Supplier life cycle management method based on dynamic portrait and AI enabling

Abstract

The invention provides a supplier life cycle management method based on a dynamic portrait and AI enabling, and the method comprises the steps: taking a Kafka message queue as a unified data bus, accessing supplier multi-source data through an API interface, log collection and other modes, employing Flink to clean and convert high-speed data streams in real time, and carrying out the batch processing of large-scale historical data through an MRS; based on GuasDWS, constructing an evaluation model containing dimensions of quality control, delivery ability and the like, calculating supplier ability scores and generating ability short board labels; analyzing demands for the purchasers, screening and sorting to recommend suppliers, and pushing customized training courses matched with a short board for the suppliers; decision feedback and training effect data are collected through portal interaction, and a driving model and algorithm increment optimization are transmitted back. The problems of static lag and extensive matching of traditional management can be solved, and the supply chain management efficiency and toughness are improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a supplier lifecycle management method based on dynamic profiling and AI empowerment. Background Technology

[0002] In modern supply chain systems, suppliers, as core participants, directly impact a company's operational efficiency, cost control, and market competitiveness through their capabilities, contract fulfillment stability, and risk profile. This is particularly true in sectors such as retail e-commerce, manufacturing, and financial services. As businesses expand and supply chains become more complex, the need for refined and dynamic supplier management is increasingly urgent. Traditional supplier management relies heavily on information architecture for basic functions, such as storing supplier business information, qualification documents, and historical transaction data in relational databases. Regular statistical reports are used to assess indicators like contract fulfillment rates and quality pass rates, aiming to standardize supplier cooperation processes. However, with the widespread adoption of big data technology and the evolving demands for supply chain collaboration, a static information management model is no longer sufficient to meet companies' needs for real-time supplier capability awareness, risk warnings, and optimal resource allocation.

[0003] While current mainstream supplier management solutions are gradually incorporating big data tools (such as offline data analysis using Hadoop) or basic machine learning models (such as supplier grouping through cluster analysis), significant technical shortcomings remain: First, capability assessment dimensions are rigid and outdated, relying heavily on static scorecards from quarterly or annual periods. This makes it difficult to integrate multi-source heterogeneous data such as order dynamics, logistics timeliness, and public opinion risks in real time, resulting in the inability to incorporate key information such as sudden fluctuations in supplier production capacity and newly acquired certifications into the assessment in a timely manner, which can easily lead to cooperation risks. Second, training and empowerment mechanisms are passive and inefficient, often relying on a unified document library or online platform to push general learning materials without customizing content based on the specific capability shortcomings of suppliers (such as delivery delays and unstable quality). Furthermore, there is a lack of correlation verification between training effectiveness and business data, making it difficult to achieve targeted improvement of supplier capabilities. Third, resource matching logic is crude, often selecting suppliers based on simple rules such as region and category, without comprehensively quantifying multi-dimensional capabilities such as supplier quality stability, cost flexibility, and innovation potential. This not only easily leads to the burying of high-quality supplier resources but may also cause insufficient supply chain resilience due to over-reliance on leading suppliers, restricting the overall supply chain's collaborative efficiency and sustainable development. Summary of the Invention

[0004] In view of this, the embodiments of this application provide a supplier lifecycle management method based on dynamic profiling and AI empowerment, which can solve the problems of static lag and extensive matching in traditional management, and improve the efficiency and resilience of supply chain management.

[0005] The technical solution of this application embodiment is implemented as follows: In a first aspect, embodiments of this application provide a supplier lifecycle management method based on dynamic profiling and AI empowerment, comprising the following steps: Using Kafka message queue as a unified data bus, we access basic static data, business dynamic data, external public opinion data and supply chain relationship data from suppliers through API interfaces, log collection, IoT devices and web crawlers. We use the Flink real-time computing engine to clean and transform high-speed data streams, and at the same time, we use the MRS big data platform to batch import and process large-scale historical data. Based on GuassDWS, a multi-dimensional capability assessment model is constructed. The model includes core dimensions such as quality control, delivery capability, cost control, technical strength, and response speed. The supplier capability score is calculated through quantifiable indicators of each dimension, and the score is compared with industry benchmarks or preset thresholds to automatically identify capability shortcomings and generate corresponding capability shortcoming labels. For buyers, we analyze their procurement needs and filter, rank, and recommend suppliers; for suppliers, we push customized training courses based on their capability gap tags. Human-computer interaction is achieved through the buyer portal and supplier portal, and procurement decision feedback data and training effect data are collected. The data is then sent back to the data processing link to drive incremental training and optimization of the capability assessment model and recommendation algorithm.

