A method, apparatus, equipment, and medium for account opening verification based on the chain of responsibility model.
By adopting the chain-of-responsibility model for account opening review, parallel completion of multi-dimensional profile data and automated delayed judgment are achieved, solving the problems of low efficiency and low accuracy of risk control in existing technologies, and improving the scalability and availability of the system.
Patent Information
- Application Number
- CN202611283533.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-24
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2046-08-24
AI Technical Summary
Existing account opening review schemes suffer from problems such as delayed judgment, low efficiency, poor scalability, low risk control accuracy, and insufficient resources. They are unable to effectively identify abnormal account opening behavior, resulting in low review efficiency and easy omissions or false interceptions.
The account opening review method based on the chain of responsibility model is adopted. By constructing a delayed judgment parameter object for parallel profile completion, and using the chain of responsibility of whitelist, personalized and general rule nodes for second review and delayed judgment, the parallel processing and automated delayed processing of multi-dimensional profile data are realized.
It improved the overall efficiency and risk control accuracy of account opening review, reduced manpower and time costs, ensured the high scalability and high availability of the system, and avoided task drop-off and false interception.
Smart Images

Figure CN122840901B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of computer technology, and in particular to an account opening verification method, apparatus, equipment and medium based on the chain of responsibility model. Background Technology
[0002] Currently, account opening services in the financial industry (such as securities account opening, bank account opening, insurance application, and loan approval) are subject to regulatory requirements. Auditing institutions (such as securities firms) typically review the information submitted by users, including their risk appetite and financial status, to ensure compliance with regulatory requirements. To improve the accuracy and security of the review process, current account opening procedures usually employ a multi-round review mechanism, including a first-round review of user materials and a second-round review process that proceeds directly to further verification after the first-round review is passed.
[0003] However, in actual account opening processing, not all users who pass the first review are suitable to immediately proceed to the second review stage. For example, some users may have opened accounts in a concentrated period of time, multiple account opening applications may be associated with the same referrer, there may be abnormal clustering at the same branch or channel, or multiple users may submit account opening requests using the same device. These situations require temporary suspension of review. Such users should not be directly entered into the second review pool. Instead, automatic screening and diversion of these users should be carried out after the first review and before the second review process.
[0004] Regarding the aforementioned abnormal account opening behavior, current mainstream solutions have various inherent flaws: The first solution is a direct processing model, where all materials are directly pushed to the second-level review pool after the first review is passed. Delayed judgment relies entirely on manual identification, which is inefficient, prone to overlooking risky customers, and lacks unified and standardized delayed control capabilities. The second solution uses hard-coded conditional judgments, hard-coding the delayed identification logic into the review process code using multi-layered conditional branches (such as if-else statements). This method requires modification of the core code and full regression testing when adding or adjusting conditions (i.e., risk control rules), resulting in poor scalability and maintainability, and low priority. The first approach is unclear and violates the open / closed principle. The second approach introduces an independent rule engine (such as Drools) to implement external rules, which has problems such as heavy component dependencies, high learning costs, difficult debugging, and coarse granularity of rule short-circuiting control. The third approach is a single-point static risk control, which relies solely on blacklists and device fingerprints to directly reject applications with hard rejections, without performing clustered statistics or soft blocking. This approach can only identify isolated / single-point anomalies and cannot identify batch and group account openings by the same referrer, the same business department, the same channel, and the same device in a short period of time, leading to false positives in risk control. Moreover, the blocking results cannot be rolled back, which can easily cause normal users to be mistakenly blocked and result in a poor customer experience.
[0005] In addition, the current account opening review scheme has the following shortcomings: The current scheme requires writing a separate strategy class for each rule, and adding new matching rules requires code modification. Furthermore, the rule dimensions are strongly coupled with user profile fields, resulting in low configuration flexibility. Delayed judgments rely on multiple external data sources (such as branch offices, referrers, banks, devices, and channels) to complete the user profile. The current scheme uses a serial query method to obtain relevant information, which increases the review processing time and affects the response efficiency of the system's review interface. The current scheme uses a uniform time window for risk assessment of all rules, without differentiating the characteristics of different business dimensions. This may lead to missed judgments due to window truncation of cross-period abnormal behaviors such as device reuse, or misjudgments due to over-statistical analysis of non-device dimensions. In high-concurrency account opening scenarios, asynchronous execution of review tasks may result in insufficient task processing resources, such as a full thread pool causing some delayed judgment tasks to fail to execute in a timely manner (e.g., directly discarding the task), thus affecting the integrity of the review process and easily causing some users to be missed. Summary of the Invention
[0006] In view of this, the purpose of this application is to provide an account opening review method, device, equipment, and medium based on the chain of responsibility model, which can ensure that no cases are missed during delayed review and realize automated second-stage delayed processing, thereby saving labor and time costs, improving the overall efficiency and risk control accuracy of account opening review, and ensuring the high scalability and high availability of the account opening review system. The specific solution is as follows: Firstly, this application discloses an account opening verification method based on the chain of responsibility model, applied to an account opening verification system, including: Obtain the account opening records of the target account opening users after the first review is approved, and obtain the user account opening records; the user account opening records include user identifiers and corresponding account opening request identifiers; Based on the user's account opening record, a delayed judgment parameter object is constructed, and a delayed judgment task for the target account opening user is submitted to a preset bounded thread pool. When the bounded thread pool throws a refusal to execute exception, the multi-dimensional profile data of the target account opening user is obtained in the current calling thread, and the delayed judgment parameter object is completed in parallel based on the multi-dimensional profile data through multiple pre-created task groups to obtain the completed user profile; the task group includes multi-dimensional user profile fields. Read the pre-configured set of delayed rules and construct a chain of responsibility containing multiple rule nodes based on the multiple delayed rules in the set; the multiple rule nodes include whitelist rule nodes, personalized rule nodes, and general rule nodes; The delayed judgment task is executed through the chain of responsibility and based on the completed user profile to obtain the second-instance delayed judgment result; the second-instance delayed judgment result includes the review status output by each rule node after executing the delayed judgment, and the review status is any one of the following: delayed status, no delayed status, and no hit status.
