Method and device for controlling and auditing content during banking business processing

Abstract

The present invention discloses a method and device for controlling audit content in banking business handling, relating to the financial field. The method includes: receiving a banking business handling request from a user; the business includes multiple business steps; in each current business step of the business handling, the following operations are performed until all business steps are completed: determining all key features that need to be audited in the current business step according to the preferred audit step sets of each key feature; the preferred audit step sets of each key feature are pre-determined based on the feature data involved in each business step of each banking business and the key feature set corresponding to each banking business, and the key feature set corresponding to each banking business is pre-determined based on the historical transaction data of the banking business; auditing the current business step according to all key features that need to be audited in the current business step, and completing the current business step when the audit is passed. The present invention can effectively control the risks in banking business handling.

Description

Technical Field

[0001] The present invention relates to the field of financial technology, and in particular, to a method and device for controlling audit content in banking business processing. Background Art

[0002] This section aims to provide background or context for the embodiments of the present invention described in the claims. The description herein is not admitted to be prior art merely by including it in this section.

[0003] Currently, banking business scenarios consist of a series of business steps. In business processing, in order to strictly control risks, audits are required. However, the current audit content and audit processes are all set manually and lack data support. Therefore, it is impossible to effectively control risks in banking business processing. Summary of the Invention

[0004] An embodiment of the present invention provides a method for controlling audit content in banking business processing to effectively control risks in banking business processing. The method includes:

[0005] Receiving a banking business processing request from a user; the banking business includes multiple business steps;

[0006] In each current business step of business processing, the following operations are performed until all business steps are completed:

[0007] Determining all key features that need to be audited in the current business step according to the preferred audit step sets of each key feature; the preferred audit step sets of each key feature are determined in advance according to the feature data involved in each business step of each banking business and the key feature set corresponding to each banking business, and the key feature set corresponding to each banking business is determined in advance according to the historical transaction data of each banking business;

[0008] Auditing the current business step according to all key features that need to be audited in the current business step, and completing the current business step when the audit is passed.

[0009] An embodiment of the present invention also provides a device for controlling audit content in banking business processing to effectively control risks in banking business processing. The device includes:

[0010] A receiving unit, configured to receive a banking business processing request from a user; the banking business includes multiple business steps;

[0011] A control unit, configured to perform the following operations in each current business step of business processing until all business steps are completed:

[0012] Determine all key features to be reviewed in the current business step according to the set of preferred review steps for each key feature; the set of preferred review steps for each key feature is determined in advance based on the feature data involved in each business step of each banking service and the set of key features corresponding to each banking service, and the set of key features corresponding to each banking service is determined in advance based on the historical transaction data of each banking service;

[0013] Conduct the review of the current business step according to all the key features to be reviewed in the current business step, and complete the current business step when the review is passed.

[0014] An embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the method for controlling the review content in the above-mentioned banking service handling is implemented.

[0015] An embodiment of the present invention also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for controlling the review content in the above-mentioned banking service handling is implemented.

[0016] An embodiment of the present invention also provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the method for controlling the review content in the above-mentioned banking service handling is implemented.

[0017] In an embodiment of the present invention, the solution for controlling the review content in banking service handling includes: receiving a business handling request from a user; the business includes multiple business steps; in each current business step of business handling, the following operations are performed until all business steps are completed: Determine all key features to be reviewed in the current business step according to the set of preferred review steps for each key feature; the set of preferred review steps for each key feature is determined in advance based on the feature data involved in each business step of each banking service and the set of key features corresponding to each banking service, and the set of key features corresponding to each banking service is determined in advance based on the historical transaction data of each banking service; Conduct the review of the current business step according to all the key features to be reviewed in the current business step, and complete the current business step when the review is passed, which can effectively control the risks in banking service handling. Description of the Drawings

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings. In the accompanying drawings:

[0019] Figure 1 It is a schematic flowchart of the method for controlling the audit content in the handling of banking business in the embodiment of the present invention;

[0020] Figure 2 It is a schematic flowchart of the method for determining the key feature set corresponding to each banking business in the embodiment of the present invention;

[0021] Figure 3 It is a schematic flowchart of the method for determining the preferred audit step set of each key feature in the embodiment of the present invention;

[0022] Figure 4 It is a schematic flowchart of the method for controlling the audit content in the handling of banking business in another embodiment of the present invention;

[0023] Figure 5 It is a schematic structural diagram of the device for controlling the audit content in the handling of banking business in the embodiment of the present invention. Detailed implementation manners

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer and more understandable, the following will further describe the embodiments of the present invention in detail with reference to the accompanying drawings. Herein, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but not to limit the present invention.

[0025] Figure 1 It is a schematic flowchart of the method for controlling the audit content in the handling of banking business in the embodiment of the present invention. As Figure 1 shown, the method includes the following steps:

[0026] Step 101: Receive a banking business handling request from a user; the banking business includes multiple business steps;

[0027] Step 102: In each current business step of the business handling, perform the following operations until all business steps are completed:

[0028] Step 1021: Determine all key features to be reviewed in the current business step according to the preferred review step sets of each key feature; the preferred review step sets of each key feature are determined in advance based on the feature data involved in each business step of each banking service and the key feature set corresponding to each banking service, and the key feature set corresponding to each banking service is determined in advance based on the historical transaction data of each banking service.