[0006] Secondly, embodiments of this application also provide a supplier lifecycle management device based on dynamic profiling and AI empowerment, the device comprising: The data processing module uses Kafka message queue as a unified data bus to access basic static data, business dynamic data, external public opinion data and supply chain relationship data from suppliers through API interfaces, log collection, IoT devices and web crawlers. It uses Flink real-time computing engine to clean and transform high-speed data streams, and at the same time uses the MRS big data platform to batch import and process large-scale historical data. The tag generation module is used to build a multi-dimensional capability assessment model based on GuassDWS. The model includes core dimensions such as quality control, delivery capability, cost control, technical strength, and response speed. The module calculates the supplier's capability score through quantifiable indicators of each dimension and compares the score with industry benchmarks or preset thresholds to automatically identify capability shortcomings and generate corresponding capability shortcoming tags. The filtering and recommendation module is used to analyze the procurement needs of buyers and filter, sort and recommend suppliers; and to push customized training courses to suppliers based on their capability gap tags. The feedback collection module is used to collect procurement decision feedback data and training effect data through human-computer interaction via the buyer portal and supplier portal, and to send the data back to the data processing link to drive incremental training and optimization of the capability assessment model and recommendation algorithm.

[0007] Thirdly, embodiments of this application also provide an electronic device, including: a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the supplier lifecycle management method based on dynamic profiling and AI empowerment as described in any of the first aspects.

[0008] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, performs the supplier lifecycle management method based on dynamic profiling and AI empowerment as described in any of the first aspects.

[0009] The embodiments of this application have the following beneficial effects: By real-time acquisition and integrated batch processing of multi-source heterogeneous data, data silos in traditional supplier management are broken down, and the problem of lagging static assessments is solved. Based on the multi-dimensional dynamic capability profiles built by GuassDWS, precise quantification and intelligent identification of weaknesses in core dimensions such as supplier quality control, delivery capabilities, and cost control are achieved, effectively improving the timeliness of risk warnings and the comprehensiveness of assessments. Leveraging a two-way AI empowerment mechanism, on the one hand, through NLP demand analysis and multi-objective optimization ranking algorithms, optimal supplier recommendations are provided to buyers, taking into account business matching, historical performance, risk level, and supply chain resilience, avoiding order allocation imbalances caused by crude resource matching. On the other hand, this provides support for high-quality small and medium-sized enterprises. On the one hand, it creates equal exposure opportunities for suppliers; on the other hand, it pushes customized training based on capability gap tags and forms an effect tracking closed loop, promoting targeted improvement of supplier capabilities and solving the problems of passive inefficiency and disconnected content in traditional training. In addition, the full-link closed-loop design of procurement decision feedback and training effect data to feed back into the model optimization enables the system to have the ability to continuously iterate and self-optimize. Ultimately, it transforms supplier management from a passive, experience-driven model to a proactive, AI-driven collaborative model, significantly improving the efficiency of supply chain resource allocation, reducing cooperation risks and operating costs, and can be widely applied to multiple industry scenarios such as retail e-commerce, manufacturing, and financial services, with strong practical application value and promotion significance. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart illustrating steps S101-S104 provided in the embodiments of this application; Figure 2 This is a flowchart illustrating steps S201-S204 provided in the embodiments of this application; Figure 3 This is a flowchart illustrating steps S301-S303 provided in the embodiments of this application; Figure 4 This is a schematic diagram of the supplier lifecycle management device based on dynamic profiling and AI empowerment provided in the embodiments of this application; Figure 5 This is a schematic diagram of the composition structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0013] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0014] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0015] In the following description, the terms "first, second, third" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first, second, third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0016] It should be noted that the term "comprising" will be used in the embodiments of this application to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application and is not intended to limit this application.