[0007] Optionally, the step of reading a pre-configured set of delayed rules and constructing a chain of responsibility containing multiple rule nodes based on multiple delayed rules in the set includes: Determine if a pre-configured set of delayed rules exists in the local cache; If the delayed rule set exists in the local cache, then read the delayed rule set from the local cache; If the delayed rule set does not exist in the local cache, the delayed rule set is read from the preset database; Filter the whitelist rule subset, personalized rule subset and general rule subset from the delayed rule set, and create whitelist rule nodes corresponding to the whitelist rule subset, personalized rule nodes corresponding to the personalized rule subset and general rule nodes corresponding to the general rule subset respectively; The whitelist rule nodes, personalized rule nodes, and general rule nodes are sorted according to execution priority to generate a chain of responsibility; the execution priority is that the whitelist rule nodes take precedence over the personalized rule nodes, and the personalized rule nodes take precedence over the general rule nodes.
[0008] Optionally, each delay rule in the delay rule set includes a main rule configuration object and a rule extension configuration object; The main rule configuration object includes switch identifiers and historical count thresholds for multiple matching dimensions, and the extended rule configuration object is a set of key-value pairs based on rule fields and a list of rule values.
[0009] Optionally, the step of performing the delayed judgment task based on the chain of responsibility and the completed user profile to obtain the second-instance delayed judgment result includes: The whitelist rule nodes in the chain of responsibility determine whether there are any delayed rules in the whitelist rule subset that match the completed user profile; If there is a delay rule in the whitelist rule subset that matches the completed user profile, then the chain of responsibility is terminated, and a second review delay judgment result with the review status being non-delayed is output. If there is no delayed rule in the whitelist rule subset that matches the completed user profile, the review status is output as "not matched". The personalized rule node is used to determine whether each delayed rule in the personalized rule subset matches the completed user profile. If any delayed rule in the subset of personalized rules matches the completed user profile, the chain of responsibility is terminated, and a second review delayed judgment result with the review status being delayed is output. If any delayed rule in the personalized rule subset does not match the completed user profile, the review status is output as "missed". The general rule node is used to determine whether each delayed rule in the general rule subset matches the completed user profile. If any postponement rule in the subset of general rules matches the completed user profile, then the second review postponement judgment result with the review status postponed is output. If any delay rule in the subset of general rules does not match the completed user profile, the second review delay judgment result with the review status set to no delay will be output.
[0010] Optionally, determining whether there exists a delayed rule in the whitelist rule subset that matches the completed user profile includes: By using reflection, all field names in the main configuration object of the rule that are not empty and belong to the preset dimension field set are read from the rule main configuration object located in the whitelist rule subset, and the constraint dimension set of the corresponding delayed rule is obtained. According to the mapping relationship between the field names in the constraint dimension set and the field names of the user profile, the corresponding field values are read from the completed user profile through the reflection mechanism to obtain the first profile field value, and it is determined whether the first profile field value is empty. If the value of the first profile field is empty, it is determined that the current delay rule does not match the completed user profile; The field values in the completed user profile are obtained through reflection mechanism to obtain the second profile field values, and it is determined whether the second profile field values belong to the rule value list in the rule extension configuration object. If the value of the second profile field does not belong to the list of rule values in the rule extension configuration object, it is determined that the current delayed rule does not match the completed user profile. If the value of the second profile field belongs to the list of rule values in the rule extension configuration object, then it is determined that the current delayed rule matches the completed user profile.
[0011] Optionally, determining whether each delayed rule in the subset of personalized rules matches the completed user profile includes: Extract field values corresponding to the switch identifiers in the main configuration object of the rules within the subset of personalized rules from the completed user profile to obtain multiple third profile field values. Combine the multiple third profile field values to obtain statistical query conditions. The third profile field values include target identifier values. If the target identifier value in the third profile field represents a device rule, then the historical account opening records across different submission time windows are statistically analyzed based on the device identifier in the completed user profile to obtain the number of historical account opening records. If the target identifier value in the third profile field is a non-device rule, then the number of historical account opening records within the time range of the front-end account opening data submission is counted, using the front-end account opening data submission time as a window constraint. Query the number of historical account opening records according to the statistical query conditions, and determine whether the number of historical account opening records is greater than the historical counting threshold; If the number of historical account opening records is greater than the historical counting threshold, then it is determined that the current postponement rule matches the completed user profile; If the number of historical account opening records is not greater than the historical counting threshold, then it is determined that the current postponement rule does not match the completed user profile.
[0012] Optionally, the process of performing parallel profile completion on the delayed judgment parameter object based on the multi-dimensional profile data through multiple pre-created task groups further includes: If no user profile data corresponding to any of the user profile fields is found from the multi-dimensional profile data, an error log is recorded for the user profile field, the field value of the corresponding user profile field is assigned to an empty value, and the completion of other user profile fields continues.