[0029] Step 1022: Conduct the review of the current business step according to all the key features to be reviewed in the current business step, and complete the current business step when the review is passed.

[0030] When the method for controlling the review content in banking service handling provided by the embodiment of the present invention works: receive a business handling request from a user; the business includes multiple business steps; in each current business step of business handling, the following operations are performed until all business steps are completed: determine all key features to be reviewed in the current business step according to the preferred review step sets of each key feature; the preferred review step sets of each key feature are determined in advance based on the feature data involved in each business step of each banking service and the key feature set corresponding to each banking service, and the key feature set corresponding to each banking service is determined in advance based on the historical transaction data of each banking service; conduct the review of the current business step according to all the key features to be reviewed in the current business step, and complete the current business step when the review is passed, which can effectively control the risks in banking service handling. The method for controlling the review content in banking service handling will be introduced in detail below.

[0031] In one embodiment, as Figure 2 shown, the method for controlling the review content in banking service handling may further include: determining the key feature set corresponding to each banking service in advance according to the historical transaction data of each banking service according to the following method:

[0032] Step 201: Select multiple machine learning models.

[0033] Step 202: Determine the features corresponding to each banking service according to the historical transaction data of each banking service.

[0034] Step 203: For each feature combination, based on the transaction data regarding the feature combination in the historical transaction data, train the selected multiple machine learning models with each risk type as the supervision identifier to obtain multiple risk prediction models corresponding to each risk type.

[0035] Step 204: For each risk type involved, determine the error vector corresponding to this risk type for this feature combination. Each component of this error vector corresponds one-to-one to the risk prediction model corresponding to this risk type, and the value of each component is equal to the misrecognition rate of the risk prediction model corresponding to this risk type corresponding to this component;

[0036] Step 205: Based on the error vector, determine the partial order of each risk type involved corresponding to the feature combination. Among them, for any two feature combinations, this partial order can be used to determine whether the first feature combination corresponding to this risk type involved for these two feature combinations is superior to the second feature combination corresponding to this risk type involved for these two feature combinations;

[0037] Step 206: According to the partial order of each risk type involved corresponding to the feature combination, determine multiple maximal feature combinations corresponding to this risk type involved. Among them, this maximal feature combination is the maximal element of this partial order;

[0038] Step 207: Determine the union of multiple maximal feature combinations corresponding to all risk types involved as the key feature set corresponding to this risk type for each banking service;

[0039] Step 208: Determine the union of the key feature sets corresponding to all risk types for each banking service as the key feature set corresponding to each banking service.

[0040] Specifically in implementation, the above implementation manner of pre-determining the key feature set corresponding to each banking service can effectively find the key features of each risk type and provide a data basis for subsequent audits. In this way, it can not only effectively control the occurrence of risks, but also avoid unnecessary audit work.

[0041] In one embodiment, based on the error vector, determining the partial order of each risk type involved corresponding to the feature combination may include:

[0042] For any two feature combinations, determine the difference between the error vector of the first feature combination corresponding to this risk type involved for these two feature combinations and the error vector of the second feature combination corresponding to this risk type involved for these two feature combinations. If each component of this difference is less than or equal to 0, then determine that the first feature combination corresponding to this risk type involved for these two feature combinations is superior to the second feature combination corresponding to this risk type involved for these two feature combinations.

[0043] It should be noted that the definition of the maximal element of the partial order is that there is no other element in the set corresponding to the partial order such that this other element is superior to this maximal element. The maximal feature combination is that there is no other feature combination in all feature combinations corresponding to the banking service that is superior to this maximal feature combination.

[0044] In one embodiment, step 206 determines multiple maximal feature combinations corresponding to the risk type involved according to the partial order of the feature combinations corresponding to each risk type involved, which may include:

[0045] Initialize the maximal authentication value corresponding to each feature combination of the banking business to possible, and initialize the corresponding comparison boolean value to yes;

[0046] Successively perform the following steps for each feature combination of the banking business:

[0047] If the maximal authentication value corresponding to this feature combination is not equal to possible, then set the comparable feature combination corresponding to this feature combination to empty; otherwise, set the comparable feature combination corresponding to this feature combination to all feature combinations of the banking business for which the corresponding comparison boolean value is yes (excluding this feature combination);

[0048] Successively determine the partial order relationship between this feature combination and each corresponding comparable feature combination: If this comparable feature combination is superior to this feature combination, then update the maximal authentication value corresponding to this feature combination to no; If this feature combination is superior to this comparable feature combination, then update the maximal authentication value corresponding to this comparable feature combination to no, and determine this comparable feature combination as the secondary feature combination of this feature combination;

[0049] If it is determined that all comparable feature combinations corresponding to this feature combination are not superior to this feature combination, then determine this feature combination as the maximal feature combination corresponding to the risk type involved, and update the comparison boolean values of all secondary feature combinations of this maximal feature combination to no.