[0018] See Figure 1 , Figure 1 This is a flowchart illustrating steps S101-S10* of the supplier lifecycle management method based on dynamic profiling and AI empowerment provided in this application embodiment, which will be combined with... Figure 1 Steps S101-S104 are explained below.

[0019] In step S101, Kafka message queue is used as a unified data bus. Through API interface, log collection, IoT devices and web crawler, basic static data, business dynamic data, external public opinion data and supply chain relationship data of suppliers are accessed. Flink real-time computing engine is used to clean and transform high-speed data streams. At the same time, large-scale historical data is imported and processed in batches through MRS big data platform. Here, Kafka message queue is used as a unified data bus because it has the characteristics of high throughput and low latency, and can simultaneously access multiple types of data (system data from API interfaces, business data from log collection, real-time status data from IoT devices, and external data from web crawlers), avoiding link chaos when multiple data sources are accessed.

[0020] Among them, the Flink real-time computing engine is used to process high-speed business dynamic data (such as order fulfillment status and logistics trajectory), and can complete data cleaning (removing outliers, such as pass rate data that exceeds the reasonable range) and transformation (unifying data format, such as standardizing the delivery time field of different systems to YYYY-MM-DD format) in seconds; the MRS big data platform is used for batch processing of large-scale historical data (such as transaction records and historical public opinion of the past year). The combination of the two realizes the integration of stream and batch data, providing a comprehensive and accurate data source for subsequent profile construction.

[0021] In step S102, a multi-dimensional capability assessment model is constructed based on GuassDWS. The model includes core dimensions such as quality control, delivery capability, cost control, technical strength, and response speed. The supplier capability score is calculated through quantifiable indicators of each dimension, and the score is compared with industry benchmarks or preset thresholds to automatically identify capability shortcomings and generate corresponding capability shortcoming labels. Here, we choose GuassDWS to build a multi-dimensional capability assessment model because it supports petabyte-level data storage and complex analysis, and can efficiently calculate multi-dimensional indicators. Among them, the quality control dimension can be quantified by product sampling pass rate and after-sales complaint rate; the delivery capability dimension can be quantified by on-time delivery rate and order fulfillment rate; the cost control dimension can be quantified by the deviation of the quotation from the industry average price and the cost reduction rate; the technical strength dimension can be quantified by the number of patents and the application rate of new technologies; and the response speed dimension can be quantified by the inquiry response time and problem resolution cycle.

[0022] By comparing the scores of each dimension with industry benchmarks (such as supply chain industry report data released by third-party institutions) or enterprise preset thresholds (such as enterprises requiring on-time delivery rate ≥95%), a capability deficiency label is automatically generated (such as generating a delivery delay risk label if on-time delivery rate is 88% < 95%). This label is updated to the profile in real time, providing a precise basis for identifying shortcomings for subsequent AI empowerment.

[0023] In step S103, for the purchaser, the purchase requirements are analyzed and suppliers are screened, sorted and recommended; for the suppliers, customized training courses are pushed based on capability gap tags. Here, the supplier recommendations for buyers address the issues of low efficiency and suboptimal product selection caused by manual screening. By analyzing needs, screening candidates, and optimizing the ranking, the recommendations are ensured to match the procurement needs and be the best overall. For customized training pushes for suppliers, the issues of training content being out of touch with their weaknesses and low participation are addressed. Based on the capability gap tags generated in the previous steps, the appropriate courses are pushed, making training go from a one-size-fits-all approach to on-demand matching.

[0024] In step S104, human-computer interaction is achieved through the buyer portal and supplier portal to collect procurement decision feedback data and training effect data. The data is then sent back to the data processing link to drive incremental training and optimization of the capability assessment model and recommendation algorithm.

