Error transaction processing method and device in bank distributed system

Abstract

The present invention discloses a method and device for processing error transactions in a bank distributed system, and relates to the technical field of computer data processing. The method comprises: constructing a transaction feature vector according to historical error transaction data for clustering, performing a first division on the historical error transaction data to obtain a plurality of first-category data sets; determining the error transaction processing method with the highest application frequency as the error transaction processing method of each first-category data set; constructing a transaction feature vector according to the transaction feature data of the error transaction to be processed; calculating the distance between the transaction feature vector of the error transaction to be processed and the cluster center of the transaction feature vector corresponding to each first-category data set; determining a plurality of first-category data sets corresponding to a plurality of distance calculation results less than a preset threshold as standard data sets; determining the weight of each standard data set according to the distance calculation result, and determining the processing method of the error transaction to be processed in combination with the error transaction processing method of each standard data set.

Description

Technical Field

[0001] The present invention relates to the technical field of computer data processing, and in particular to a method and device for processing error transactions in a bank distributed system. Background Art

[0002] This section is intended to provide a background or context to the embodiments of the invention recited in the claims. No admission is made that the description herein is prior art by inclusion in this section.

[0003] In a bank's distributed system, transaction data often exists in multiple different databases, and distributed transactions are required to ensure transaction consistency among all parties. When an error occurs, it is often necessary to determine how to handle the error transaction in the current scenario. Currently, all error determinations are handled manually by the middle office, which is inefficient and may cause an avalanche of multiple errors due to an error that is not handled in time, which can easily affect the user experience.

[0004] To address the above problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiment of the present invention provides a method for processing error transactions in a bank distributed system, which is used to quickly determine the processing method of error transactions, improve the bank's error transaction processing efficiency, reduce the impact scope of error transactions, and improve user experience. The method includes:

[0006] Acquire historical error transaction data of the bank's distributed system, wherein the historical error transaction data includes transaction feature data of each historical error transaction and a corresponding error transaction processing method;

[0007] Constructing transaction feature vectors of historical error transactions, wherein each transaction feature vector of a historical error transaction corresponds to transaction feature data of each historical error transaction; clustering the transaction feature vectors of multiple historical error transactions, and performing a first division on the historical error transaction data according to the clustering result to obtain multiple first-category data sets, wherein the first-category data sets are data sets obtained by taking transaction features as classification criteria;

[0008] In each first-category data set, the error transaction processing method with the highest application frequency is determined as the error transaction processing method of the first-category data set;

[0009] When there is an error transaction to be processed in the bank distributed system, the transaction feature data of the error transaction to be processed is obtained; a transaction feature vector of the error transaction to be processed is constructed, and the transaction feature vector of the error transaction to be processed corresponds to the transaction feature data of the error transaction to be processed; and the distance between the transaction feature vector of the error transaction to be processed and the cluster center of the transaction feature vector corresponding to each first type of data set is calculated;

[0010] When there are multiple distance calculation results less than a preset threshold, determine multiple first-category data sets corresponding to the multiple distance calculation results as standard data sets; determine the weight of each standard data set in ascending order according to the multiple distance calculation results less than the preset threshold;

[0011] According to the weight of each standard data set and the error transaction processing method corresponding to each standard data set, the error transaction processing method of the error transaction to be processed is determined.

[0012] The embodiment of the present invention also provides an error transaction processing device in a bank distributed system, which is used to quickly determine the processing method of the error transaction, improve the bank's error transaction processing efficiency, reduce the impact range of the error transaction, and improve the user experience. The device includes:

[0013] A transaction data acquisition module, used to acquire historical error transaction data of the bank's distributed system, wherein the historical error transaction data includes transaction feature data of each historical error transaction and a corresponding error transaction processing method;

[0014] A first division module is used to construct a transaction feature vector of a historical error transaction, wherein each transaction feature vector of a historical error transaction corresponds to transaction feature data of each historical error transaction; clustering the transaction feature vectors of a plurality of historical error transactions, and performing a first division on the historical error transaction data according to the clustering result to obtain a plurality of first-category data sets, wherein the first-category data set is a data set obtained by taking the transaction feature as a classification standard;

[0015] A first determination module is used to determine, in each first-category data set, the error transaction processing method with the highest application frequency as the error transaction processing method of the first-category data set;

[0016] A calculation module is used to obtain transaction feature data of the error transaction to be processed when there is an error transaction to be processed in the bank distributed system; construct a transaction feature vector of the error transaction to be processed, wherein the transaction feature vector of the error transaction to be processed corresponds to the transaction feature data of the error transaction to be processed; and calculate the distance between the transaction feature vector of the error transaction to be processed and the cluster center of the transaction feature vector corresponding to each first type of data set;

[0017] A weight determination module, used for determining, when multiple distance calculation results are less than a preset threshold, multiple first-category data sets corresponding to the multiple distance calculation results as standard data sets; and determining the weight of each standard data set in ascending order according to multiple distance calculation results less than the preset threshold;

[0018] The second determination module is used to determine the error transaction processing method of the error transaction to be processed according to the weight of each standard data set and the error transaction processing method corresponding to each standard data set.