[0050] In one embodiment, as Figure 3 shown, the above method for controlling the review content in the handling of banking business may further include: predetermining a set of preferred review steps for each key feature according to the feature data involved in each business step of each banking business and the key feature set corresponding to each banking business according to the following method:

[0051] Step 301: Determine the dependency relationship between each business step of the banking business;

[0052] Step 302: Determine the key features involved in each business step of each banking business according to the feature data involved in each business step of each banking business and the key feature set corresponding to each banking business;

[0053] Step 303: Determine the preferred review steps for each key feature according to the key features involved in each business step of each banking business and the dependency relationship between the business steps of each banking business;

[0054] Step 304: Upload the set of preferred review steps for each key feature to the blockchain.

[0055] In specific implementation, the above implementation manner of pre-determining the set of preferred review steps for each key feature can ensure timely control of the occurrence of banking business risks and avoid unnecessary review work.

[0056] In one embodiment, determining the preferred review steps for each key feature according to the key features involved in each business step of each banking business and the dependency relationship between each business step of each banking business may include:

[0057] Determine the set of potential review steps for each key feature according to the key features involved in each business step of each banking business;

[0058] For the set of potential review steps for each key feature, determine whether there is a node in the set of potential review steps that is an ancestor node of other nodes; where each potential review step corresponds to a node;

[0059] When there is a node in the set of potential review steps that is an ancestor node of other nodes, remove the ancestor node from the set of potential review steps until there is no ancestor node of other nodes in the set of potential review steps, and obtain multiple preferred review steps for each key feature.

[0060] In specific implementation, the above other nodes refer to the existence of a certain other node such that the ancestor node is the ancestor node of the other node, rather than the ancestor node of all other nodes.

[0061] In one embodiment, determining the preferred review steps for each key feature according to the key features involved in each business step of each banking business and the dependency relationship between each business step of each banking business may include:

[0062] Based on the dependency relationship between business steps, construct a step relationship graph corresponding to each banking business. Each node of the step relationship graph is a business step of each banking business. There is an edge between two business steps if and only if there is a dependency relationship between the two business steps, and the direction of the edge is from the dependent step to the dependent step, and set the distance of each edge to 1;

[0063] Determine the set of potential review steps for each key feature according to the key features involved in each business step of each banking business;

[0064] For each key feature, initialize the set of alternative business steps as the set of potential review steps for the key feature;

[0065] Loop and execute the following steps until there are no two steps in the set of alternative business steps such that the shortest distance between the two steps in the step relationship graph is finite:

[0066] Find two steps in the set of alternative business steps such that the shortest distance from the first step to the second step among the two steps in the step relationship graph is finite, and delete the first step from the set of alternative business steps;

[0067] Select business steps that meet the following conditions from all business steps of each banking business: such that the distance from each alternative business step in the set of alternative business steps to the business step is finite;

[0068] Among the selected multiple business steps, determine the business steps that do not depend on other business steps among the selected multiple business steps, and determine the determined business steps as the preferred review steps for the key feature.

[0069] In specific implementation, the above implementation manner of determining the preferred review steps for each key feature can quickly and accurately find the preferred review steps for the key feature by using the graph structure.

[0070] In one embodiment, as Figure 4 shown, the method for controlling the review content in the above-mentioned banking business handling may further include:

[0071] Step 103: Real-time monitor each banking business according to a pre-established risk warning model;

[0072] Step 104: When a new risk type that does not exist in the historical risk type set is monitored, add the new risk type to the multiple risk types involved in each banking business to obtain the updated multiple risk types involved in each banking business;

[0073] Step 105: According to the updated multiple risk types involved in each banking business, update the key feature set corresponding to each banking business and the preferred review step set for each key feature; the updated preferred review step set for each key feature is used to control the review content in banking business handling.

[0074] In specific implementation, the further preferred implementation manner of the method for controlling the review content in the above-mentioned banking business handling further improves the security of banking business handling.

[0075] In one embodiment, the method for controlling the review content in the above-mentioned banking business handling may further include: establishing the risk warning model in advance according to the following method:

[0076] Establish a risk early warning model based on the historical risk data involved in each risk type related to each banking business;

[0077] Upload the risk early warning model to the blockchain.

[0078] In specific implementation, the implementation method of pre - establishing the risk early warning model further improves the security of handling banking business.

[0079] In specific implementation, the method further includes:

[0080] For each edge of the step relationship diagram corresponding to the banking business, set the execution distance corresponding to this edge as the estimated processing time of the dependent step corresponding to this edge;

[0081] For each step, according to the step relationship diagram and the execution distances corresponding to each edge, determine the longest distance between this step and the initial node of the step relationship diagram, and determine the sum of this longest distance and the estimated processing time of this step as the completion time of this step;

[0082] Before the completion time of each step, send the review prompt information of this step to the review system.