[0025] Finally, human-computer interaction is achieved through the buyer portal and supplier portal. Buyers can make decisions based on the recommendation list, and suppliers can view profiles and learning courses. At the same time, feedback on purchasing decisions (such as not selecting the top supplier or marking inaccurate recommendations) and training effectiveness data (such as changes in pass rates after learning) are sent back to the data processing link to drive the capability assessment model to adjust the indicator weights (such as increasing the weight of public opinion data in risk level calculation if risk level assessments are repeatedly marked as inaccurate) and incrementally train the recommendation algorithm (such as optimizing the ranking logic based on purchasing decision preferences), ultimately achieving a virtuous cycle of becoming smarter with use.

[0026] In some embodiments, the basic static data includes supplier business information, qualification certifications, and registered capital, which are derived from ERP or SCRM systems; The business dynamic data includes order fulfillment rate, on-time delivery rate, product quality pass rate, transaction frequency and scale, which are collected in real time through a log collector; The external public opinion data is obtained by crawling industry association announcements, penalty notices, news public opinion and social media evaluations, and processed using NLP technology to perceive external risks to suppliers. The supply chain relationship data is extracted from the enterprise database and includes supplier equity structure, investment relationships, and information on senior management connections.

[0027] The detailed descriptions of each data type are as follows: Basic static data: Data content: Supplier business information (such as registered capital, business scope, and years of establishment), qualification certifications (such as ISO9001 quality certification and industry access qualifications). This type of data is a prerequisite for judging the basic compliance of suppliers (for example, suppliers with registered capital of less than 5 million may not meet the requirements for large-scale procurement). Data source: ERP (Enterprise Resource Planning) system or SCRM (Supplier Relationship Management) system. These two systems are the core systems for storing basic supplier information within the enterprise, ensuring the authority and accuracy of the data. Function: To provide basic attribute tags for dynamic profiles, such as suppliers with complete qualifications and large suppliers (registered capital ≥ 10 million), as basic screening conditions for subsequent resource matching.

[0028] Business dynamic data: Data content includes: order fulfillment rate (number of fulfilled orders / total number of orders), on-time delivery rate (number of on-time delivered orders / total number of orders), product quality pass rate (number of qualified products / number of sampled products), and transaction frequency and scale (such as the number of transactions per month and the amount of a single transaction). This type of data directly reflects the supplier's current cooperation capabilities. Data source: Real-time inflow through log collectors, such as transaction logs from the order system and inspection logs from the quality inspection system, ensuring that data is generated synchronously with business activities and avoiding delays and errors from manual entry; Function: It serves as the core basis for calculating the dynamic profile capability dimension. For example, a delivery on-time rate of 92% is directly used for the score calculation of the delivery capability dimension.

[0029] External public opinion data: Data content includes: industry association announcements (such as industry violation notices), penalty notices, news and public opinion (such as negative news about suppliers), and social media reviews (such as customer complaints about supplier products). This type of data is used to perceive external risks to suppliers. Processing method: After crawling data through web crawlers, semantic analysis is performed using NLP technology. For example, texts about supplier A being penalized for product quality issues are automatically categorized as quality risks; information about customer complaints caused by supplier B's delayed delivery is extracted and associated with delivery risks. Purpose: To supplement the risk dimensions of the profile and avoid risk omissions caused by relying solely on internal data (e.g., if a supplier has good internal performance data but has recently been subject to significant quality penalties, this needs to be reflected in the risk level).

[0030] Supply chain relationship data: Data content: Supplier equity structure (e.g., whether there are related companies), investment relationships (e.g., upstream and downstream companies invested in), and senior management relationships (e.g., whether senior management holds positions in other risky companies) extracted from enterprise databases (e.g., industrial and commercial enterprise information database, enterprise credit information disclosure system); Function: Used to build supply chain knowledge graphs, identify hidden risk associations (such as the controlling shareholder of supplier A also controlling a company with a bad credit record, which requires vigilance against joint risks), and provide supply chain synergy references for resource matching (such as giving priority to suppliers that have no associated risks with the company's existing partners).

[0031] In some embodiments, the quantitative calculation of the delivery capability dimension is based on the supplier's on-time delivery rate and order fulfillment rate. When the score of either dimension is consistently lower than a preset threshold, the system automatically generates a corresponding capability deficiency label and updates it to the supplier's dynamic profile in real time.

