Supplier abnormity identification and supervision system

Abstract

The invention relates to the technical field of enterprise management, in particular to a supplier anomaly identification and supervision system. The system comprises an IP abnormity identification module which is used for identifying whether an IP login address of a supplier account is abnormal or not and executing quotation operation when the IP login address is abnormal, and if yes, triggering first risk early warning and sending first early warning information to a supervision unit; the subject identification module is used for identifying whether suppliers in the purchase contract meet performance conditions or not, and if not, triggering second risk early warning and sending second early warning information to the supervision unit; and the associated risk identification module is used for identifying whether the supplier is associated with a staff or other suppliers and identifying whether tendency exists between a supplier evaluator and the supplier evaluated by the supplier evaluator, and if the identification result of any one of the supplier evaluator and the supplier evaluated by the supplier evaluator is yes, third risk early warning is triggered, and third early warning information is sent to the supervision unit. The supplier supervision efficiency is improved, and the risk management and control level of the purchasing link is improved.

Description

Technical Field

[0001] This application relates to the field of enterprise management technology, and in particular to a supplier anomaly identification and monitoring system. Background Technology

[0002] In bidding activities, suppliers are the core entities for project implementation, and their qualifications, integrity records, and performance directly affect the quality, cost, and schedule of the bidding project. With the widespread application of bidding activities in government procurement, engineering construction, and material procurement, the size and complexity of the supplier group have increased significantly—including large state-owned enterprises, listed companies, and other highly standardized entities, as well as diverse participants such as small and medium-sized enterprises and emerging technology companies.

[0003] However, current supplier management often suffers from problems such as "emphasizing entry requirements over process" and "emphasizing results over supervision," which poses multiple challenges to the fairness, transparency, and final project outcomes of the bidding market.

[0004] Existing supervision relies heavily on manual verification (such as expert review of documents and on-site inspections), which is inefficient and highly subjective. Although some projects have introduced electronic bidding systems, their functions are limited to online processes (such as uploading bid documents and bid opening and evaluation), and they lack the ability to intelligently analyze and provide risk warnings for the supplier's behavior throughout the entire lifecycle (such as abnormal bidding strategies and logistics / progress data during the performance of the contract). Summary of the Invention

[0005] This application provides a supplier anomaly identification and monitoring system, including: The IP anomaly identification module is used to identify whether the supplier account has an abnormal IP login address and to perform a quotation operation when the IP login address is abnormal. If so, the first risk warning is triggered and the first warning information is sent to the supervision unit. The abnormal IP login address is defined as the IP login address being an address other than the frequently used IP address. The entity identification module is used to identify whether the supplier in the procurement contract meets the performance conditions. If not, it triggers a second risk warning and sends a second warning message to the supervision unit. The associated risk identification module is used to identify whether the supplier is associated with the employee or other suppliers, and to identify whether there is any bias between the supplier evaluator and the supplier being evaluated. If the identification result of either is yes, a third risk warning is triggered and a third warning message is sent to the supervision unit.

[0006] In some embodiments, the associated risk identification module further includes: The internal and external correlation identification unit is used to obtain employee information from the human resources management system and enterprise personnel information of the supplier, identify whether the employee information and enterprise personnel information are consistent, and if so, determine that there is an internal and external correlation. The external association identification unit is used to obtain newly added supplier information in the supplier management system and compare whether the newly added supplier information is consistent with the existing supplier information in the supplier management system. If so, it is determined that there is an external association. The bias identification unit is used to identify whether, within a preset period, the supplier evaluator and the supplier they evaluate appear simultaneously a preset number of times and the frequency of the highest evaluation score exceeds a preset frequency threshold. If so, it is determined that there is a bias between the supplier evaluator and the supplier.

[0007] In some embodiments, the bias identification unit associates products in the procurement platform with their respective suppliers and supplier evaluators, and then performs grouped queries on suppliers and supplier evaluators that appear together a preset number of times. Specifically, a three-element relationship structure of product-supplier-supplier evaluator is formed. The structure is then grouped and statistically analyzed using SQL or distributed query language to filter out evaluators who appear together with a certain supplier more than a preset number of times within a preset period. Based on their rating distribution, it is determined whether there is a biased behavior where the scores are concentrated at the highest level.