[0019] An embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the error transaction processing method in the above-mentioned bank distributed system when executing the computer program.

[0020] An embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the error transaction processing method in the above-mentioned bank distributed system is implemented.

[0021] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the error transaction processing method in the above-mentioned bank distributed system is implemented.

[0022] In an embodiment of the present invention, historical error transaction data of a bank distributed system is obtained; transaction feature vectors of historical error transactions are constructed, wherein each transaction feature vector of a historical error transaction corresponds to transaction feature data of each historical error transaction; transaction feature vectors of a plurality of historical error transactions are clustered, and the historical error transaction data are first divided according to the clustering result to obtain a plurality of first-category data sets; in each first-category data set, the error transaction processing method with the highest application frequency is determined as the error transaction processing method of the first-category data set; when there is an error transaction to be processed in the bank distributed system, the transaction feature data of the error transaction to be processed is obtained; a transaction feature vector of the error transaction to be processed is constructed, wherein the transaction feature vector of the error transaction to be processed corresponds to The invention relates to a method for processing error transactions, comprising: calculating the transaction feature data of the error transaction to be processed; calculating the distance between the transaction feature vector of the error transaction to be processed and the cluster center of the transaction feature vector corresponding to each first-category data set; determining, when multiple distance calculation results are less than a preset threshold, multiple first-category data sets corresponding to the multiple distance calculation results as standard data sets; determining the weight of each standard data set in order from small to large according to multiple distance calculation results less than the preset threshold; determining the error transaction processing method of the error transaction to be processed according to the weight of each standard data set and the error transaction processing method corresponding to each standard data set, so as to quickly determine the processing method of the error transaction, improve the bank's error transaction processing efficiency, reduce the impact range of the error transaction, and improve the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] 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 prior art descriptions. 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 creative work. In the drawings:

[0024] Figure 1 It is a processing flow chart of the error transaction processing method in the bank distributed system in an embodiment of the present invention;

[0025] Figure 2 It is a structural schematic diagram of an error transaction processing device in a bank distributed system according to an embodiment of the present invention;

[0026] Figure 3 It is a schematic diagram of a specific example of the structure of an error transaction processing device in a bank distributed system in an embodiment of the present invention;

[0027] Figure 4 It is a schematic diagram of a specific example of the structure of an error transaction processing device in a bank distributed system in an embodiment of the present invention;

[0028] Figure 5 The figure is a schematic diagram of the structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0029] To make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0030] Figure 1 FIG. 1 is a processing flow chart of a method for processing error transactions in a bank distributed system according to an embodiment of the present invention. Figure 1 As shown, the error transaction processing method in the bank distributed system in the embodiment of the present invention may include:

[0031] Step 101: Acquire historical error transaction data of a bank distributed system, wherein the historical error transaction data includes transaction feature data of each historical error transaction and a corresponding error transaction processing method;

[0032] Step 102: construct transaction feature vectors of historical error transactions, wherein each transaction feature vector of a historical error transaction corresponds to transaction feature data of each historical error transaction; cluster the transaction feature vectors of multiple historical error transactions, and perform a first division on the historical error transaction data according to the clustering result to obtain multiple first-category data sets, wherein the first-category data sets are data sets obtained by taking transaction features as classification criteria;

[0033] Step 103: In each first-category data set, the error transaction processing method with the highest application frequency is determined as the error transaction processing method of the first-category data set;

[0034] Step 104: when there is an error transaction to be processed in the bank distributed system, obtain transaction feature data of the error transaction to be processed; construct a transaction feature vector of the error transaction to be processed, wherein the transaction feature vector of the error transaction to be processed corresponds to the transaction feature data of the error transaction to be processed; calculate the distance between the transaction feature vector of the error transaction to be processed and the cluster center of the transaction feature vector corresponding to each first type of data set;

[0035] Step 105: when there are multiple distance calculation results less than a preset threshold, determine multiple first-category data sets corresponding to the multiple distance calculation results as standard data sets; and determine the weight of each standard data set in ascending order according to the multiple distance calculation results less than the preset threshold;

[0036] Step 106: Determine the error transaction processing method for the error transaction to be processed according to the weight of each standard data set and the error transaction processing method corresponding to each standard data set.