[0013] Secondly, this application discloses an account opening verification device based on the chain of responsibility model, applied to an account opening verification system, comprising: The first acquisition module is used to acquire the account opening records of the target account opening users after the first review has been approved, and to obtain the user account opening records; the user account opening records include user identifiers and corresponding account opening request identifiers. The object construction module is used to construct a delayed judgment parameter object based on the user's account opening record; The submission module is used to submit a delayed judgment task for the target account opening user to a preset bounded thread pool. The second acquisition module is used to acquire multi-dimensional profile data of the target account opening user in the current calling thread when the bounded thread pool throws a refusal to execute exception, and to perform parallel profile completion on the delayed judgment parameter object based on the multi-dimensional profile data through multiple pre-created task groups to obtain the completed user profile; the task group includes multi-dimensional user profile fields. The responsibility chain construction module is used to read a pre-configured set of delayed rules and construct a responsibility chain containing multiple rule nodes based on multiple delayed rules in the set of delayed rules; the multiple rule nodes include whitelist rule nodes, personalized rule nodes, and general rule nodes; The execution module is used to perform the delayed judgment task through the chain of responsibility and based on the completed user profile to obtain the second-instance delayed judgment result; the second-instance delayed judgment result includes the review status output by each rule node after performing the delayed judgment, and the review status is any one of the following: delayed status, no delayed status, and no hit status.
[0014] Thirdly, this application discloses an electronic device, including a processor and a memory; wherein, when the processor executes a computer program stored in the memory, it implements the aforementioned account opening verification method based on the chain of responsibility model.
[0015] Fourthly, this application discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned account opening verification method based on the chain of responsibility model.
[0016] As can be seen, this application is applied to an account opening review system, including: obtaining the account opening record of a target account opening user after the first-level review has passed, and obtaining a user account opening record containing a user identifier and a corresponding account opening request identifier; constructing a delayed judgment parameter object based on the user account opening record, and submitting a delayed judgment task for the target account opening user to a preset bounded thread pool; when the bounded thread pool throws a refusal to execute exception, obtaining multi-dimensional profile data of the target account opening user in the current calling thread, and performing parallel profile completion on the delayed judgment parameter object based on the multi-dimensional profile data through multiple pre-created task groups to obtain a supplementary profile. The task group includes a multi-dimensional user profile field; a pre-configured set of delayed rules is read, and a responsibility chain containing multiple rule nodes is constructed based on multiple delayed rules in the set; the multiple rule nodes include whitelist rule nodes, personalized rule nodes, and general rule nodes; the delayed judgment task is executed through the responsibility chain and based on the completed user profile to obtain the second-instance delayed judgment result; the second-instance delayed judgment result includes the review status output by each rule node after performing the delayed judgment, and the review status is any one of delayed status, no delayed status, or no hit status.
[0017] After the first-instance review is passed, this application constructs a delayed judgment parameter object based on the user's account opening record and submits the delayed judgment task to a preset bounded thread pool. When the thread pool throws a refusal to execute exception, the multi-dimensional profile data of the current account opening user is obtained in the current calling thread. Based on the multi-dimensional profile data, the delayed judgment parameter object is completed in parallel through multiple task groups. It can be seen that when the thread pool throws an exception (such as when the thread pool is full), the review system will not directly discard the current delayed judgment task, but will continue to execute the subsequent delayed judgment task in the current calling thread. In this way, it is ensured that no case is missed in the delayed judgment. Furthermore, this application executes the delayed judgment task based on the completed user profile and through the responsibility chain created based on the delayed rule set, and obtains a second-instance delayed judgment result containing one of the three review states. The above method realizes automated second-instance delayed processing, saves manpower and time costs, and improves the overall efficiency and risk control accuracy of account opening review. In addition, this application does not require external components / engines and does not require code modification, thus ensuring the high scalability and high availability of the account opening review system. Attached Figure Description
[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0019] Figure 1 This application discloses a flowchart of an account opening review method based on the chain of responsibility model. Figure 2 This is a schematic diagram of an account opening verification device based on the chain of responsibility model disclosed in this application; Figure 3 This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation
[0020] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0021] This application discloses an account opening review method based on the chain of responsibility model, applied to an account opening review system. See [link to relevant documentation]. Figure 1 As shown, the method includes: Step S11: Obtain the account opening record of the target account opening user after the first review is approved, and obtain the user account opening record; the user account opening record includes the user identifier and the corresponding account opening request identifier.
[0022] It should be noted that the account opening review method proposed in this application is specifically applied to the account opening review scenario after the first review and before the second review, and the entire review process can be realized through an account opening review system deployed on a server.
[0023] In this embodiment, after the review system passes the first review, it obtains the account opening record of the target account opening user who has passed the review, such as the user identifier (userId) and the user's account opening request identifier (reqId), and obtains the user account opening record.
[0024] Step S12: Construct a delayed judgment parameter object based on the user account opening record, and submit the delayed judgment task for the target account opening user to a preset bounded thread pool.
[0025] In this embodiment, after obtaining the user account opening record containing the user identifier (userId) and the account opening request identifier (reqId), a delayed judgment parameter object can be constructed for the target account opening user based on the user account opening record, and the delayed judgment task for the target account opening user can be submitted to the preset bounded thread pool.
[0026] Specifically, a delayed judgment task for the target user can be generated based on the user's account opening record. Then, the delayed judgment task is encapsulated into an asynchronous task object to obtain a delayed judgment parameter object. Finally, the delayed judgment parameter object (containing the delayed judgment task) is submitted to a bounded thread pool (such as a thread pool with 5 core threads, 10 maximum threads, a work queue capacity of 1024, and a named thread factory of "second-audit-delay-").
[0027] Step S13: When the bounded thread pool throws a refusal to execute exception, the multi-dimensional profile data of the target account opening user is obtained in the current calling thread, and the delayed judgment parameter object is completed in parallel based on the multi-dimensional profile data through multiple pre-created task groups to obtain the completed user profile; the task group includes multi-dimensional user profile fields.
[0028] In this embodiment, when the account opening review system detects a Rejected Execution Exception thrown by the bounded thread pool due to the thread pool being full, it does not discard the current task. Instead, it synchronously performs degradation processing within the current calling thread to ensure that no orders are missed during delayed judgment. Specifically, the system first obtains multi-dimensional profile data of the target account opening user within the current calling thread, such as branch number, referrer's mobile phone number, account opening identifier, third-party custodian bank number, customer acquisition type, etc. Then, based on the multi-dimensional profile data, it performs parallel profile completion on the aforementioned delayed judgment parameter objects through three pre-created parallel task groups to obtain the completed user profile; the task groups include multi-dimensional user profile fields.