[0008] In some embodiments, the subject identification module is further configured to: Obtain the supplier's registered capital and number of employees covered by social security; The registered capital amount is compared with the purchase contract amount corresponding to the supplier. If the registered capital amount is lower than the product of the purchase contract amount and a preset ratio, the first performance capability is determined to be insufficient. The purchase contract amount is based on the pre-stored information in the contract management system and can be obtained by indexing information such as the contract party, contract number, and contract name, or by extracting it based on the contract text combined with OCR recognition technology.

[0009] If the number of people covered by social security is lower than the set threshold, then the second performance capacity is deemed insufficient. If the first or second performance capability is insufficient, the supplier is deemed not to meet the performance conditions.

[0010] In some embodiments, the system establishes a data interface with an enterprise information query platform and a contract management system to obtain the supplier's enterprise information. The entity identification module sends the supplier's name to the enterprise information query platform to retrieve its registered capital, number of employees covered by social security, and enterprise personnel information.

[0011] In some embodiments, the employee information is a set of employee names and the names of their close relatives, and the enterprise personnel information includes a set of legal representative names, shareholder names, and senior executive names.

[0012] In some embodiments, if the number of names that match any employee's name and the names of their close relatives with the names in the company's personnel information exceeds a preset number, then the employee information and the company's personnel information are identified as consistent.

[0013] In some embodiments, the new supplier information includes the bank account and contact information provided by the new supplier during the registration phase; If the bank account information or contact information in the existing supplier information is consistent with that in the newly added supplier information, it is determined that there is an external connection.

[0014] In another embodiment, the newly added supplier information may also include a set of shareholder names in the equity relationship. The set of shareholder names can be obtained from the enterprise information query platform. The parameter "new supplier name" is passed to the enterprise information query platform to retrieve its equity relationship. The equity relationship of the new supplier and the equity relationship of the existing supplier are analyzed using a graph neural network to reveal the hidden supplier association chain, thereby further improving the identification accuracy.

[0015] In some embodiments, the supervision unit includes: The first supervision unit is used to receive the first early warning information and trigger the business supervision processing flow. The second supervisory unit is used to receive the second early warning information and trigger the compliance processing procedure. The third supervisory unit is used to receive the third early warning information and trigger the audit processing flow.

[0016] In some embodiments, the IP anomaly identification module, the subject identification module, and the associated risk identification module are each configured with an execution frequency, which is configured to be once a day.

[0017] Compared to related technologies, the supplier anomaly identification and monitoring system provided in this application, through cross-comparison of multi-source data, can achieve full-process monitoring of supplier behavior, improving the risk control level of the procurement process; the IP anomaly identification module can promptly detect suppliers maliciously changing login locations to forge bidding behavior, preventing bid rigging or human-machine separation operations; the entity identification module focuses on the supplier's capability issues and abnormal activities during the performance of cooperation agreements to ensure the smooth operation of procurement and ensure the quality of performance; the association risk identification module identifies the internal relationships between suppliers and procurement enterprises, effectively preventing risks such as the transfer of benefits and unfair transactions caused by related relationships. The entire system is data-driven, improving regulatory efficiency, reducing the burden of manual review, and enhancing the timeliness and accuracy of early warnings.

[0018] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0019] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a structural block diagram of the supplier anomaly identification and monitoring system according to an embodiment of this application; Figure 2 This is a structural block diagram of the associated risk identification module according to an embodiment of this application; Figure 3 This is a logical schematic diagram of the subject identification module in an embodiment of this application; Figure 4 This is a connection diagram of the supplier anomaly identification and monitoring system according to an embodiment of this application.

[0020] In the picture: 1. IP anomaly detection module; 2. Subject identification module; 3. Association risk identification module; 301. Internal and external association identification unit; 302. External association identification unit; 303. Tendency identification unit. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0022] Obviously, the accompanying drawings described below are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar scenarios based on these drawings without any inventive effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, any changes to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.

[0023] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0024] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “a,” “an,” “an,” “the,” and similar words used in this application do not indicate quantity limitation and may indicate singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. “Multiple” in this application refers to two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. The character “ / ” generally indicates that the preceding and following related objects are in an “or” relationship. The terms “first,” “second,” “third,” etc., used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects.