[0037] In order to determine the processing methods of different error transactions in the bank's distributed system, the historical error transaction data of the bank's distributed system can be obtained first, where the historical error transaction data includes the transaction feature data of each historical error transaction and the corresponding error transaction processing method. The error transaction processing method may include, for example, forward execution, backward compensation, etc.

[0038] In one embodiment, the transaction characteristic data may include characteristic data of at least one of the following dimensions: transaction channel, transaction category, transaction location, transaction account category, and transaction error code.

[0039] Since there is a correlation between the transaction feature data of an error transaction and its corresponding error transaction processing method, the historical error transaction data can be classified according to the transaction features of the historical error transactions. First, a transaction feature vector of the historical error transaction can be constructed, wherein each transaction feature vector of the historical error transaction corresponds to the transaction feature data of each historical error transaction; then, the transaction feature vectors of multiple historical error transactions can be clustered, and the historical error transaction data can be divided for the first time according to the clustering results to obtain multiple first-category data sets, wherein the first-category data sets are data sets obtained by taking transaction features as classification criteria; finally, in each first-category data set, the error transaction processing method with the highest application frequency can be determined as the error transaction processing method of the first-category data set.

[0040] In specific implementation, K-means can be selected to cluster the transaction feature vectors of multiple historical error transactions, thereby obtaining multiple first-category data sets. Since the transaction category, as the most important factor in transaction features, can directly affect the division of the first-category data set, in order to obtain a more accurate clustering result, for each of the multiple first-category data sets obtained above, the transaction category of each transaction data in the first-category data set can be determined; from all transaction categories, the transaction category with the largest number of corresponding transactions is selected, and the transaction category is determined as the transaction category of the first-category data set. For each first-category data set, determine whether the set meets condition p: whether the ratio of the number of transactions corresponding to the transaction category of the first-category data set to the number of transactions of the first-category data set is greater than the specified ratio. If not, continue to perform cluster analysis on the transaction feature vector corresponding to the first-category data set until each newly generated first-category data set meets the above condition p.

[0041] After obtaining multiple first-category data sets, the error transaction processing method with the highest application frequency in each first-category data set can be determined as the error transaction processing method of the first-category data set. For example, if the processing method of 70% of the historical error transactions in the first-category data set H is forward execution, and the processing method of 30% of the historical error transactions is backward compensation, then the error transaction processing method of the first-category data set H can be determined to be forward execution.

[0042] When there is an error transaction to be processed in a bank distributed system, the transaction feature data of the error transaction to be processed can be obtained first; a transaction feature vector of the error transaction to be processed can be constructed, wherein the transaction feature vector of the error transaction to be processed corresponds to the transaction feature data of the error transaction to be processed; the distance between the transaction feature vector of the error transaction to be processed and the cluster center of the transaction feature vector corresponding to each first-category data set is calculated; when there are multiple distance calculation results less than a preset threshold, the multiple first-category data sets corresponding to the multiple distance calculation results can be determined as standard data sets; and according to the multiple distance calculation results less than the preset threshold, the weight of each standard data set is determined in order from small to large; according to the weight of each standard data set and the error transaction processing method corresponding to each standard data set, the error transaction processing method of the error transaction to be processed can be determined.

[0043] In specific implementation, the Euclidean distance between the transaction feature vector of the error transaction to be processed and the cluster center of the transaction feature vector corresponding to each first type of data set can be calculated. The Euclidean distance calculation formula is as follows:

[0044]

[0045] Where d(x,y) is the Euclidean distance between the transaction feature vector of the error transaction to be processed and the cluster center of the transaction feature vector corresponding to each first-category data set, x k is the transaction feature vector of the kth error transaction to be processed, y k is the cluster center of the transaction feature vector corresponding to the kth first category data set, and n is an integer.

[0046] In specific implementation, after determining that multiple first-category data sets are standard data sets, the weight of each standard data set can be determined according to the following formula. Assuming that there are m standard data sets, the weight of the i-th standard data set can be set to:

[0047]

[0048] Among them, W i is the weight of the i-th standard data set, d iis the distance corresponding to the i-th standard data set (i.e., the Euclidean distance between the transaction feature vector of the error transaction to be processed and the cluster center of the transaction feature vector corresponding to the i-th standard data set).