[0083] Among them, the estimated processing time of each step can be determined according to the following method:

[0084] Obtain the historical processing data of the banking business;

[0085] For each business step of the banking business, obtain the processing time data of this business step from the historical processing data of the banking business, and determine each processing time of this business step as the processing time sample of this business step;

[0086] Set multiple probability values, and determine the estimated processing time corresponding to each probability value for each business step of the banking business. Among them, the estimated processing time corresponding to each probability value for each business step is determined such that the proportion of samples less than or equal to this estimated processing time in the processing time samples of this business step is equal to this probability value;

[0087] According to the historical processing data of the banking business and the step relationship diagram, determine the maximum business step sequence, where the first business step of this maximum business step sequence does not depend on any business step, and the last business step is not depended on by any other business step, and for every two adjacent business steps in the sequence, the next business step depends on the previous business step, and there is no waiting time (the time waiting for the completion of other business steps) in the processing time of each business step of this maximum business step sequence;

[0088] Based on the maximum business process sequence and the historical processing data of banking operations, determine the cumulative error of each probability value; specifically, the probability value can be determined according to the following formula 's cumulative error:

[0089] (or ),

[0090] where E p is the cumulative error corresponding to the probability value p , is the estimated processing time of the th business process step in the maximum business process sequence corresponding to the probability value p , t j is the business processing duration of the j th processing data in this historical processing data (only including the processing time of each step, and the processing time of parallel steps is not accumulated);

[0091] Select the probability value with the smallest corresponding cumulative error from multiple probability values;

[0092] For each business process step, determine the estimated processing time of this business process step as the estimated processing time corresponding to the selected probability value.

[0093] To facilitate understanding of how the present invention is implemented, an example is given below for introduction.

[0094] The application scenario of how the present invention is implemented can be: The banking business scenario consists of a series of transactions. For example, the foreign currency exchange business has the following steps: customer identity verification, creating a temporary account, performing foreign currency exchange, business review, and giving the customer foreign currency notes. In the application, it can be abstracted as s1, s2,... sn.

[0095] There may be different risks in different steps, and the factors triggering such risks may be certain key features, such as teller operation errors, incorrect customer face recognition data, teller illegal authorization, etc. In this way, it is necessary for the bank to manually review each key feature (key information) in the business handling process, so that risks can be discovered in a timely manner and the occurrence of risks can be reduced.

[0096] Risks: Various risks that may be encountered in bank business handling, such as staff operation risks, customer loss risks, transaction failure risks, etc.

[0097] In the method for controlling and auditing content in banking business processing provided by the embodiments of the present invention, the dependency relationships between business steps are determined; for a certain banking business, the risk types involved in the banking business are determined; based on the historical business data corresponding to each risk type, feature selection is performed to obtain the key feature set corresponding to each risk type; it is determined whether there are key features of the risk types involved in the business step in the feature data of each business step, and if so, it is determined that the business step is a potential auditing step for the key feature; according to the set of potential auditing steps of each key feature, and then based on the dependency relationships between business steps, the set of auditing steps of each key feature is determined; when the business proceeds to a certain step, all the key features that need to be audited at this step are determined, so that the bank audits all these key features at this step.

[0098] Regarding the historical business data, it is the historical data of customers handling business. Each piece of historical data includes business time, location, customer information, business type, business channel, business specific data (such as amount, currency, whether it is a large amount, etc.), customer face verification data, teller operation data, teller authorization data, risk identification (whether there is a risk), risk type, and the corresponding steps where the risk category appears.

[0099] Each dimension of the above historical data can be called a feature, such as customer information, business information, time, etc. The key feature is the feature with the highest correlation with risk. For example, the key features of teller violation risk may be teller operation data and teller authorization data.

[0100] Feature selection can be a feature selection method in machine learning. This method can find the most critical features and ignore the unimportant features, for the purpose of dimensionality reduction and improving prediction accuracy.

[0101] Each of the above data contains the source step information of the data. For example, there is an identifier in the teller operation data indicating that the data comes from the teller operation step.

[0102] The method for controlling and auditing content in banking business processing provided by the embodiments of the present invention includes the following steps:

[0103] 1. For a certain banking business, determine the dependency relationships between multiple steps, that is, the execution of one step must depend on the completion of other steps. For example, for two nodes a and b, if a depends on b, a can be called a child node and b can be called a parent node. If there is a linear dependency link, any other node in the line is an ancestor node of the end node, and the end node is a descendant node of any other node. For example, if a depends on b, b depends on c, c depends on d, and d depends on e, then a, b, c, and d are all ancestor nodes of the e node, and e is their descendant node.

[0104] 2. For a certain banking business, obtain the historical business data involved, determine the risk types involved in the banking business, as well as the risk types involved in each business step, the business steps involved in each risk type, and the characteristic data involved in each business step;

[0105] 3. Based on the historical business data corresponding to each risk type, perform feature selection to obtain multiple key features corresponding to each risk type. The union of the key features of all risk types constitutes the set of key features to be audited for this banking business;

[0106] 4. Determine whether there are key features of the risk types involved in the characteristic data of each business step. If there are, determine that this business step is a potential audit step for this key feature, and then obtain the set of potential audit steps for each key feature;

[0107] 5. According to the set of potential audit steps for each key feature, and then based on the dependency relationship between business steps, determine the set of audit steps for each key feature. And store this information in the blockchain.

[0108] 5.1 It can be to determine whether there is a node in the set of potential audit steps for each key feature that is an ancestor node of other nodes. If there is, then remove this ancestor node from the set of potential audit steps until there is no node that is an ancestor node of other nodes. This set of potential audit steps is the set of audit steps.

[0109] 5.2 The above set of audit steps can be further optimized. Determine whether there is a certain step among multiple steps of the entire business that is a common descendant node of the above set of audit steps. If there is, then replace all ancestor nodes of this descendant node in the set of audit steps with this descendant node.

[0110] 6. When the business proceeds to a step in this set of steps, based on the set of audit steps for each key feature stored in the blockchain, determine all the key features that need to be audited at this step, so that the bank can audit all these key features at this step.