[0025] This embodiment provides a supplier anomaly identification and monitoring system, which is used to implement the above embodiments and preferred embodiments. Details already described will not be repeated. As used below, the terms "module," "unit," "subunit," etc., can refer to a combination of software and / or hardware that performs a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0026] Figure 1 This is a structural block diagram of a supplier anomaly identification and monitoring system according to an embodiment of this application, applied to an enterprise material procurement platform. (Refer to...) Figure 1 As shown, the system includes: IP anomaly identification module 1 is used to identify whether the supplier account has an abnormal IP login address and to perform a quotation operation when the IP login address is abnormal. If so, it triggers the first risk warning and sends the first warning information to the supervision unit. Herein, an abnormal IP login address means that the IP login address is an address other than the frequently used IP address. The entity identification module 2 is used to identify whether the supplier in the procurement contract meets the performance conditions. If not, it triggers a second risk warning and sends a second warning message to the supervision unit. The associated risk identification module 3 is used to identify whether the supplier is associated with the employee or other suppliers, and to identify whether there is any bias between the supplier evaluator and the supplier being evaluated. If the identification result of either is yes, a third risk warning is triggered and a third warning message is sent to the supervision unit.

[0027] In some embodiments, the IP anomaly identification module 1, the subject identification module 2, and the associated risk identification module 3 are each configured with an execution frequency, which is configured to be once a day. Optionally, the execution frequency of each module and its sub-units can also be configured separately.

[0028] Based on the above steps, this application achieves full-process supervision of supplier behavior through cross-comparison of multi-source data, improving the risk control level of the procurement process; the IP anomaly identification module can promptly detect suppliers maliciously changing login locations to forge bidding behavior, preventing bid rigging or human-machine separation operations; the entity identification module monitors the supplier's capabilities and abnormal activities during the performance of cooperation agreements to ensure the smooth operation of procurement and guarantee the quality of performance; the association risk identification module identifies the internal relationships between suppliers and procurement companies, effectively preventing risks such as the transfer of benefits and unfair transactions caused by related relationships. The entire system is data-driven, improving regulatory efficiency, reducing the burden of manual review, and enhancing the timeliness and accuracy of early warnings.

[0029] In the above embodiments, the frequently logged-in IP address is obtained by acquiring the historical IP login address of the supplier account. If there are multiple IP login addresses, the province to which the IP login address with the largest number of logins belongs is identified, and the province is used as the frequently logged-in address. If the IP login address is from a province other than the frequently logged-in address, an alert is issued.

[0030] In one possible implementation, the IP anomaly identification module 1 identifies the frequently used login addresses of supplier accounts by combining the number of consecutive active days of the IP login address and the login time period for weighted analysis. This includes: obtaining all historical IP login addresses of the target supplier account within a certain number of days; for each IP address, counting the number of consecutive active days it appears and the number of logins during typical working hours (such as 9:00 to 18:00), and setting weight coefficients for different indicators for comprehensive scoring; determining the province of the IP address with the highest score as the frequently used address, and triggering the first risk warning when a login occurs outside that province and a quotation operation is performed.

[0031] This implementation method comprehensively considers the number of consecutive active days of a supplier account on a certain IP address and whether the login behavior is concentrated on weekdays and during working hours. The system sets a preset number of days (e.g., 30 days) and counts the login behavior of each IP in turn: if an IP has login behavior for N consecutive days (e.g., ≥3 days), or its login behavior is mainly concentrated in the working hours of 9:00-18:00, then the IP is determined to be a relatively real or natural login address of the supplier; corresponding weights are assigned according to "consecutive active days" and "login frequency during working hours" (e.g., active days weight is 0.6, login time weight is 0.4), and the comprehensive score of each IP address is calculated; the province corresponding to the IP address with the highest score is selected as the "frequently used address".

[0032] In this way, this implementation method can more accurately reflect the supplier's true operating location and behavioral characteristics, avoiding misjudgments caused by temporary bulk logins. By introducing continuity and time period characteristics, the ability to distinguish abnormal behavior can be significantly enhanced, making it particularly suitable for identifying covert operations such as spoofed locations and logins using other people's accounts. At the same time, this method improves the robustness and reliability of "frequently used address" identification, making the early warning mechanism more intelligent and accurate, thereby improving the overall risk identification level of the system.