[0049] The following is an illustration using a specific example: for example, when three Euclidean distance calculation results are less than a preset threshold, the first type of data sets corresponding to the three Euclidean distance calculation results can be set to: standard data set A, standard data set B, and standard data set C; wherein the processing method for historical error transactions of standard data set A is forward execution, the processing method for historical error transactions of standard data set B is forward execution, and the processing method for historical error transactions of standard data set C is backward compensation. The weights of standard data set A, standard data set B, and standard data set C are calculated respectively using the above-mentioned weight calculation formula. If the sum of the weight of standard data set A and the weight of standard data set B is greater than the weight of standard data set C, it can be determined that the error transaction processing method for the error transaction to be processed is forward execution.

[0050] Compared with the prior art of first determining that the pending error transaction belongs to a certain first-category data set, and then using the error transaction processing method corresponding to the first-category data set as the processing method for the pending error transaction, the embodiment of the present invention conducts a comprehensive analysis of multiple first-category data sets that meet the restriction conditions by referencing weights, and then determines the processing method for the pending error transaction based on the analysis results. This can improve the accuracy of determining the processing method for the pending error transaction and avoid affecting the reliability of the selection of the processing method for the pending error transaction due to the limitations of a certain first-category data set itself.

[0051] In one embodiment, it may also include: when only one distance calculation result is less than a preset threshold, determining the error transaction processing method for the error transaction to be processed according to the error transaction processing method of the first type of data set corresponding to the distance calculation result; when no distance calculation result is less than the preset threshold, uploading the error transaction to be processed to the management end of the bank distributed system, and the management end determines the error transaction processing method for the error transaction to be processed.

[0052] In specific implementation, if no distance calculation result is less than the preset threshold, it means that the transaction characteristics of the pending error transaction and the historical error transaction are not similar, and the error transaction processing method of the pending error transaction cannot be obtained based on the error transaction processing method of any first-category data set. At this time, the pending error transaction can be uploaded to the management end of the bank's distributed system, and the staff will determine the error transaction processing method of the pending error transaction. After the error transaction processing is completed, the processed error transaction and its processing method can also be saved as new sample data to the historical error transaction data, which is used to continue to construct the transaction feature vector of the historical error transaction for clustering and division to obtain a new first-category data set.

[0053] In the actual operation process, as the complexity of the bank's distributed system becomes higher and higher, there will always be pending error transactions that have never been processed. When the number of samples of historical error transaction data is small, it is easy to cause there to be no first-class data set whose distance calculation result with the pending error transaction is less than the preset threshold. At this time, if the pending error transaction is still processed according to the historical error transaction data, although the efficiency can be improved, it is easy to cause errors in the processing of error transactions, resulting in greater losses. Therefore, the automatic processing of error transactions based on historical error transaction data can be combined with the manual processing of error transactions. In the actual operation process, the preset threshold of the distance calculation result (that is, the preset threshold of the distance between the transaction feature vector of the pending error transaction and the cluster center of the transaction feature vector corresponding to each first-class data set) can be dynamically adjusted according to the number of samples of historical error transaction data, which can improve the processing efficiency of error transactions while ensuring that bank losses are avoided.

[0054] In one embodiment, the method may further include: taking the error transaction processing method as a category identifier, performing a second division on the historical error transaction data to obtain multiple second-category data sets, wherein the error transaction processing method in each second-category data set is the same; after determining the error transaction processing method for the error transaction to be processed, calculating the similarity between the transaction feature data of the error transaction to be processed and the transaction feature data in each second-category data set; when the error transaction processing method in the second-category data set with the highest similarity to the transaction feature data of the error transaction to be processed is not the determined error transaction processing method for the error transaction to be processed, uploading the error transaction to be processed to the management end of the bank distributed system, and the management end re-determines the error transaction processing method for the error transaction to be processed.

[0055] In one embodiment, calculating the similarity between the transaction feature data of the pending error transaction and the transaction feature data in each second category data set may include: pre-setting the priorities of the transaction feature data of different dimensions, and calculating, in turn, the proportions of the transaction feature data of each dimension of the pending error transaction in the transaction feature data of the corresponding dimension of each second category data set according to the priorities of the transaction feature data of different dimensions; determining the similarity between the transaction feature data of the pending error transaction and the transaction feature data in each second category data set according to the proportions of the transaction feature data of each dimension of the pending error transaction in the transaction feature data of the corresponding dimension of each second category data set.