[0111] It can also be audited based on the risk prediction result:

[0112] 7. For each risk type of this banking business, based on the obtained risk data, establish a risk early warning model and store it in the blockchain.

[0113] 8. The above risk early warning model monitors the risks of the banking business in real time. When a certain risk type is monitored, add this risk type to the set of risks to be controlled.

[0114] When the business reaches a certain step in the set of steps, determine whether there is a risk in the set of risks to be controlled among the risk types corresponding to this step. If so, from all the key features that need to be reviewed in this step, screen out the key features belonging to the key feature set of this risk type. At this step, review the screened key feature data.

[0115] The advantage of the method for controlling and reviewing content in banking business handling provided by the embodiments of the present invention is that it can efficiently determine the review content and review time points, control the risks in business handling, and achieve digital risk control of the bank.

[0116] An apparatus for controlling and reviewing content in banking business handling is also provided in the embodiments of the present invention, as described in the following embodiments. Since the principle of the apparatus for solving problems is similar to that of the method for controlling and reviewing content in banking business handling, the implementation of the apparatus can refer to the implementation of the method for controlling and reviewing content in banking business handling, and the repeated parts will not be elaborated.

[0117] Figure 5 It is a schematic structural diagram of the apparatus for controlling and reviewing content in banking business handling in the embodiments of the present invention, as Figure 5 shown. The apparatus includes:

[0118] A receiving unit 01 for receiving a banking business handling request from a user; the banking business includes a plurality of business steps;

[0119] A control unit 02 for performing the following operations in each current business step of business handling until all business steps are completed:

[0120] Determine all the key features that need to be reviewed in the current business step according to the preferred review step sets of each key feature; the preferred review step sets of each key feature are determined in advance according to the feature data involved in each business step of each banking business and the key feature set corresponding to each banking business, and the key feature set corresponding to each banking business is determined in advance according to the historical transaction data of each banking business;

[0121] Review the current business step according to all the key features that need to be reviewed in the current business step, and complete the current business step when the review is passed.

[0122] In one embodiment, the apparatus for controlling and reviewing content in banking business handling may further include: a key feature set determination unit for determining the key feature set corresponding to each banking business in advance according to the historical transaction data of each banking business according to the following method:

[0123] Select a plurality of machine learning models;

[0124] Determine the features corresponding to each banking service based on the historical transaction data of each banking service;

[0125] For each feature combination, based on the transaction data regarding the feature combination in the historical transaction data, train multiple selected machine learning models with each risk type as the supervision identifier to obtain multiple risk prediction models corresponding to each risk type;

[0126] For each risk type involved, determine the error vector corresponding to the feature combination for this risk type. Each component of the error vector corresponds one-to-one with the risk prediction model corresponding to this risk type, and the value of each component is equal to the misrecognition rate of the risk prediction model corresponding to this risk type corresponding to this component;

[0127] Based on the error vector, determine the partial order of the feature combinations corresponding to each risk type involved. Among them, for any two feature combinations, this partial order can be used to determine whether the first feature combination corresponding to this risk type involved in these two feature combinations is superior to the second feature combination corresponding to this risk type involved in these two feature combinations;

[0128] According to the partial order of the feature combinations corresponding to each risk type involved, determine multiple maximal feature combinations corresponding to this risk type involved, where the maximal feature combination is the maximal element of this partial order;

[0129] Determine the union of the multiple maximal feature combinations corresponding to all risk types involved as the key feature set corresponding to this risk type for each banking service;

[0130] Determine the union of the key feature sets corresponding to all risk types corresponding to each banking service as the key feature set corresponding to each banking service.

[0131] In one embodiment, based on the error vector, determining the partial order of the feature combinations corresponding to each risk type involved may include:

[0132] For any two feature combinations, determine the difference between the error vector of the first feature combination corresponding to this risk type involved in these two feature combinations and the error vector of the second feature combination corresponding to this risk type involved in these two feature combinations. If each component of this difference is less than or equal to 0, then determine that the first feature combination corresponding to this risk type involved is superior to the second feature combination corresponding to this risk type involved.

[0133] In one embodiment, the device for controlling the review content in the above banking service handling may further include: a preferred review step set determination unit, which is used to pre-determine the preferred review step sets of each key feature according to the following method based on the feature data involved in each business step of each banking service and the key feature set corresponding to each banking service:

[0134] Determine the dependencies between the various business steps of banking services;

[0135] Based on the characteristic data involved in the various business steps of each banking service and the key characteristic set corresponding to each banking service, determine the key characteristics involved in the various business steps of each banking service;

[0136] According to the key characteristics involved in the various business steps of each banking service and the dependencies between the business steps of each banking service, determine the preferred review steps for each key characteristic;

[0137] Upload the set of preferred review steps for each key characteristic to the blockchain.

[0138] In one embodiment, according to the key characteristics involved in the various business steps of each banking service and the dependencies between the business steps of each banking service, determining the preferred review steps for each key characteristic may include:

[0139] Based on the key characteristics involved in the various business steps of each banking service, determine the set of potential review steps for each key characteristic;

[0140] For the set of potential review steps for each key characteristic, determine whether there is a node in the set of potential review steps that is an ancestor node of other nodes; where each potential review step corresponds to a node;

[0141] When there is a node in the set of potential review steps that is an ancestor node of other nodes, remove the ancestor node from the set of potential review steps until there is no ancestor node of other nodes in the set of potential review steps, and obtain multiple preferred review steps for each key characteristic.