[0032] Here, on-time delivery rate and order fulfillment rate are chosen as core quantitative indicators because they comprehensively reflect delivery capabilities from two dimensions: time compliance and quantity compliance. On-time delivery rate focuses on whether the goods are delivered on time (e.g., if the delivery date is agreed to be the 10th but it is delivered on the 12th, it is considered as not being on time). Order fulfillment rate focuses on whether the goods are delivered in the agreed quantity (e.g., if the delivery date is agreed to be 1000 pieces but only 800 pieces are delivered, the fulfillment rate is 80%). Combining the two can avoid the one-sidedness of a single indicator (e.g., if a supplier has an on-time delivery rate of 100% but an order fulfillment rate of only 70%, it is still considered as insufficient delivery capability).

[0033] The specific calculation method can be exemplified as follows: Delivery capability score = (On-time delivery rate × 0.6 + Order fulfillment rate × 0.4) × 100. The weight can be adjusted according to the company's procurement needs (for example, in procurement scenarios that are sensitive to delivery time, the weight of the on-time delivery rate can be increased to 0.7).

[0034] Continuous judgment criteria can be set in conjunction with business cycles, such as three consecutive purchase order cycles or monthly average scores, to avoid misjudgments caused by a single abnormal data (e.g., a supplier's delay due to a sudden logistics disaster is not directly judged as a weakness); the preset threshold is set based on two categories: one is the industry benchmark (e.g., if the industry average delivery capability score is 85 points, then the threshold can be set to 85 points), and the other is the enterprise's customized requirements (e.g., if the enterprise's core procurement business requires a delivery capability score ≥ 90 points, then the threshold is set to 90 points).

[0035] When the delivery capability score remains below the threshold, the system automatically generates corresponding tags (such as insufficient on-time delivery rate, low order fulfillment rate, or comprehensive tag indicating delivery capability needs improvement) and updates them to the supplier dynamic profile in real time. The logic for real-time updates is as follows: after each delivery transaction is completed, the system automatically recalculates the delivery capability score. If the score drops from above the threshold to below, or remains below the threshold, tag generation / update is triggered immediately. If the score subsequently rises back above the threshold, the system automatically deletes the weakness tag, ensuring that the profile changes synchronously with the supplier's actual capabilities.

[0036] The generated delivery capability gap labels will directly serve as trigger conditions for the AI-enabled center's intelligent growth empowerment module. For example, after the system identifies high-frequency delivery delay labels, it will automatically push customized training courses such as efficient supply chain scheduling plans and logistics collaborative management skills, providing precise directions for subsequent improvement of gaps.

[0037] In some embodiments, see Figure 2 , Figure 2 This is a flowchart illustrating steps S201-S204 provided in the embodiments of this application. The specific process of analyzing the purchaser's purchase needs and screening, sorting and recommending suppliers includes steps S201-S204, which will be explained in conjunction with each step.

[0038] In step S201, key information in the procurement requirement text is identified using NLP technology, and constraints such as product category, process, delivery date, and budget are automatically extracted. In step S202, based on the extracted constraints, suppliers that meet the qualification and category requirements are retrieved from the supplier pool of the Elasticsearch index to form a candidate supplier set; In step S203, a multi-objective optimization ranking algorithm is used to score and rank the suppliers in the candidate set to generate a recommendation list.

[0039] Procurement requirements are often expressed in natural language (e.g., urgently need a batch of high-precision PCBs, delivery within 15 days, budget of 500,000 yuan). The core purpose of using NLP technology is to automatically extract structured constraints, avoiding the errors and time consumption of manual extraction.

[0040] A specific implementation example is as follows: The NLP model uses keyword recognition and semantic analysis to extract constraints such as category = PCB board, process = high precision, delivery time ≤ 15 days, and budget ≤ 500,000 yuan from the requirement text. At the same time, it standardizes ambiguous expressions (such as urgent needs can be converted into high delivery time priority, and the weight of delivery time matching degree is increased in subsequent sorting). If there is missing information in the requirement text (such as the budget is not mentioned), the system can automatically prompt the purchasing personnel to supplement it, or default to the budget range of similar purchases in the enterprise.