[0056] In specific implementation, since the historical error transaction data includes the transaction feature data of each historical error transaction and the corresponding error transaction processing method, in addition to the first division of the historical error transaction data using the transaction feature as the classification standard, the error transaction processing method can also be used as the category identifier to perform a second division on the historical error transaction data to obtain multiple second-category data sets, wherein the error transaction processing method in each second-category data set is the same, which can be used to reversely verify the processing method of the error transaction to be processed determined according to the first-category data set, that is, by calculating the similarity between the transaction feature data of the error transaction to be processed and the transaction feature data in each second-category data set, if the processing method of the error transaction in the second-category data set with the highest similarity to the transaction feature data of the error transaction to be processed is not the determined error transaction processing method of the error transaction to be processed, it means that the processing method of the error transaction to be processed determined according to the first-category data set is inaccurate. If it is still processed according to the determined error transaction processing method, it is easy to cause multiple subsequent wrong accounts, causing serious losses.

[0057] Since not all error transactions can be handled based on historical error transaction data, the second type of data set can also be used to reversely verify the error transaction handling method for the pending error transactions previously determined based on the first type of data set. This can detect previous erroneous handling methods and take timely remedial measures to avoid further losses. During specific implementation, the pending error transactions can be uploaded to the management end of the bank's distributed system, and the staff can re-determine the error transaction handling method for the pending error transactions.

[0058] Still taking the above content as an example, the second data set can be divided into two, the second data set M corresponding to forward execution, and the second data set N corresponding to backward compensation; for example, the priorities of transaction feature data of different dimensions are pre-set as follows: transaction error code>transaction category>transaction account category>transaction channel>transaction channel; among which the error transactions to be processed are: transaction error code: 10001 (user authentication failed), transaction category: transfer, account category: ordinary user, transaction channel: ATM, transaction location: P place; if the data with transaction error code 10001 accounts for 60% of the transaction error code data in the second data set M, and also accounts for 60% of the transaction error code data in the second data set N, then continue to compare according to priority. Transaction category: if the data with transaction category of transfer accounts for 80% of the transaction category data in the second data set M and 50% of the transaction category data in the second data set N, there is no need to continue to compare the transaction account category, transaction channel, and the proportion of the transaction channel in the transaction feature data of the corresponding dimensions in the second data sets M and N. It can be determined that the similarity between the pending error transaction and the second data set M is the highest, and the error transaction processing method of the second data set M is forward execution, which is the same as the error transaction processing method of the determined pending error transaction. It can be proved that the processing method of the pending error transaction determined according to the first data set is accurate, and it is finally determined to process the pending error transaction in a forward execution manner.

[0059] In the specific implementation, in addition to using the transaction characteristics and the error transaction processing method as variables in the aforementioned method to obtain the first type of data set and the second type of data set, in order to improve the efficiency of the clustering operation, the transaction feature vectors of multiple historical error transactions can also be clustered according to the following method: (through one clustering, it can be ensured that the transaction feature similarity of each data set meets the preset requirements, and the error transaction processing method of each data set is the same)

[0060] 1. Select K cluster centers in the transaction feature vector space of historical error transactions, each cluster center corresponds to a third type of data set, and the initial elements of the third type of data set only contain the historical error transaction data corresponding to the cluster center;

[0061] 2. For each historical error transaction data, multiple cluster centers that are consistent with the error transaction processing method of the historical error transaction data are selected from all cluster centers; the distance between each selected cluster center and the transaction feature vector of the historical error transaction data is calculated; the minimum value among the multiple distances is selected as the minimum distance of the transaction features of the historical error transaction data with the same processing method; the distance between each unselected cluster center and the transaction feature vector of the historical error transaction data is calculated; the minimum value among the multiple distances is selected as the minimum distance of the transaction features of the historical error transaction data with different processing methods;

[0062] 3. For each historical error transaction data, if the minimum distance of transaction features with the same processing method is less than the minimum distance of transaction features with different processing methods, the historical error transaction data is divided into the third category data set corresponding to the cluster center corresponding to the minimum distance of transaction features with the same processing method; if the minimum distance of transaction features with the same processing method is greater than the minimum distance of transaction features with different processing methods, a new cluster center is created, and the new cluster center corresponds to a new third category data set, and the initial element of the new third category data set only contains the historical error transaction data;

[0063] 4. After executing the above steps 2 and 3 for all historical error transaction data, for each third-category data set, the cluster center corresponding to the third-category data set is updated to the mean value of all transaction feature vectors corresponding to the third-category data set;

[0064] 5. Repeat the above steps 2, 3 and 4 until the change of the cluster center is very small (the absolute value of the difference between the two values ​​of the cluster center before and after the update is less than the threshold), thereby obtaining multiple third-category data sets.