[0142] In one embodiment, according to the key characteristics involved in the various business steps of each banking service and the dependencies between the business steps of each banking service, determining the preferred review steps for each key characteristic may include:

[0143] Based on the dependencies between the business steps, construct a step relationship graph corresponding to each banking service, where each node of the step relationship graph is a business step of each banking service, and there is an edge between two business steps if and only if there is a dependency between the two business steps, and the direction of the edge is from the dependent step to the dependent step, and set the distance of each edge to 1;

[0144] Based on the key characteristics involved in the various business steps of each banking service, determine the set of potential review steps for each key characteristic;

[0145] For each key feature, initialize the set of alternative business steps as the set of potential review steps for that key feature;

[0146] Loop through the following steps until there are no two steps in the set of alternative business steps such that the shortest distance between the two steps in the step relationship diagram is finite:

[0147] Find two steps in the set of alternative business steps such that the shortest distance from the first step to the second step among the two steps in the step relationship diagram is finite, and remove the first step from the set of alternative business steps;

[0148] From all the business steps of each banking service, select the business steps that meet the following conditions: the distance from each alternative business step in the set of alternative business steps to this business step is finite;

[0149] From the selected multiple business steps, determine the business steps that do not depend on other business steps among the selected multiple business steps, and determine the determined business steps as the preferred review steps for that key feature.

[0150] In one embodiment, the device for controlling the review content in the above-mentioned banking service handling may further include:

[0151] A monitoring unit for performing real-time risk monitoring on each banking service according to a pre-established risk warning model;

[0152] A first update unit for adding a new risk type to the multiple risk types involved in each banking service when a new risk type not existing in the historical risk type set is monitored, to obtain the updated multiple risk types involved in each banking service;

[0153] A second update unit for updating the key feature set corresponding to each banking service and the preferred review step set of each key feature according to the updated multiple risk types involved in each banking service; the updated preferred review step set of each key feature is used for controlling the review content in banking service handling.

[0154] In one embodiment, the device for controlling the review content in the above-mentioned banking service handling may further include: a establishing unit for pre-establishing the risk warning model according to the following method:

[0155] Establish a risk warning model according to the historical risk data involved in each risk type involved in each banking service;

[0156] Upload the risk warning model to the blockchain.

[0157] An embodiment of the present invention further provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the method for controlling the review content in the above-mentioned banking business handling is implemented.

[0158] An embodiment of the present invention further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the method for controlling the review content in the above-mentioned banking business handling is implemented.

[0159] An embodiment of the present invention further provides a computer program product. The computer program product includes a computer program. When the computer program is executed by a processor, the method for controlling the review content in the above-mentioned banking business handling is implemented.

[0160] In an embodiment of the present invention, the solution for controlling the review content in banking business handling includes: receiving a business handling request from a user; the business includes multiple business steps; in each current business step of the business handling, the following operations are performed until all business steps are completed: determining all key features that need to be reviewed in the current business step according to the preferred review step sets of each key feature; the preferred review step sets of each key feature are determined in advance according to the feature data involved in each business step of each banking business and the key feature set corresponding to each banking business, and the key feature set corresponding to each banking business is determined in advance according to the historical transaction data of each banking business; performing the review of the current business step according to all key features that need to be reviewed in the current business step, and completing the current business step when the review is passed, which can effectively control the risks in banking business handling.

[0161] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0162] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows 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 the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices generate means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0163] These computer program instructions can 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, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0164] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0165] The specific embodiments described above further elaborate on the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for controlling audit content in banking business processing, characterized in that, it includes: Receiving a banking business processing request from a user; the banking business includes multiple business steps; In each current business step of business processing, the following operations are performed until all business steps are completed: According to the preferred audit step sets of each key feature, determine all the key features that need to be audited in the current business step; the preferred audit step sets of each key feature are determined in advance based on the feature data involved in each business step of each banking business and the key feature set corresponding to each banking business, and the key feature set corresponding to each banking business is determined in advance based on the historical transaction data of each banking business; According to all the key features that need to be audited in the current business step, conduct the audit of the current business step, and complete the current business step when the audit is passed; The method for controlling audit content in banking business processing further includes: determining the key feature set corresponding to each banking business in advance according to the following method based on the historical transaction data of each banking business: Select multiple machine learning models; according to the historical transaction data of each banking business, determine the features corresponding to each banking business; for each feature combination, based on the transaction data regarding the feature combination in the historical transaction data, and using each risk type as a supervision identifier, train the selected multiple machine learning models to obtain multiple risk prediction models corresponding to each risk type; for each risk type involved, determine the error vector corresponding to the feature combination for this risk type, each component of the error vector corresponds one-to-one with the risk prediction model corresponding to this risk type, and the value of each component is equal to the misrecognition rate of the risk prediction model corresponding to this component for this risk type; based on the error vector, determine the partial order of the feature combinations corresponding to each risk type involved, wherein, for any two feature combinations, this partial order can be used to determine whether the first feature combination corresponding to this risk type involved in these two feature combinations is superior to the second feature combination corresponding to this risk type involved in these two feature combinations; according to the partial order of the feature combinations corresponding to each risk type involved, determine multiple maximal feature combinations corresponding to this risk type involved, wherein, this maximal feature combination is the maximal element of this partial order; determine the union of the multiple maximal feature combinations corresponding to all the risk types involved as the key feature set corresponding to this risk type for each banking business; determine the union of the key feature sets corresponding to all the risk types for each banking business as the key feature set corresponding to each banking business.