[0029] The task group is divided into three groups: Task Group 1, Task Group 2, and Task Group 3. Task Group 1 is used to complete user profile fields including but not limited to branch number (branchNo), referrer's mobile phone number (brokerMobile), and channel value (channelValue). Task Group 2 is used to complete user profile fields including but not limited to third-party custodian bank number (bankNo), device identifier (uuid), and Shanghai A account opening identifier (shaSelect). Task Group 3 is used to complete user profile fields including but not limited to whether the user is from another province (outerProvinces), customer acquisition type (vestIn), and front-end account opening information submission time (submitDate).
[0030] It's worth noting that the review system creates an asynchronous task (CompletableFuture) for each task group. The three task groups execute in parallel and merge after all groups complete (allOf().join()). By executing the profile completion operation in parallel across multiple task groups, the efficiency of the completion process is significantly improved, thereby enhancing the overall efficiency of the account opening review.
[0031] It should be noted that asynchronous processing in normal paths will not block the audit interface or reduce response latency; while in abnormal paths, when degrading to synchronous processing, the current task is not discarded and there is no need to introduce message queue middleware.
[0032] It should be noted that the parallel profile completion process for the delayed judgment parameter object based on the multi-dimensional profile data through multiple pre-created task groups specifically includes: if no profile data corresponding to any of the user profile fields is found from the multi-dimensional profile data, an error log is recorded for that user profile field, and the field value of the corresponding user profile field is set to null, while continuing the completion of other user profile fields. In this embodiment, within the profile completion operation of each dimension, the review system adopts independent exception capture; when an exception occurs in the query of the external data source (profile data) for a certain dimension, only an error log is recorded and the field values of all user profile fields for that dimension are set to null, without interrupting the profile completion operation of other dimensions or the overall judgment process, and the subsequent rule matching is performed with that dimension as null. If no profile data corresponding to any of the user profile fields is found from the multi-dimensional profile data, an error log is recorded for that user profile field, and the field value of the corresponding user profile field is set to null, while continuing the completion of other user profile fields.
[0033] In addition, during the user profile completion process, dirty data can be normalized. For example, when the device identifier (UUID) is a placeholder "-", the auditing system normalizes it to a null value, thereby preventing dirty data from causing false hits in subsequent rule matching.
[0034] Step S14: Read the pre-configured set of delayed rules and construct a chain of responsibility containing multiple rule nodes based on the multiple delayed rules in the set of delayed rules; the multiple rule nodes include whitelist rule nodes, personalized rule nodes, and general rule nodes.
[0035] In this embodiment, after performing parallel profile completion on the delayed judgment parameter object to obtain the completed user profile, a pre-configured delayed rule set is further read from the local cache, and then a responsibility chain containing multiple rule nodes is constructed based on multiple delayed rules in the delayed rule set; wherein, the multiple rule nodes include, but are not limited to, whitelist rule nodes, personalized rule nodes, general rule nodes, and other nodes.
[0036] In one specific implementation, the step of reading a pre-configured set of delayed rules and constructing a chain of responsibility containing multiple rule nodes based on multiple delayed rules in the set may specifically include: determining whether a pre-configured set of delayed rules exists in a local cache; if the set of delayed rules exists in the local cache, reading it from the local cache; if the set of delayed rules does not exist in the local cache, reading it from a preset database; filtering a whitelist rule subset, a personalized rule subset, and a general rule subset from the set of delayed rules, and creating whitelist rule nodes corresponding to the whitelist rule subset, personalized rule nodes corresponding to the personalized rule subset, and general rule nodes corresponding to the general rule subset, respectively; sorting the whitelist rule nodes, personalized rule nodes, and general rule nodes according to execution priority to generate a chain of responsibility; wherein the execution priority is that whitelist rule nodes take precedence over personalized rule nodes, and personalized rule nodes take precedence over general rule nodes. In this embodiment, it is first determined whether a pre-configured set of delayed rules exists in the local cache. Each delayed rule in this set (which can be flexibly and dynamically configured and adjusted according to actual application needs) includes a main rule configuration object and extended rule configuration objects. The main rule configuration object includes switch identifiers for multiple matching dimensions (such as whether it is a referrer, a branch, a channel, a bank, a device, a Shanghai A-share, an out-of-province customer, customer acquisition type, etc.) and a historical count threshold. The extended rule configuration object is specifically a set of key-value pairs based on rule fields and a list of rule values. By reading delayed rules from the cache, database pressure can be reduced. It should be noted that if the delayed rule set is not read from the cache, for example, if the cache is empty, the review system can query all delayed rules from the preset database and return them. Next, a whitelist rule subset, a personalized rule subset, and a general rule subset are selected from the delayed rule set. Then, nodes corresponding to each rule subset are created. For example, whitelist rule nodes are created for the whitelist rule subset, personalized rule nodes for the personalized rule subset, and general rule nodes for the general rule subset. The priority of the three nodes is whitelist rule nodes > personalized rule nodes > general rule nodes. Finally, the three nodes are sorted and assembled according to the above priority to obtain the chain of responsibility. For example, the next node pointer of the general rule node is set to null, the next node of the personalized rule node is set to the general rule node, and the next node of the whitelist rule node is set to the personalized rule node, thus forming the chain of responsibility.
