data item, the time difference data item, the fourth data item, the fifth data item and the sixth data item.

3. The method according to claim 2, characterized in that The step of obtaining a target model in the asset data association impact model based on the target historical data information of the target asset, the associated historical data information of the associated assets, the corresponding historical interest rate, and the initial target model includes: Based on the historical target data of multiple historical time points in the target historical data information, obtaining the historical target data changes corresponding to the historical time points; Substituting historical target data at multiple historical time points in the target historical data information, historical target data changes, historical associated data at multiple historical time points in the associated historical data information, the historical interest rate, and historical time differences between corresponding adjacent historical time points into the initial target model to obtain multiple undetermined target models; Solve the multiple undetermined target models to determine the corresponding first coefficient of the first data item, the second coefficient of the second data item and the third coefficient of the third data item, and obtain the target model based on the first coefficient, the second coefficient, the third coefficient and the initial target model.

4. The method according to claim 2, characterized in that The obtaining of the association model in the asset data association impact model based on the target historical data information, the associated historical data information, the historical interest rate and the initial association model includes: Based on the historical associated data of multiple historical time points in the associated historical data information, obtaining the historical associated data changes corresponding to the historical time points; Substituting historical target data at multiple historical time points in the target historical data information, historical associated data at multiple historical time points in the associated historical data information, changes in historical associated data, the historical interest rate, and historical time differences corresponding to adjacent historical time points into the initial association model to obtain multiple undetermined association models; Solve the multiple undetermined association models to determine the corresponding fourth coefficient of the fourth data item, the fifth coefficient of the fifth data item, and the sixth coefficient of the sixth data item, and obtain the association model based on the fourth coefficient, the fifth coefficient, the sixth coefficient, and the initial association model.

5. The method according to claim 1, wherein Further including: Before obtaining the pre-change future data of the target asset at multiple future time points based on the asset data association impact model corresponding to the target asset and the associated assets corresponding to the target asset, the current target data of the target asset, the current associated data of the associated assets, and the current interest rate, Based on the verification target data of multiple verification time points of the target asset, the verification association data of multiple verification time points of the associated assets, the corresponding verification interest rates and the verification time differences of the corresponding adjacent verification time points, it is determined whether the asset data association impact model is applicable. If not, a model applicability alarm is issued.

6. The method according to claim 5, characterized in that The determining whether the asset data association impact model is applicable based on the verification target data of the target asset at multiple verification time points, the verification association data of the associated assets at multiple verification time points, the corresponding verification interest rates, and the verification time differences of the corresponding adjacent verification time points includes: Input the corresponding verification time difference, verification rate, verification target data corresponding to the earliest verification time point, and verification association data into the target model and association model in the asset data association impact model for cross-iteration calculation to obtain test output data of multiple verification time points of the target asset; Determining error rates corresponding to a plurality of the verification time points based on the corresponding verification target data and test output data; Determine whether there is no verification time point whose corresponding error rate is greater than or equal to the preset error rate threshold; if not, issue a model applicability alarm.

7. The method according to claim 1, characterized in that The method of obtaining the pre-change future data of the target asset at multiple future time points based on the asset data association impact model corresponding to the target asset and the associated assets corresponding to the target asset, the current target data of the target asset, the current association data of the associated assets, and the current interest rate includes: The preset predicted time difference, the current target data, the current associated data and the current interest rate are input into the target model and the associated model in the asset data associated impact model for cross-iterative calculation to obtain the future data of the target asset before the change at multiple future time points.

8. The method according to claim 7, characterized in that The method of obtaining the target asset's future data after changes at multiple future time points based on the expected interest rate change range, the asset data association impact model, the current target data, the current association data, and the current interest rate includes: Based on the interest rate change range and the current interest rate, obtain the expected interest rate after the change; The predicted time difference, the current target data, the current associated data and the expected interest rate after the change are input into the target model and the associated model in the asset data association impact model for cross-iteration calculation to obtain the future data of the target asset after the change at multiple future time points.

9. The method according to claim 1, characterized in that The determining whether there is a transaction anomaly in the target asset based on the pre-change future data and the post-change future data corresponding to the plurality of future time points includes: Obtaining a target data change range corresponding to the future time point based on the corresponding future data before and after the change; Determine whether there is a future time point where the corresponding target data change amplitude is greater than or equal to a preset data change amplitude threshold. If so, issue a transaction abnormality alarm.

10. The method according to claim 1, characterized in that The determining whether there is a transaction anomaly in the target asset based on the pre-change future data and the post-change future data corresponding to the plurality of future time points includes: Obtaining a target data change range corresponding to the future time point based on the corresponding future data before and after the change; Obtaining a target data change rate corresponding to the future time point based on the target data change amplitude and the future data before the change or the future data after the change; Determine whether there is a future time point where the corresponding target data change rate is greater than or equal to a preset data change rate threshold. If so, issue a transaction abnormality alarm.

11. A transaction abnormality alarm device, characterized in that: include: A first prediction module is configured to obtain pre-change future data of the target asset at multiple future time points based on an asset data association impact model corresponding to a target asset and associated assets corresponding to the target asset, current target data of the target asset, current associated data of the associated assets, and a current interest rate; A second prediction module is configured to obtain the target asset's future data after changes at a plurality of future time points based on the expected interest rate change range, the asset data association impact model, the current target data, the current association data, and the current interest rate; an abnormality alarm module, configured to determine whether there is a transaction abnormality in the target asset based on the future data before and after the change corresponding to the plurality of future time points, and if so, issue a transaction abnormality alarm; Model building modules for: Before obtaining the pre-change future data of the target asset at multiple future time points based on the asset data association impact model corresponding to the target asset and the associated assets corresponding to the target asset, the current target data of the target asset, the current association data of the associated assets, and the current interest rate, an initial target model of the target asset and initial association models of the associated assets corresponding to the target asset are constructed based on a preset data dynamic function; Obtaining a target model in the asset data association impact model based on the target historical data information of the target asset, the associated historical data information of the associated assets, the corresponding historical interest rates, and the initial target model; Based on the target historical data information, the associated historical data information, the historical interest rate and the initial association model, an association model in the asset data association impact model is obtained.

12. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method according to any one of claims 1 to 10 is implemented.

13. A computer-readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 10 is implemented.

Images