[0033] In the above scheme, the threshold for consecutive active days can be adaptively adjusted according to different business scenarios. For example, the cycle can be shortened and sensitivity improved during centralized procurement of projects. The range of login time periods can also be flexibly set according to industry characteristics. For example, it can be extended to 21:00 for nighttime operation scenarios.

[0034] In some embodiments, reference Figure 2 As shown, the associated risk identification module 3 further includes: The internal and external association identification unit 301 is used to obtain employee information in the human resources management system and enterprise personnel information of the supplier, identify whether the employee information and enterprise personnel information are consistent, and if so, determine that there is an internal and external association. The external association identification unit 302 is used to obtain newly added supplier information in the supplier management system and compare whether the newly added supplier information is consistent with the existing supplier information in the supplier management system. If so, it is determined that there is an external association. The bias identification unit 303 is used to identify whether, within a preset period, the supplier evaluator and the supplier being evaluated appear simultaneously a preset number of times and the frequency of the highest evaluation score exceeds a preset frequency threshold, and if so, to determine whether there is a bias between the supplier evaluator and the supplier.

[0035] This implementation method constructs a refined and traceable supplier association identification mechanism through multi-dimensional analysis of internal and external personnel information and evaluation behavior. Internal and external association identification enhances the monitoring capability of internal personnel using affiliated companies to obtain orders; external association identification can effectively detect behaviors such as suppliers falsifying independent identities and circumventing bidding restrictions; and the bias identification mechanism reveals potential subjective manipulation risks in the supplier evaluation process. The integration of these three identification mechanisms helps prevent commercial bribery, false independent bidding, and unfair evaluations, comprehensively improving the transparency and compliance level of supply chain management.

[0036] In some embodiments, the bias identification unit 303, after associating products in the procurement platform with their respective suppliers and supplier evaluators, performs grouped queries on suppliers and supplier evaluators that appear simultaneously a preset number of times. Specifically, a three-element relationship structure of product-supplier-supplier evaluator is formed. This structure is then grouped and statistically analyzed using SQL or a distributed query language to filter out evaluators who co-occur with a supplier more than a preset number of times within a preset period. Combined with their rating distribution, it is determined whether there is biased behavior where the scores are concentrated at the highest. Optionally, the preset period is 365 days, and the preset number of times can be flexibly configured according to actual needs, such as 5 times, or any other number of times. For example, but not limited to, the preset frequency threshold can be 80%. If the names of suppliers and supplier evaluators appear simultaneously in the same product evaluation 5 times within 365 days, and the scores in 4 of those evaluations are the highest, then it is considered that the supplier evaluator may have a bias towards the supplier, triggering an early warning. This embodiment improves the accuracy of identifying scoring manipulation behavior through product cascading and grouped co-occurrence statistics.

[0037] In some embodiments, reference Figure 3 As shown, the subject recognition module 2 is further configured as follows: Obtain the supplier's registered capital and number of employees covered by social security; The registered capital amount is compared with the purchase contract amount corresponding to the supplier. If the registered capital amount is lower than the product of the purchase contract amount and a preset ratio, the supplier is deemed to have insufficient performance capability. The purchase contract amount is pre-stored in the contract management system and can be obtained by indexing information such as the contract party, contract number, and contract name, or by extracting it based on the contract text combined with OCR recognition technology. Optionally, the preset ratio is configured as 20%, meaning that the supplier's registered capital is less than 20% of the total amount of the purchase contract project it has signed. By analyzing the historical performance records of suppliers, it is found that suppliers with registered capital less than 20% of the contract amount have a significantly higher default rate than suppliers with sufficient funds. Based on the preset ratio of this application, it can ensure that suppliers have sufficient financial buffer to cope with contract risks, effectively screen out suppliers with matching financial strength, and reduce the risk of default.

[0038] If the number of employees covered by social security is lower than a set threshold, then the second performance capability is deemed insufficient. The set threshold can be configured to 3 or any integer greater than 0. This is because suppliers with too few employees covered by social security (e.g., less than 3 people) often lack a stable team and have poor performance capability (e.g., delivery delays, quality issues). If the first or second performance capability is insufficient, the supplier is deemed not to meet the performance conditions.

[0039] In the above embodiments, the preset proportion of long-term contracts can be higher than the preset proportion of one-time purchase contracts, so as to ensure the financial stability of the supplier and evaluate whether its financial strength is sufficient to support the performance of the contract. In another embodiment, large, well-known suppliers may enjoy a lower preset ratio, while emerging or smaller suppliers may require a higher ratio.