[0037] It should be noted that after performing the rule judgment operation, each rule node will return any one of the following three states: First, it returns a delayed state (true): indicating that a hit has been achieved and a delay has been triggered, at which point the responsibility chain is short-circuited and terminated; second, it returns a non-delayed state (false): indicating that an exemption has been achieved (such as a whitelist hit), at which point the responsibility chain is short-circuited and terminated, and the final result is no delay; third, it returns a missed state (null): indicating that the current node has not been hit, and the responsibility chain continues to be passed to the next node.
[0038] Step S15: Execute the delayed judgment task through the chain of responsibility and based on the completed user profile to obtain the second-instance delayed judgment result; the second-instance delayed judgment result includes the review status output by each rule node after performing the delayed judgment, and the review status is any one of the following: delayed status, no delayed status, and no hit status.
[0039] In this embodiment, a responsibility chain including whitelist rule nodes, personalized rule nodes, and general rule nodes is used to perform a delayed judgment task based on the user profile that has been completed, thereby outputting a second-instance delayed judgment result. This result can be any one of the three review states (i.e., true, false, null).
[0040] In this embodiment, the step of performing the delayed judgment task based on the completed user profile through the chain of responsibility to obtain the second-instance delayed judgment result may specifically include: determining whether there is a delayed rule matching the completed user profile in the whitelist rule subset through the whitelist rule nodes in the chain of responsibility; if there is a delayed rule matching the completed user profile in the whitelist rule subset, then the chain of responsibility is terminated, and the second-instance delayed judgment result with a review status of no delay is output; if there is no delayed rule matching the completed user profile in the whitelist rule subset, then the review status is output as a missed state, and the delayed rule in each of the personalized rule subsets is determined by the personalized rule nodes to determine whether it matches the completed user profile. If any delayed rule in the personalized rule subset matches the completed user profile, the chain of responsibility terminates, and a second-review delayed judgment result with a delayed review status is output. If any delayed rule in the personalized rule subset does not match the completed user profile, the review status is output as "missed," and the general rule nodes determine whether each delayed rule in the general rule subset matches the completed user profile. If any delayed rule in the general rule subset matches the completed user profile, a second-review delayed judgment result with a delayed review status is output. If any delayed rule in the general rule subset does not match the completed user profile, a second-review delayed judgment result with a non-delayed review status is output. In this embodiment, the whitelist rule nodes in the responsibility chain first match each whitelist rule in the whitelist rule subset with the completed user profile. If the user profile matches any whitelist rule, the whitelist rule node returns a no-delay state (false), at which point the responsibility chain short-circuits and outputs a no-delay judgment result for the second review. If the whitelist rule subset is empty or none of them match, the user profile is written into the delay statistics record table (for subsequent threshold counting), and the whitelist rule node returns a no-hit state (null), at which point the responsibility chain is passed to the personalized rule node. The personalized rule node first determines whether any delayed rule in the personalized rule subset matches the completed user profile. If the completed user profile matches any delayed rule in the personalized rule subset, the responsibility chain terminates and outputs a second-review delay judgment result with a delayed status (true); if the completed user profile does not match any delayed rule in the personalized rule subset, the responsibility chain is passed to the general rule node. The general rule node uses the same judgment method as the personalized rule node to determine whether each delayed rule in the general rule subset matches the completed user profile. If it matches, it outputs the second review delayed judgment result with the review status of delayed status (true). If it does not match, it outputs the second review delayed judgment result with the review status of no delay status (false).
[0041] In this embodiment, the step of determining whether there is a deferral rule in the whitelist rule subset that matches the supplemented user profile may specifically include: reading, through a reflection mechanism, all field names that have non-null values and belong to a preset dimension field set in the rule main configuration objects located in the whitelist rule subset, to obtain a constraint dimension set of the corresponding deferral rule; according to the mapping relationship between the field names in the constraint dimension set and the user profile field names, reading the corresponding field values from the supplemented user profile through the reflection mechanism to obtain first profile field values, and determining whether the first profile field values are null values; if the first profile field value is a null value, determining that the current deferral rule does not match the supplemented user profile; acquiring the field values in the supplemented user profile through the reflection mechanism to obtain second profile field values, and determining whether the second profile field values belong to the rule value list in the rule extended configuration object; if the second profile field value does not belong to the rule value list in the rule extended configuration object, determining that the current deferral rule does not match the supplemented user profile; if the second profile field value belongs to the rule value list in the rule extended configuration object, determining that the current deferral rule matches the supplemented user profile. In this embodiment, the reflection mechanism is first used to read all field names that have non-null values and belong to the preset dimension field set in each rule main configuration object located in the whitelist rule subset, so as to obtain the constraint dimension set of the corresponding deferral rule. Then, according to the mapping relationship between the field names in the constraint dimension set (excluding whether the user is from another province) and the user profile field names, the reflection mechanism is used to obtain the corresponding field values in the supplemented user profile to obtain the first profile field value, and it is determined whether the first profile field value is a null value. If the field value is a null value, it is determined that the current deferral rule is not matched. Next, the rule extended configuration objects are grouped according to rule fields into a mapping relationship of rule field-rule value list, and for each group, the reflection mechanism is used to obtain the corresponding field value in the user profile, and it is determined whether the field value belongs to the rule value list in the extended configuration object. If it does not belong, it is determined that there is no match; if it belongs, it is determined that the user profile matches the current deferral rule.
[0042] It should be noted that before the audit system performs the comparison of the rule main configuration objects, semantic filtering needs to be performed first. For example, when the user customer acquisition type is an offline channel (such as an employee or a construction site, etc.), and the rule does not configure a referrer but the user profile shows that there is a referrer, it is determined as a mismatch; when the rule configures the customer acquisition type but the customer acquisition type shown in the user profile is not within the configuration, it is determined as a mismatch; when the rule configures the Shanghai A license plate identifier but the Shanghai A license plate identifier shown in the user profile does not match, it is determined as a mismatch. In addition, the rule main configuration object and the extended configuration object can be flexibly configured and adjusted according to actual application scenarios and business requirements.