2. The method for controlling audit content in banking business processing according to claim 1, characterized in that, Determining the partial order of the feature combinations corresponding to each risk type involved based on the error vector includes: For any two feature combinations, determine the difference between the error vector of the first feature combination corresponding to the risk type involved for these two feature combinations and the error vector of the second feature combination corresponding to the risk type involved for these two feature combinations. If each component of this difference is less than or equal to 0, determine that the first feature combination corresponding to the risk type involved is superior to the second feature combination corresponding to the risk type involved.

3. The method for controlling the audit content in banking business handling as claimed in claim 1, wherein, it further includes: pre-determining a set of preferred audit steps for each key feature according to the feature data involved in each business step of each banking business and the set of key features corresponding to each banking business by the following method: Determine the dependency relationships between the various business steps of the banking business; According to the feature data involved in each business step of each banking business and the set of key features corresponding to each banking business, determine the key features involved in each business step of each banking business; According to the key features involved in each business step of each banking business and the dependency relationships between the business steps of each banking business, determine the preferred audit steps for each key feature; Upload the set of preferred audit steps for each key feature to the blockchain.

4. The method for controlling the audit content in banking business handling as claimed in claim 3, wherein, determining the preferred audit steps for each key feature according to the key features involved in each business step of each banking business and the dependency relationships between the business steps of each banking business includes: According to the key features involved in each business step of each banking business, determine a set of potential audit steps for each key feature; For the set of potential audit steps for each key feature, determine whether there is a node in the set of potential audit steps that is an ancestor node of other nodes; wherein, each potential audit step corresponds to a node; When there is a node in the set of potential audit steps that is an ancestor node of other nodes, remove this ancestor node from the set of potential audit steps until there is no ancestor node of other nodes in the set of potential audit steps, and obtain multiple preferred audit steps for each key feature.

5. The method for controlling the audit content in banking business handling as claimed in claim 3, wherein, determining the preferred audit steps for each key feature according to the key features involved in each business step of each banking business and the dependency relationships between the business steps of each banking business includes: According to the dependency relationships between the business steps, construct a step relationship graph corresponding to each banking business, each node of this step relationship graph is a business step of each banking business, there is an edge between two business steps if and only if there is a dependency relationship between the two business steps, and the direction of this edge is from the dependent step to the dependent step, and set the distance of each edge to 1; According to the key features involved in each business step of each banking business, determine a set of potential audit steps for each key feature; For each key feature, initialize the set of alternative business steps as the set of potential audit steps for this key feature; Loop and execute the following steps until there are no two steps in the set of alternative business steps such that the shortest distance between the two steps in the step relationship graph is finite: Find two steps in the set of alternative business steps such that the shortest distance from the first step to the second step among the two steps in the step relationship graph is finite, and delete the first step from the set of alternative business steps; Select business steps that meet the following conditions from all business steps of each banking business: such that the distance from each alternative business step in the set of alternative business steps to this business step is finite; Among the selected multiple business steps, determine the business steps that do not depend on other business steps among the selected multiple business steps, and determine the determined business steps as the preferred review steps for this key feature.

6. The method for controlling review content in banking business handling as described in claim 1, characterized in that, further comprising: Real-time risk monitoring is performed on each banking business according to a pre-established risk warning model; When a new risk type that does not exist in the historical risk type set is monitored, add the new risk type to the multiple risk types involved in each banking business to obtain the updated multiple risk types involved in each banking business; According to the updated multiple risk types involved in each banking business, update the key feature set corresponding to each banking business and the preferred review step set of each key feature; the updated preferred review step set of each key feature is used to control the review content in banking business handling.

7. The method for controlling review content in banking business handling as described in claim 6, characterized in that, further comprising: establishing the risk warning model in advance according to the following method: Establish a risk warning model according to the historical risk data involved in each risk type involved in each banking business; Upload the risk warning model to the blockchain.