[0040] In another embodiment, a preset ratio can be adjusted to reflect different risk tolerance levels based on the risk level of the contract (such as delivery time requirements).

[0041] In the above embodiments, the threshold for the number of people covered by social security can be adaptively modified according to the needs of the cooperative project, or it can be dynamically updated.

[0042] In another embodiment, a threshold can be set based on the total number and size of the supplier's employees; larger suppliers may require a higher number of social security employees to meet the needs of large-scale contracts.

[0043] In another embodiment, this application embodiment can select suppliers with strong performance capabilities in the industry as benchmarks based on the comparison of internal and external data, analyze the correspondence between their social security personnel and contract size, and configure reasonable threshold settings.

[0044] This implementation method comprehensively assesses suppliers' performance capabilities based on dual standards, enhancing the scientific rigor and practicality of risk identification. By introducing a ratio verification between registered capital and contract amount, it objectively reflects whether the supplier's financial strength supports the corresponding contract performance requirements. The number of employees covered by social security serves as a direct indicator of a company's human resource allocation, reflecting whether it possesses the basic personnel capacity for contract performance. Combining these two methods avoids misjudgments based on a single indicator, improving identification accuracy. Both the ratio and threshold settings support custom configuration, giving the system strong adaptability and dynamic adjustment capabilities, making it widely applicable to supplier audit and early warning mechanisms across different industries, project sizes, and risk management strategies.

[0045] In the above embodiments, the registered capital can be further refined into subscribed and paid-in amounts, and the timeliness of the capital contribution can be verified in conjunction with the time of contribution.

[0046] In some embodiments, reference Figure 4 As shown, the system of this application establishes data interfaces with the enterprise information query platform and the contract management system to obtain the enterprise information of the supplier. The entity identification module sends the supplier name to the enterprise information query platform to retrieve its registered capital amount, number of employees with social security, and enterprise personnel information. This system calls the open interface of the enterprise information query platform. For example, but not limited to, the enterprise information query platform can be Qichacha or Aiqicha. The contract management system interface can adopt GraphQL or direct database connection mode, and supports on-demand field extraction and automatic synchronization.

[0047] Similarly, refer to Figure 4 As shown, the system in this application also establishes data interfaces with the human resources management system and the supplier management system.

[0048] The aforementioned contract management system, human resource management system, and supplier management system can be subsystems under the enterprise's material procurement platform or external independent systems, and data transmission can be achieved through open interfaces.

[0049] In some embodiments, the employee information is a set of employee names and the names of their close relatives, and the enterprise personnel information includes a set of legal representative names, shareholder names, and senior executive names.

[0050] In some embodiments, if the number of names that match any employee's name and the names of their close relatives with the names in the company's personnel information exceeds a preset number, then the employee information and the company's personnel information are identified as having consistency. Optionally, the preset number is 2.

[0051] The above embodiments broaden the identification scope by cross-referencing multiple fields of employee and enterprise personnel information, which can uncover hidden interpersonal relationships and is especially suitable for preventing kinship ties.

[0052] In the above embodiments, the number of names that match can also be counted by fuzzy matching or hash comparison.

[0053] In some embodiments, name matching rules can be used to match the names of employees and their close relatives with the names in the company's personnel information. These name matching rules can incorporate name entity recognition models from natural language processing to handle complex situations such as pinyin, aliases, and people with the same name but different identities. Specifically, this includes: The system performs entity standardization on the names of employees and their close relatives, as well as the name field in the company's personnel information. Each name is vectorized using a name word vector model (such as BERT-NER or FastText), and the cosine similarity between word vectors is calculated to determine the semantic similarity. The semantic similarity of names is scored, and similar names with similar pinyin are identified by combining speech spelling conversion rules, such as "Zhang Wei" and "Zhang Wei". Common aliases are matched using an alias dictionary, such as "Zhang San" and "Lao Zhang". Contextual information from a knowledge graph is used to help determine whether they belong to the same actual person. Ultimately, the system achieves cross-field and cross-source data consistency identification, helping to confirm whether they belong to the same natural person or family member.