[0043] In this embodiment, determining whether each delayed rule in the personalized rule subset matches the completed user profile may specifically include: extracting field values corresponding to the switch identifier in the main configuration object of the rules within the personalized rule subset from the completed user profile to obtain multiple third profile field values, and combining the multiple third profile field values to obtain statistical query conditions; the third profile field values include target identifier values; if the target identifier value in the third profile field values represents a device rule, then the historical account opening records across different submission time windows are statistically analyzed based on the device identifier in the completed user profile. The historical account opening record count is obtained. If the target identifier value in the third profile field represents a non-device rule, the historical account opening record count within the front-end account opening data submission time range is counted using the front-end account opening data submission time as a window constraint. The historical account opening record count is queried according to the statistical query conditions, and it is determined whether the historical account opening record count is greater than the historical counting threshold. If the historical account opening record count is greater than the historical counting threshold, it is determined that the current postponement rule matches the completed user profile. If the historical account opening record count is not greater than the historical counting threshold, it is determined that the current postponement rule does not match the completed user profile. In this embodiment, when a user profile matches a delayed rule within a subset of personalized rules, the corresponding field values (including the target identifier value) can be extracted from the user profile based on the various dimension switches in the main rule configuration object (such as any combination of one or more of the following: referrer, business department, channel, bank, device, Shanghai A, out-of-province account, and customer acquisition type). This yields the third profile field value, and the extracted third profile field values are then combined in multiple dimensions to construct statistical query conditions, i.e., the statistical dimension DTO (Data Transfer Object). Next, it is determined whether the target identifier value (uuid) in the third profile field value represents a device rule. If so, no submission time window constraint is set. In this case, all historical account opening records across different submission time windows can be counted based on the device identifier (i.e., all historical account opening records are counted across periods for the device identifier, since device reuse is a cross-period behavior), to obtain the number of historical account opening records. If the target identifier value (uuid) in the third profile field value represents a non-device rule, the submission date of the front-end account opening data is used as the window constraint to count the number of historical account opening records within that time range.
[0044] It should be noted that when the channel switch of the main rule configuration object is turned on, the list of channel values in the extended rule configuration object can be used as the channel condition for statistical queries (i.e., multi-channel combined counting); when the channel switch is turned off, the channel value corresponding to the user profile can be used as the condition (i.e., single-channel counting).
[0045] Furthermore, the system queries the historical account opening record count according to the statistical query conditions and determines whether the historical account opening record count is greater than the historical counting threshold. If the historical account opening record count is greater than the historical counting threshold, the current postponement rule is directly determined to match the completed user profile, that is, the postponement condition is met. At this time, a soft interception operation can be performed. For example, the review status field of the user account opening process record is updated to the postponement status (true), so that it does not enter the second review pool; the review log is written to record the user identifier, request identifier, review round, system identifier, and the second review postponement judgment result as the postponement status (true); the current rule node returns the postponement status (true), and the responsibility chain is short-circuited and terminated. If none of the rules of the current node reach the threshold, such as the historical account opening record count not being greater than the historical counting threshold, the current postponement rule is determined to not match the completed user profile. At this time, the node (such as the personalized rule node) returns the miss status (null), or the general rule node (chain tail node) returns the no postponement status (false).
[0046] In this embodiment, batch / group account opening is identified by combining multi-dimensional combinations with historical counting thresholds, and soft blocking (setting a delayed state) is used instead of hard rejection, which can simultaneously take into account the risk control needs of securities account opening business and user experience; the counting window is set according to the semantic differences of the dimensions, so that the cross-period behavior of device reuse is not missed and non-device dimensions are not over-counted; in addition, the channel merging count supports flexible expression of business.
[0047] In addition, the review system provides a manual cancellation interface. When reviewers manually review a delayed status record and execute the cancellation operation, the review system will change its review status from delayed (true) to normal, allowing it to enter the second review pool. This demonstrates that the review status can be effectively terminated. Therefore, a delayed status is an intermediate state that can be cancelled, while a rejection is a final state requiring a new application.
[0048] Furthermore, the account opening review method proposed in this application is not limited to securities account opening scenarios, but can also be extended to multiple review scenarios such as bank account opening, insurance application, and credit approval.
[0049] As can be seen, in this embodiment, after the first-instance review is passed, a delayed judgment parameter object is constructed based on the user's account opening record, and a delayed judgment task is submitted to a preset bounded thread pool. When the thread pool throws a refusal to execute exception, the multi-dimensional profile data of the current account opening user is obtained in the current calling thread, and the delayed judgment parameter object is completed in parallel based on the multi-dimensional profile data through multiple task groups. It can be seen that in this embodiment, when the thread pool throws an exception (such as when the thread pool is full), the review system will not directly discard the current delayed judgment task, but will continue to execute the subsequent delayed judgment task in the current calling thread. In this way, it is ensured that no case is missed in the delayed judgment. Furthermore, in this embodiment, the delayed judgment task is executed according to the completed user profile and through the responsibility chain created based on the delayed rule set, to obtain a second-instance delayed judgment result containing one of the three review states. The above method realizes automated second-instance delayed processing, saves manpower and time costs, and improves the overall efficiency and risk control accuracy of account opening review. In addition, this embodiment does not require the use of external components / engines and does not require code modification, thereby ensuring the high scalability and high availability of the account opening review system.