[0065] The embodiment of the present invention also provides an error transaction processing device in a bank distributed system, as described in the following embodiment. Since the principle of solving the problem by the device is similar to the error transaction processing method in the bank distributed system, the implementation of the device can refer to the implementation of the error transaction processing method in the bank distributed system, and the repeated parts will not be repeated.

[0066] Figure 2 FIG. 1 is a schematic diagram of the structure of the error transaction processing device in the bank distributed system in an embodiment of the present invention. Figure 2 As shown, the error transaction processing device in the bank distributed system in the embodiment of the present invention may specifically include:

[0067] The transaction data acquisition module 201 is used to acquire the historical error transaction data of the bank distributed system, wherein the historical error transaction data includes the transaction feature data of each historical error transaction and the corresponding error transaction processing method;

[0068] The first division module 202 is used to construct a transaction feature vector of a historical error transaction, wherein each transaction feature vector of a historical error transaction corresponds to transaction feature data of each historical error transaction; cluster the transaction feature vectors of a plurality of historical error transactions, and perform a first division on the historical error transaction data according to the clustering result to obtain a plurality of first-category data sets, wherein the first-category data sets are data sets obtained by taking transaction features as classification criteria;

[0069] A first determination module 203 is used to determine, in each first-category data set, the error transaction processing method with the highest application frequency as the error transaction processing method of the first-category data set;

[0070] The calculation module 204 is used to obtain transaction feature data of the error transaction to be processed when there is an error transaction to be processed in the bank distributed system; construct a transaction feature vector of the error transaction to be processed, wherein the transaction feature vector of the error transaction to be processed corresponds to the transaction feature data of the error transaction to be processed; and calculate the distance between the transaction feature vector of the error transaction to be processed and the cluster center of the transaction feature vector corresponding to each first type of data set;

[0071] The weight determination module 205 is used to determine, when multiple distance calculation results are less than a preset threshold, multiple first-category data sets corresponding to the multiple distance calculation results as standard data sets; and determine the weight of each standard data set in ascending order according to the multiple distance calculation results less than the preset threshold;

[0072] The second determination module 206 is used to determine the error transaction processing method of the error transaction to be processed according to the weight of each standard data set and the error transaction processing method corresponding to each standard data set.

[0073] In one embodiment, the transaction feature data includes feature data of at least one of the following dimensions:

[0074] Transaction channel, transaction category, transaction location, transaction account type, transaction error code.

[0075] Figure 3 FIG. 1 is a schematic diagram of a specific example of the structure of an error transaction processing device in a bank distributed system according to an embodiment of the present invention. Figure 3 As shown, in the embodiment, Figure 2 The error transaction processing device in the bank distributed system shown also includes:

[0076] The third determining module 301 is used to:

[0077] When only one distance calculation result is less than the preset threshold, determining the error transaction processing method of the error transaction to be processed according to the error transaction processing method of the first type of data set corresponding to the distance calculation result;

[0078] When no distance calculation result is less than the preset threshold, the pending error transaction is uploaded to the management end of the bank's distributed system, and the management end determines the error transaction processing method for the pending error transaction.

[0079] Figure 4 FIG. 1 is a schematic diagram of a specific example of the structure of an error transaction processing device in a bank distributed system according to an embodiment of the present invention. Figure 4 As shown, in the embodiment, Figure 2 The error transaction processing device in the bank distributed system shown also includes:

[0080] The second division module 401 is used for:

[0081] Taking the error transaction processing method as the category identifier, the historical error transaction data is divided for the second time to obtain multiple second-category data sets, wherein the error transaction processing method in each second-category data set is the same;

[0082] After determining the error transaction processing method of the error transaction to be processed, calculating the similarity between the transaction feature data of the error transaction to be processed and the transaction feature data in each second type of data set;

[0083] When the processing method of the error transaction in the second type of data set with the highest similarity to the transaction feature data of the error transaction to be processed is not the determined error transaction processing method of the error transaction to be processed, the error transaction to be processed will be uploaded to the management end of the bank's distributed system, and the management end will re-determine the error transaction processing method of the error transaction to be processed.

[0084] Figure 4 Can include Figure 3 The third determination module 301 in.

[0085] In one embodiment, the second division module 401 is specifically used for:

[0086] The priorities of transaction feature data of different dimensions are pre-set, and according to the priorities of transaction feature data of different dimensions, the proportion of transaction feature data of each dimension of the error transaction to be processed in the transaction feature data of the corresponding dimension of each second type of data set is calculated in sequence;

[0087] According to the proportion of the transaction feature data of each dimension of the error transaction to be processed in the transaction feature data of the corresponding dimension of each second category data set, the similarity between the transaction feature data of the error transaction to be processed and the transaction feature data in each second category data set is determined.