[0054] When faced with scenarios involving misspelled names, transliterated pinyin, common names, or people with the same name but different identities, the system can more intelligently infer true relationships, thereby reducing the risk of misjudgment and omission. Especially in real-world, large-scale supplier data, it exhibits higher fault tolerance and robustness in identification, contributing to improved comprehensiveness and depth in supplier auditing and oversight.

[0055] The human resources management system pre-stores the names of close relatives (including spouses, children, parents, etc.). The employee names are the names of current employees, and may also include former employees and retired employees. Former employees and retired employees can select a certain range of personnel data, such as data within the last three years, so as to discover whether current employees, retired employees, former employees and their close relatives are associated with suppliers through the internal and external association identification unit.

[0056] In some embodiments, the new supplier information includes the bank account and contact information provided by the new supplier during the registration phase; If the bank account information or contact information in the existing supplier information is consistent with that in the newly added supplier information, it is determined that there is an external connection.

[0057] Based on this, the external association identification unit determines whether there is a connection between the bank account information, contact information and other offline information of the actual suppliers, so as to prevent the same actual controller from using multiple accounts to participate in bidding or manipulate market behavior, further reduce the space for gray operation, and ensure the openness, fairness and traceability of the bidding and procurement behavior.

[0058] In another embodiment, the newly added supplier information may also include a set of shareholder names in the equity relationship. The set of shareholder names can be obtained from the enterprise information query platform. The parameter "new supplier name" is passed to the enterprise information query platform to retrieve its equity relationship. The equity relationship of the new supplier and the equity relationship of the existing supplier are analyzed using a graph neural network to reveal the hidden supplier association chain, thereby further improving the identification accuracy.

[0059] Specifically, when a new supplier submits its name information during the registration or data update process, the system automatically calls the open interface of the enterprise information query platform to retrieve its current enterprise structure, including the legal representative, controlling shareholder, actual controller, etc., using the new supplier's name as the search parameter, and extracts its set of shareholder names. The system simultaneously aggregates the shareholder names of all existing suppliers in the supplier management system, constructing a relationship graph among suppliers. A graph neural network model (such as GCN, GAT, or GraphSAGE) is used to train and embed the graph's representation, enabling the system to identify implicit supplier association chains hidden in multi-hop relationships (such as common parent shareholders or cross-shareholding structures). If the GNN analysis results show a high-confidence equity path connection between a new supplier and an existing supplier (such as shared shareholders within two hops or shared control nodes), an external association is identified, triggering an alert.

[0060] In another embodiment, the set of shareholder names can also be expanded to fields such as shareholder ID, shareholding ratio, and investment time for weighted graph construction; the graph neural network can select a lightweight architecture based on the graph complexity, such as FastGCN based on adjacency matrix sparsification, to reduce computational complexity.

[0061] This implementation method addresses the issue of overlooking shareholder control relationships when relying solely on contact person or bank account comparisons by introducing the shareholder structure of newly added suppliers as a key data field into the external correlation identification process. This significantly enhances the intelligence and depth of supplier equity correlation identification, making it particularly suitable for identifying hidden control relationships that circumvent explicit field comparisons. By introducing graph neural networks, it overcomes the limitations of traditional planar field matching, identifying cross-level and cross-entity shareholder control paths within complex corporate equity structures. It is especially effective in identifying risky behaviors such as "shell companies" and "shell suppliers" that conceal the actual controller through structural design. This mechanism strengthens the prevention of commercial fraud, internal bid-rigging, and supplier collusion, enhancing the intelligent risk discovery capabilities of the regulatory system.

[0062] In some embodiments, the supervision unit includes: a first supervision unit, a second supervision unit, and a third supervision unit. The first supervision unit is used to receive the first warning information and trigger a business supervision process; the second supervision unit is used to receive the second warning information and trigger a compliance process; and the third supervision unit is used to receive the third warning information and trigger an audit process.

[0063] The first warning message may include the following: As of the [Trigger Time], the supplier [Supplier Name] logged into the account [System Name] at the address [Address Name] and submitted a quote. However, its frequently used address is [Frequently Used Address Name]. This has triggered a warning. Please confirm if there is a problem. The "Trigger Time" is determined based on the daily trigger time of the IP anomaly identification module 1. The "Supplier Name" is obtained based on the supplier account associations pre-stored in the supplier management system. The "Address Name" can be the province of the identified IP login address, and the "Frequently Used Address Name" is the name of the aforementioned frequently used address.