[0050] Accordingly, this application also discloses an account opening verification device based on the chain of responsibility model, applied to an account opening verification system. See [link to relevant documentation]. Figure 2 As shown, the device includes: The first acquisition module 11 is used to acquire the account opening record of the target account opening user after the first review is approved, and obtain the user account opening record; the user account opening record includes a user identifier and a corresponding account opening request identifier. Object construction module 12 is used to construct a delayed judgment parameter object based on the user account opening record; Submission module 13 is used to submit a delayed judgment task for the target account opening user to a preset bounded thread pool; The second acquisition module 14 is used to acquire multi-dimensional profile data of the target account opening user in the current calling thread when the bounded thread pool throws a refusal to execute exception, and to perform parallel profile completion on the delayed judgment parameter object based on the multi-dimensional profile data through multiple pre-created task groups to obtain the completed user profile; the task group includes multi-dimensional user profile fields. The responsibility chain construction module 15 is used to read a pre-configured set of delayed rules and construct a responsibility chain containing multiple rule nodes based on multiple delayed rules in the set of delayed rules; the multiple rule nodes include whitelist rule nodes, personalized rule nodes, and general rule nodes. Execution module 16 is used to execute the delayed judgment task through the chain of responsibility and based on the completed user profile to obtain the second review delayed judgment result; the second review delayed judgment result includes the review status output by each rule node after executing the delayed judgment, and the review status is any one of the following: delayed status, no delayed status, and no hit status.
[0051] The specific workflow of each of the above modules can be found in the relevant content disclosed in the foregoing embodiments, and will not be repeated here.
[0052] Furthermore, embodiments of this application also disclose an electronic device, Figure 3 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application.
[0053] Figure 3 This is a schematic diagram of the structure of an electronic device 20 provided in an embodiment of this application. Specifically, the electronic device 20 may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the account opening verification method based on the chain of responsibility model disclosed in any of the foregoing embodiments. Alternatively, the electronic device 20 in this embodiment may specifically be an electronic computer.
[0054] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.
[0055] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.
[0056] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the account opening verification method based on the chain of responsibility model executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.
[0057] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned account opening verification method based on the chain of responsibility model. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.
[0058] Furthermore, embodiments of this application also disclose a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the account opening review method based on the chain of responsibility model disclosed above.
[0059] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.
[0060] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0061] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.
[0062] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0063] The foregoing has provided a detailed description of the account opening review method, apparatus, equipment, and medium based on the chain of responsibility model provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.
Claims
1. An account opening verification method based on the chain of responsibility model, characterized in that, Applications to the account opening verification system include: Obtain the account opening records of the target account opening users after the first review is approved, and obtain the user account opening records; the user account opening records include user identifiers and corresponding account opening request identifiers; Based on the user's account opening record, a delayed judgment parameter object is constructed, and a delayed judgment task for the target account opening user is submitted to a preset bounded thread pool. When the bounded thread pool throws a refusal to execute exception, the multi-dimensional profile data of the target account opening user is obtained in the current calling thread, and the delayed judgment parameter object is completed in parallel based on the multi-dimensional profile data through multiple pre-created task groups to obtain the completed user profile; the task group includes multi-dimensional user profile fields. Read a pre-configured set of delayed rules and construct a chain of responsibility containing multiple rule nodes based on multiple delayed rules in the set; the multiple rule nodes include whitelist rule nodes, personalized rule nodes, and general rule nodes; the execution priority of the multiple rule nodes is that the whitelist rule nodes take precedence over the personalized rule nodes, and the personalized rule nodes take precedence over the general rule nodes. The delayed judgment task is executed through the chain of responsibility and based on the completed user profile to obtain the second-instance delayed judgment result; the second-instance delayed judgment result includes the review status output by each rule node after executing the delayed judgment, and the review status is any one of the following: delayed status, no delayed status, and no hit status; The step of performing the delayed judgment task based on the completed user profile through the chain of responsibility to obtain the second-instance delayed judgment result includes: determining whether there is a delayed rule matching the completed user profile in the whitelist rule subset through the whitelist rule nodes in the chain of responsibility; if there is a delayed rule matching the completed user profile in the whitelist rule subset, then the chain of responsibility is terminated, and the second-instance delayed judgment result with a review status of no delay is output; if there is no delayed rule matching the completed user profile in the whitelist rule subset, then the review status is output as a missed state, and each delayed rule in the personalized rule subset is determined by the personalized rule nodes to match the completed user profile; if the If any delayed rule in the personalized rule subset matches the completed user profile, the chain of responsibility terminates, and a second-review delayed judgment result with a delayed review status is output. If any delayed rule in the personalized rule subset does not match the completed user profile, the review status is output as "missed," and the general rule nodes determine whether each delayed rule in the general rule subset matches the completed user profile. If any delayed rule in the general rule subset matches the completed user profile, a second-review delayed judgment result with a delayed review status is output. If any delayed rule in the general rule subset does not match the completed user profile, a second-review delayed judgment result with a non-delayed review status is output.
2. The account opening verification method based on the chain of responsibility model according to claim 1, characterized in that, The step of reading a pre-configured set of delayed rules and constructing a chain of responsibility containing multiple rule nodes based on multiple delayed rules in the set includes: Determine if a pre-configured set of delayed rules exists in the local cache; If the delayed rule set exists in the local cache, then read the delayed rule set from the local cache; If the delayed rule set does not exist in the local cache, the delayed rule set is read from the preset database; Filter the whitelist rule subset, personalized rule subset and general rule subset from the delayed rule set, and create whitelist rule nodes corresponding to the whitelist rule subset, personalized rule nodes corresponding to the personalized rule subset and general rule nodes corresponding to the general rule subset respectively; The whitelist rule nodes, personalized rule nodes, and general rule nodes are sorted according to the execution priority to generate a chain of responsibility.
3. The account opening verification method based on the chain of responsibility model according to claim 2, characterized in that, Each delay rule in the delay rule set contains a main rule configuration object and a rule extension configuration object; The main rule configuration object includes switch identifiers and historical count thresholds for multiple matching dimensions, and the extended rule configuration object is a set of key-value pairs based on rule fields and a list of rule values.