[0088] Based on the above invention concept, Figure 5 As shown, the present invention also proposes a computer device 500, including a memory 510, a processor 520, and a computer program 530 stored in the memory 510 and executable on the processor 520, wherein the processor 520 implements the error transaction processing method in the aforementioned bank distributed system when executing the computer program 530.

[0089] An embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the error transaction processing method in the above-mentioned bank distributed system is implemented.

[0090] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the error transaction processing method in the above-mentioned bank distributed system is implemented.

[0091] In summary, in an embodiment of the present invention, historical error transaction data of a bank distributed system is obtained; transaction feature vectors of historical error transactions are constructed, wherein each transaction feature vector of a historical error transaction corresponds to transaction feature data of each historical error transaction; transaction feature vectors of multiple historical error transactions are clustered, and the historical error transaction data are first divided according to the clustering result to obtain multiple first-category data sets; in each first-category data set, the error transaction processing method with the highest application frequency is determined as the error transaction processing method of the first-category data set; when there is an error transaction to be processed in the bank distributed system, the transaction feature data of the error transaction to be processed is obtained; a transaction feature vector of the error transaction to be processed is constructed, wherein the transaction feature vector of the error transaction to be processed corresponds to the transaction feature data of the error transaction to be processed. Transaction feature data corresponding to the error transaction to be processed; calculating the distance between the transaction feature vector of the error transaction to be processed and the cluster center of the transaction feature vector corresponding to each first-category data set; when multiple distance calculation results are less than a preset threshold, determining multiple first-category data sets corresponding to the multiple distance calculation results as standard data sets; determining the weight of each standard data set in order from small to large according to multiple distance calculation results less than the preset threshold; determining the error transaction processing method of the error transaction to be processed according to the weight of each standard data set and the error transaction processing method corresponding to each standard data set, which can quickly determine the processing method of the error transaction, improve the bank's error transaction processing efficiency, reduce the impact range of the error transaction, and improve user experience.

[0092] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0093] 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 flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, 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 flowchart and / or block diagram. 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.

[0094] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate 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 A function specified in one or more boxes.

[0095] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0096] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for handling error transactions in a bank distributed system, characterized in that: include: Acquire historical error transaction data from the bank's distributed system, wherein the historical error transaction data includes transaction feature data of each historical error transaction and a corresponding error transaction processing method; Constructing transaction feature vectors for historical error transactions, where each transaction feature vector corresponds to transaction feature data for each historical error transaction; clustering the transaction feature vectors of multiple historical error transactions, and performing a first division of the historical error transaction data based on the clustering results to obtain multiple first-category data sets, where the first-category data sets are data sets obtained using transaction features as classification criteria; In each first-category data set, the error transaction processing method with the highest frequency of application is determined as the error transaction processing method for the first-category data set; When there is an error transaction to be processed in the bank distributed system, transaction feature data of the error transaction to be processed is obtained; a transaction feature vector of the error transaction to be processed is constructed, wherein the transaction feature vector of the error transaction to be processed corresponds to the transaction feature data of the error transaction to be processed; and a distance between the transaction feature vector of the error transaction to be processed and a cluster center of the transaction feature vector corresponding to each first type of data set is calculated; When multiple distance calculation results are smaller than a preset threshold, determining multiple first-category data sets corresponding to the multiple distance calculation results as standard data sets; Determine the weight of each standard data set in ascending order based on multiple distance calculation results that are less than a preset threshold; Determine the error transaction processing method for the error transaction to be processed based on the weight of each standard data set and the error transaction processing method corresponding to each standard data set; Using the error transaction processing method as a category identifier, the historical error transaction data is divided a second time to obtain multiple second-category data sets, where the error transactions in each second-category data set are processed in the same way; After determining the error transaction processing method for the error transaction to be processed, calculating the similarity between the transaction feature data of the error transaction to be processed and the transaction feature data in each second type of data set; If the handling method of the error transaction in the second type of data set with the highest similarity to the transaction feature data of the error transaction to be processed is not the determined error transaction handling method for the error transaction to be processed, the error transaction to be processed is uploaded to the management terminal of the bank's distributed system, and the management terminal re-determines the error transaction handling method for the error transaction to be processed; Calculate the similarity between the transaction feature data of the pending error transaction and the transaction feature data in each second type of data set, including: Priorities for transaction feature data of different dimensions are pre-set. Based on the priorities of the transaction feature data of different dimensions, the proportion of the transaction feature data of each dimension of the error transaction to be processed in the transaction feature data of the corresponding dimension in each second-category data set is calculated in sequence; Based on the proportion of the transaction feature data of each dimension of the error transaction to be processed in the transaction feature data of the corresponding dimension of each second-category data set, the similarity between the transaction feature data of the error transaction to be processed and the transaction feature data in each second-category data set is determined.