8. A device for controlling review content in banking business handling, characterized in that, comprising: A receiving unit, configured to receive a banking business handling request from a user; the banking business includes multiple business steps; A control unit, configured to perform the following operations in each current business step of business handling until all business steps are completed: Determine all key features that need to be reviewed in the current business step according to the preferred review step set of each key feature; the preferred review step set of each key feature is determined in advance according to the feature data involved in each business step of each banking business and the key feature set corresponding to each banking business, and the key feature set corresponding to each banking business is determined in advance according to the historical transaction data of each banking business; According to all key features that need to be reviewed in the current business step, perform the review of the current business step, and complete the current business step when the review is passed. The device for controlling and auditing content in banking business handling further includes: a key feature set determination unit, which is used to pre-determine the key feature set corresponding to each banking business according to the following method based on the historical transaction data of each banking business: select multiple machine learning models; determine the features corresponding to each banking business based on the historical transaction data of each banking business; for each feature combination, based on the transaction data regarding the feature combination in the historical transaction data, and using each risk type as a supervision identifier, train the selected multiple machine learning models to obtain multiple risk prediction models corresponding to each risk type; for each risk type involved, determine the error vector corresponding to the feature combination for this risk type, each component of the error vector corresponds one-to-one with the risk prediction model corresponding to this risk type, and the value of each component is equal to the misrecognition rate of the risk prediction model corresponding to this component for this risk type; based on the error vector, determine the partial order of the feature combinations corresponding to each risk type involved, where, for any two feature combinations, this partial order can be used to determine whether the first feature combination corresponding to this risk type involved in these two feature combinations is superior to the second feature combination corresponding to this risk type involved in these two feature combinations; according to the partial order of the feature combinations corresponding to each risk type involved, determine multiple maximal feature combinations corresponding to this risk type involved, where the maximal feature combination is the maximal element of this partial order; determine the union of the multiple maximal feature combinations corresponding to all risk types involved as the key feature set corresponding to this banking business for this risk type; determine the union of the key feature sets corresponding to all risk types of this banking business as the key feature set corresponding to this banking business.

9. The device for controlling and auditing content in banking business handling according to claim 8, wherein, based on the error vector, determining the partial order of the feature combinations corresponding to each risk type involved includes: for any two feature combinations, determine the difference between the error vector of the first feature combination corresponding to this risk type involved in these two feature combinations and the error vector of the second feature combination corresponding to this risk type involved in these two feature combinations. If each component of this difference is less than or equal to 0, then determine that the first feature combination corresponding to this risk type involved is superior to the second feature combination corresponding to this risk type involved.

10. The device for controlling and auditing content in banking business handling according to claim 8, wherein, it further includes: a preferred audit step set determination unit, which is used to pre-determine the preferred audit step set of each key feature according to the following method based on the feature data involved in each business step of each banking business and the key feature set corresponding to each banking business: determine the dependency relationship between each business step of the banking business; based on the feature data involved in each business step of each banking business and the key feature set corresponding to each banking business, determine the key features involved in each business step of each banking business; Determine the preferred review steps for each key feature based on the key features involved in each business step of each banking service and the dependency relationships between the business steps of each banking service; Upload the set of preferred review steps for each key feature to the blockchain.

11. The device for controlling review content in banking service handling according to claim 10, characterized in that determining the preferred review steps for each key feature based on the key features involved in each business step of each banking service and the dependency relationships between the business steps of each banking service includes: Determine the set of potential review steps for each key feature according to the key features involved in each business step of each banking service; For the set of potential review steps for each key feature, determine whether there is a node in the set of potential review steps that is an ancestor node of other nodes; where each potential review step corresponds to a node; When there is a node in the set of potential review steps that is an ancestor node of other nodes, remove the ancestor node from the set of potential review steps until there is no ancestor node of other nodes in the set of potential review steps, and obtain multiple preferred review steps for each key feature.

12. The device for controlling review content in banking service handling according to claim 10, characterized in that determining the preferred review steps for each key feature based on the key features involved in each business step of each banking service and the dependency relationships between the business steps of each banking service includes: Construct a step relationship graph for each banking service based on the dependency relationships between business steps. Each node of the step relationship graph is a business step of each banking service. There is an edge between two business steps if and only if there is a dependency relationship between the two business steps, and the direction of the edge is from the dependent step to the dependent step, and set the distance of each edge to 1; Determine the set of potential review steps for each key feature according to the key features involved in each business step of each banking service; For each key feature, initialize the set of alternative business steps as the set of potential review steps for the key feature; Loop and execute the following steps until there are no two steps in the set of alternative business steps such that the shortest distance between the two steps in the step relationship graph is finite: Find two steps in the set of alternative business steps such that the shortest distance from the first step to the second step in the step relationship graph is finite, and delete the first step from the set of alternative business steps; Select business steps from all business steps of each banking service that meet the following conditions: the distance from each alternative business step in the set of alternative business steps to the business step is finite; Determine the business steps that do not depend on other business steps among the selected multiple business steps, and determine the determined business steps as the preferred review steps for the key feature.

13. The device for controlling review content in banking service handling according to claim 8, characterized in that further includes: A monitoring unit, configured to perform real-time risk monitoring on each banking business according to a pre-established risk warning model; A first updating unit, configured to add a new risk type, which is not present in the historical risk type set, to the multiple risk types involved in each banking business when the new risk type is monitored, so as to obtain the updated multiple risk types involved in each banking business; A second updating unit, configured to update the key feature set corresponding to each banking business and the preferred review step set of each key feature according to the updated multiple risk types involved in each banking business; the updated preferred review step set of each key feature is used to control the review content during the handling of banking business.

14. The device for controlling review content in banking business handling according to claim 13, wherein, it further includes: a building unit, configured to pre-establish the risk warning model according to the following method: Establish a risk warning model according to the historical risk data involved in each risk type involved in each banking business; Upload the risk warning model to the blockchain.

15. A computer device, including a memory, a processor, and a computer program stored on 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 7 is implemented.

16. 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 method according to any one of claims 1 to 7 is implemented.

17. A computer program product, wherein, the computer program product includes a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images