4. The account opening verification method based on the chain of responsibility model according to claim 3, characterized in that, The step of determining whether there is a delayed rule in the whitelist rule subset that matches the completed user profile includes: By using reflection, all field names in the main configuration object of the rule that are not empty and belong to the preset dimension field set are read from the rule main configuration object located in the whitelist rule subset, and the constraint dimension set of the corresponding delayed rule is obtained. According to the mapping relationship between the field names in the constraint dimension set and the field names of the user profile, the corresponding field values are read from the completed user profile through the reflection mechanism to obtain the first profile field value, and it is determined whether the first profile field value is empty. If the value of the first profile field is empty, it is determined that the current delay rule does not match the completed user profile; The field values in the completed user profile are obtained through reflection mechanism to obtain the second profile field values, and it is determined whether the second profile field values belong to the rule value list in the rule extension configuration object. If the value of the second profile field does not belong to the list of rule values in the rule extension configuration object, it is determined that the current delayed rule does not match the completed user profile. If the value of the second profile field belongs to the list of rule values in the rule extension configuration object, then it is determined that the current delayed rule matches the completed user profile.
5. The account opening verification method based on the chain of responsibility model according to claim 1, characterized in that, The step of determining whether each delayed rule in the personalized rule subset matches the completed user profile includes: Extract field values corresponding to the switch identifiers in the main configuration object of the rules within the subset of personalized rules from the completed user profile to obtain multiple third profile field values. Combine the multiple third profile field values to obtain statistical query conditions. The third profile field values include target identifier values. If the target identifier value in the third profile field represents a device rule, then the historical account opening records across different submission time windows are statistically analyzed based on the device identifier in the completed user profile to obtain the number of historical account opening records. If the target identifier value in the third profile field is a non-device rule, then the number of historical account opening records within the time range of the front-end account opening data submission is counted, using the front-end account opening data submission time as a window constraint. Query the number of historical account opening records according to the statistical query conditions, and determine whether the number of historical account opening records is greater than the historical counting threshold; If the number of historical account opening records is greater than the historical counting threshold, then it is determined that the current postponement rule matches the completed user profile; If the number of historical account opening records is not greater than the historical counting threshold, then it is determined that the current postponement rule does not match the completed user profile.
6. The account opening verification method based on the chain of responsibility model according to any one of claims 1 to 5, characterized in that, The process of performing parallel profile completion on the delayed judgment parameter object based on the multi-dimensional profile data through multiple pre-created task groups also includes: If no user profile data corresponding to any of the user profile fields is found from the multi-dimensional profile data, an error log is recorded for the user profile field, the field value of the corresponding user profile field is assigned to an empty value, and the completion of other user profile fields continues.
7. An account opening verification device based on the chain of responsibility model, characterized in that, Applications to the account opening verification system include: The first acquisition module is used to acquire the account opening records of the target account opening users after the first review has been approved, and to obtain the user account opening records; the user account opening records include user identifiers and corresponding account opening request identifiers. The object construction module is used to construct a delayed judgment parameter object based on the user's account opening record; The submission module is used to submit a delayed judgment task for the target account opening user to a preset bounded thread pool. The second acquisition module is used to acquire multi-dimensional profile data of the target account opening user in the current calling thread when the bounded thread pool throws a refusal to execute exception, and to perform parallel profile completion on the delayed judgment parameter object based on the multi-dimensional profile data through multiple pre-created task groups to obtain the completed user profile; the task group includes multi-dimensional user profile fields. The responsibility chain construction module is used to read a pre-configured set of delayed rules and construct a responsibility chain containing multiple rule nodes based on multiple delayed rules in the set of delayed rules. The multiple rule nodes include whitelist rule nodes, personalized rule nodes, and general rule nodes. The execution priority of the multiple rule nodes is that the whitelist rule nodes take precedence over the personalized rule nodes, and the personalized rule nodes take precedence over the general rule nodes. The execution module is used to execute the delayed judgment task through the chain of responsibility and based on the completed user profile to obtain the second review delayed judgment result; the second review delayed judgment result includes the review status output by each rule node after executing the delayed judgment, and the review status is any one of the following: delayed status, no delayed status, and no hit status; The execution module is specifically used to determine, through the whitelist rule nodes in the responsibility chain, whether there is a delayed rule in the whitelist rule subset that matches the completed user profile; if there is a delayed rule in the whitelist rule subset that matches the completed user profile, then the responsibility chain is terminated, and a second review delayed judgment result with a review status of no delay is output; if there is no delayed rule in the whitelist rule subset that matches the completed user profile, then the review status is output as a missed result, and the personalized rule nodes are used to determine whether each delayed rule in the personalized rule subset matches the completed user profile; if any delayed rule in the personalized rule subset matches the completed user profile... If the completed user profile matches the completed user profile, the chain of responsibility is terminated, and a second-review postponement judgment result with a postponed review status is output. If any postponement rule in the personalized rule subset does not match the completed user profile, the review status is output as a missed match, and the general rule nodes are used to determine whether each postponement rule in the general rule subset matches the completed user profile. If any postponement rule in the general rule subset matches the completed user profile, a second-review postponement judgment result with a postponed review status is output. If any postponement rule in the general rule subset does not match the completed user profile, a second-review postponement judgment result with a non-postponed review status is output.
8. An electronic device, characterized in that, It includes a processor and a memory; wherein, when the processor executes a computer program stored in the memory, it implements the account opening review method based on the chain of responsibility model as described in any one of claims 1 to 6.
9. A computer-readable storage medium, characterized in that, Used to store computer programs; wherein, when the computer programs are executed by a processor, they implement the account opening review method based on the chain of responsibility model as described in any one of claims 1 to 6.
Citation Information
Patent Citations
User service auditing request processing method and device, equipment, medium and product
CN121436617A
Enterprise human property data and business scene linkage method
CN121810239A