[0064] The second warning message may include the following: As of the [trigger time], the [Contract A] has signed a contract with the number [Contract Number] and the contract amount is [Contract Amount]. However, the registered capital of its contracting party, the [Supplier Name], is only [Registered Capital]. This has triggered a warning. Please confirm whether there is a problem.

[0065] The third warning message may include: As of the [trigger time], the names of the relatives of the employee [employee name] in the [employee's unit] [relative name 1, relative name 2] are consistent with the names of the personnel related to the supplier [supplier name] [personnel name 1, personnel name 2], which has triggered a warning. Please confirm whether there is a problem. The third warning message may also include: As of the [trigger time], the bank account information of suppliers [Supplier Name 1 and Supplier Name 2] in the [Supplier Information System] is consistent, and a warning has been triggered. Please confirm whether there is a problem. Here, "Supplier Information System" refers to the supplier management system. As of the [trigger time], the contact information for suppliers [Supplier Name 1 and Supplier Name 2] in the [Supplier Information System] is identical, with both names and phone numbers being [Contact Person Name, Contact Person Phone Number]. This has triggered an alert. Please confirm if there is a problem. The third warning information may also include: As of the [trigger time], the [personnel name] of the [personnel's unit] and the [supplier name] of the supplier have appeared in the same evaluation at the same time 5 times within 365 days, of which 4 times the person gave the highest score. The warning has been triggered. Please check if there is a problem.

[0066] In the above examples, the fields inside the “【】” symbols can all be configured according to specific content. The examples in this embodiment are not intended to limit the content of their internal fields.

[0067] In the above embodiments, when the supervision unit receives the first, second, or third early warning information pushed by the system, it automatically triggers the response processing flow. Based on the fields recorded in the early warning information, such as the early warning type, supplier name, triggering module, and affiliated unit, it automatically matches the corresponding responsible department (e.g., the first early warning information corresponds to the business department, the second early warning information corresponds to the legal department, and the third early warning information corresponds to the discipline inspection department) and generates an early warning processing work order. It supports notifying relevant personnel through in-system notifications, emails, or mobile reminders.

[0068] The business supervision process, compliance process, and audit process include the early warning triggering step, early warning issuance step, early warning response step, early warning review step, and early warning cancellation step, respectively.

[0069] During the warning triggering process, responsible personnel from the business department, legal department, and discipline inspection department log in to the system and view the warning details through the warning ledger interface. If the current warning information needs to be forwarded to other responsible personnel, the warning issuance step can be carried out to forward the current warning information to the corresponding responsible personnel for processing. In the warning response step, the responsible personnel respond to the current warning information. Specifically, a text box and attachment upload function are provided in the next level interface of the warning work order interface, allowing the responsible personnel to fill in an explanation of the warning situation and support attachment upload; then, it is submitted to the warning approver.

[0070] During the early warning review process, the early warning approver clicks the review button on the early warning ledger interface to review the text content and attachments in the early warning response step. They can also select a processing opinion based on the actual situation, such as "approved", "further investigation required", or "approved", and upload verification materials or remarks.

[0071] If the processing opinion in the warning review step is "approved", then the warning cancellation step will be executed, and a cancellation message reminder will be sent to the corresponding responsible personnel.

[0072] If the feedback in the early warning review step is "further investigation required" or "review not approved", you can jump back to the early warning review step.

[0073] All operations record information such as the person handling the operation, the time of operation, and the content of the comments. The system automatically updates the review status.

[0074] The aforementioned compliance and audit processes also include an early warning review step, which is used by the responsible personnel at the higher level to review the content of the current early warning review step.

[0075] After verification, the monitoring unit supports marking the warning as "processed" or "resolved". When necessary, the system can also perform secondary verification logic (such as confirming that the information source has been updated or that the IP address is no longer abnormal), and allows the uploading of relevant attachments (such as supplementary contract materials and employee investigation conclusions) to form a complete processing record for subsequent audit tracking.