2. The method according to claim 1, wherein The transaction feature data includes feature data of at least one of the following dimensions: Transaction channel, transaction type, transaction location, transaction account type, and transaction error code.

3. The method according to claim 1, wherein Also includes: When only one distance calculation result is less than the preset threshold, determining the error transaction processing method for the error transaction to be processed based on the error transaction processing method of the first type of data set corresponding to the distance calculation result; When no distance calculation result is less than the preset threshold, the pending error transaction is uploaded to the management end of the bank's distributed system, and the management end determines the error transaction processing method for the pending error transaction.

4. A device for processing error transactions in a bank distributed system, characterized in that: include: A transaction data acquisition module is used to acquire historical error transaction data from the bank's distributed system, wherein the historical error transaction data includes transaction feature data of each historical error transaction and the corresponding error transaction processing method; a first partitioning module for constructing transaction feature vectors for historical error transactions, wherein each transaction feature vector for a historical error transaction corresponds to transaction feature data for each historical error transaction; clustering the transaction feature vectors of multiple historical error transactions, and performing a first partitioning of the historical error transaction data based on the clustering results to obtain multiple first-category data sets, wherein the first-category data sets are data sets obtained by using transaction features as classification criteria; a first determining module configured to determine, in each first-category data set, the error transaction processing method with the highest application frequency as the error transaction processing method for the first-category data set; a calculation module configured to, when there is an error transaction to be processed in the bank distributed system, obtain transaction feature data of the error transaction to be processed; construct a transaction feature vector for the error transaction to be processed, the transaction feature vector corresponding to the transaction feature data of the error transaction to be processed; and calculate a distance between the transaction feature vector of the error transaction to be processed and a cluster center of the transaction feature vector corresponding to each first category of data set; a weight determination module, configured to determine, when a plurality of distance calculation results are less than a preset threshold, a plurality of first-category data sets corresponding to the plurality of distance calculation results as standard data sets; Determine the weight of each standard data set in ascending order based on multiple distance calculation results that are less than a preset threshold; a second determination module, configured to determine an error transaction processing method for an error transaction to be processed based on the weight of each standard data set and the error transaction processing method corresponding to each standard data set; The second division module is used for: Using the error transaction processing method as a category identifier, the historical error transaction data is divided a second time to obtain multiple second-category data sets, where the error transactions in each second-category data set are processed in the same way; After determining the error transaction processing method for the error transaction to be processed, calculating the similarity between the transaction feature data of the error transaction to be processed and the transaction feature data in each second type of data set; If the handling method of the error transaction in the second type of data set with the highest similarity to the transaction feature data of the error transaction to be processed is not the determined error transaction handling method for the error transaction to be processed, the error transaction to be processed is uploaded to the management terminal of the bank's distributed system, and the management terminal re-determines the error transaction handling method for the error transaction to be processed; The second division module is specifically used for: Priorities for transaction feature data of different dimensions are pre-set. Based on the priorities of the transaction feature data of different dimensions, the proportion of the transaction feature data of each dimension of the error transaction to be processed in the transaction feature data of the corresponding dimension in each second-category data set is calculated in sequence; Based on the proportion of the transaction feature data of each dimension of the error transaction to be processed in the transaction feature data of the corresponding dimension of each second-category data set, the similarity between the transaction feature data of the error transaction to be processed and the transaction feature data in each second-category data set is determined.

5. The device according to claim 4, characterized in that The transaction feature data includes feature data of at least one of the following dimensions: Transaction channel, transaction type, transaction location, transaction account type, and transaction error code.

6. The device according to claim 4, characterized in that The device further includes a third determining module, configured to: When only one distance calculation result is less than the preset threshold, determining the error transaction processing method for the error transaction to be processed based on the error transaction processing method of the first type of data set corresponding to the distance calculation result; When no distance calculation result is less than the preset threshold, the pending error transaction is uploaded to the management end of the bank's distributed system, and the management end determines the error transaction processing method for the pending error transaction.

7. 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 computer program, the method according to any one of claims 1 to 3 is implemented.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 3 is implemented.

9. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 3 is implemented.

Images