[0076] This implementation method establishes a standardized early warning processing flow, achieving accountability and process traceability from system alerts to manual intervention. Automatic dissemination of early warning information avoids omissions and delays caused by manual assignment; the review mechanism clarifies the responsibilities of each party, improving problem response efficiency; and the delisting mechanism ensures that all early warnings are addressed conclusively, forming a closed data loop within the system. Simultaneously, all processing records can be accumulated as supervisory data assets, providing a basis for future risk model training, early warning optimization, and regulatory compliance assessments. This processing mechanism enhances the operability, transparency, and audit friendliness of the supervisory system.

[0077] It should be noted that the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0078] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0079] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A supplier anomaly identification and monitoring system, characterized in that, include: The IP anomaly identification module is used to identify whether the supplier account has an abnormal IP login address and to perform a quotation operation when the IP login address is abnormal. If so, the first risk warning is triggered and the first warning information is sent to the supervision unit. The entity identification module is used to identify whether the supplier in the procurement contract meets the performance conditions. If not, it triggers a second risk warning and sends a second warning message to the supervision unit. The associated risk identification module is used to identify whether the supplier is associated with the employee or other suppliers, and to identify whether there is any bias between the supplier evaluator and the supplier being evaluated. If the identification result of either is yes, a third risk warning is triggered and a third warning message is sent to the supervision unit.

2. The supplier anomaly identification and monitoring system according to claim 1, characterized in that, The associated risk identification module further includes: The internal and external correlation identification unit is used to obtain employee information from the human resources management system and enterprise personnel information of the supplier, identify whether the employee information and enterprise personnel information are consistent, and if so, determine that there is an internal and external correlation. The external association identification unit is used to obtain newly added supplier information in the supplier management system and compare whether the newly added supplier information is consistent with the existing supplier information in the supplier management system. If so, it is determined that there is an external association. The bias identification unit is used to identify whether, within a preset period, the supplier evaluator and the supplier they evaluate appear simultaneously a preset number of times and the frequency of the highest evaluation score exceeds a preset frequency threshold. If so, it is determined that there is a bias between the supplier evaluator and the supplier.

3. The supplier anomaly identification and monitoring system according to claim 2, characterized in that, In the aforementioned bias identification unit, after associating the products in the procurement platform with their respective suppliers and supplier evaluators, a grouped query is performed to simultaneously show suppliers and supplier evaluators a preset number of times.

4. The supplier anomaly identification and monitoring system according to claim 3, characterized in that, The subject identification module is further configured as follows: Obtain the supplier's registered capital and number of employees covered by social security; Compare the amount of the registered capital with the amount of the purchase contract corresponding to the supplier. If the amount of the registered capital is lower than the product of the purchase contract amount and a preset ratio, then the first performance capability is determined to be insufficient. If the number of people covered by social security is lower than the set threshold, then the second performance capacity is deemed insufficient. If the first or second performance capability is insufficient, the supplier is deemed not to meet the performance conditions.

5. The supplier anomaly identification and monitoring system according to claim 4, characterized in that, The system establishes data interfaces with the enterprise information query platform and the contract management system to obtain the enterprise information of the suppliers. The entity identification module sends the supplier name to the enterprise information query platform to retrieve its registered capital amount, number of employees covered by social security and enterprise personnel information.

6. The supplier anomaly identification and monitoring system according to claim 2, characterized in that, The employee information is a collection of employee names and the names of their close relatives, and the enterprise personnel information is a collection of the names of the legal representative, shareholders, and senior executives.

7. The supplier anomaly identification and monitoring system according to claim 6, characterized in that, If the number of names matching any employee's name and the names of their close relatives in the company's personnel information exceeds a preset number, then the employee information and the company's personnel information are considered to be consistent.

8. The supplier anomaly identification and monitoring system according to claim 2, characterized in that, The newly added supplier information includes the bank account and contact information provided by the new supplier during the registration phase; If the bank account information or contact information in the existing supplier information is consistent with that in the newly added supplier information, it is determined that there is an external connection.

9. The supplier anomaly identification and monitoring system according to any one of claims 1 to 8, characterized in that, The monitoring unit includes: The first supervision unit is used to receive the first early warning information and trigger the business supervision processing flow. The second supervisory unit is used to receive the second early warning information and trigger the compliance processing procedure. The third supervisory unit is used to receive the third early warning information and trigger the audit processing flow.

10. The supplier anomaly identification and monitoring system according to any one of claims 1 to 8, characterized in that, The IP anomaly identification module, the subject identification module, and the associated risk identification module are each configured with an execution frequency.

Images