[0041] Elasticsearch (ES) was chosen as the supplier pool retrieval tool because of its distributed real-time retrieval capabilities. It can quickly filter out candidates that meet hard constraints from a large-scale supplier database (such as 100,000 suppliers). Hard constraints are uncompromising conditions such as qualifications and product categories (e.g., when purchasing high-precision PCBs, suppliers with PCB production qualifications and the ability to provide high-precision processes must be selected first). Suppliers that do not meet the requirements (such as suppliers that only produce ordinary circuit boards) are excluded, reducing the computational load of subsequent AI ranking.

[0042] The initial screening logic is as follows: the supplier attributes (qualifications, categories, process range, etc.) in the ES index are matched with the constraints extracted from the demand parsing. Suppliers that fully meet the constraints are included in the candidate set. For example, 200 candidate suppliers that meet the requirements of PCB board + high-precision process + delivery time ≤ 15 days are selected from a pool of 100,000 suppliers.

[0043] The core of the multi-objective optimization ranking algorithm is to comprehensively weigh multiple dimensions of factors and output a globally optimal recommendation list, rather than ranking based on a single dimension (such as sorting only by price from low to high). During the ranking process, the system calculates a comprehensive matching score for each candidate supplier, with the supplier with the highest score ranked first. The specific score calculation can be carried out in combination with multiple dimensions, and finally outputs a recommendation list with scores and explanations of advantages and disadvantages to the purchasing personnel (e.g., Supplier A: comprehensive score 92 points, advantages: on-time delivery rate 98%, budget matching degree 100%, disadvantage: price is slightly higher than the industry average by 5%), providing intuitive basis for purchasing decisions.

[0044] In some embodiments, the scoring criteria of the multi-objective optimization ranking algorithm include: business matching degree, historical performance, risk level, value return and supply chain resilience, and a list of alternative suppliers is generated simultaneously during the ranking process to reduce the risk of dependence on a single supplier.

[0045] Here, the meaning and weighting of each scoring criterion are set as follows: Business matching degree: This includes product fit (the degree to which the supplier's products match the specifications and processes required by the procurement needs, such as whether the supplier has micron-level processing capabilities if a high-precision PCB board is required) and capacity matching degree (whether the supplier's current capacity can meet the procurement volume, such as a high matching degree if the demand is 100,000 pieces / month and the supplier's monthly capacity is 200,000 pieces). The weight can be set to 0.3 (the core basis to ensure that the supplier can meet the basic needs). Historical performance: Based on dynamic business data, including order fulfillment rate and quality pass rate, with a weight of 0.25 (reflecting the reliability of past cooperation with suppliers and reducing cooperation risks); Risk level: Based on external public opinion data and supply chain relationship data, including real-time public opinion risk (whether there is recent negative information) and financial health (such as whether there is a risk of tight cash flow), the weight is set to 0.2 (to control cooperation risk and avoid cooperating with high-risk suppliers). Value gain: This includes procurement costs (comparison of supplier quotes with industry averages and company budgets) and logistics costs (differences in logistics costs due to the distance between the supplier's location and the buyer's warehouse), with a weight of 0.15 (balancing cost and quality). Supply chain resilience: Assess the suitability of suppliers as alternative resources, with a weight of 0.1 (to ensure supply chain stability).

[0046] Overall score = Business matching degree × 0.3 + Historical performance × 0.25 + Risk level × 0.2 + Value gain × 0.15 + Supply chain resilience × 0.1. The weights can be dynamically adjusted according to the procurement scenario (e.g., in strategic procurement scenarios, the risk level weight can be increased to 0.3).

[0047] Generating a list of alternative suppliers is to avoid the risk of relying on a single supplier (such as a sudden production stoppage by the main recommended supplier, leading to procurement disruption). The specific logic is as follows: after the AI ​​comprehensive ranking, in addition to outputting the top 3-5 main recommended suppliers, 2-3 additional suppliers with slightly lower comprehensive scores than the main recommendation, but who fully match the key constraints, are selected as alternatives. For example, the main recommended supplier A has a comprehensive score of 92 points, while alternative suppliers B and C have scores of 88 and 87 points respectively. Both B and C meet the requirements for product category, delivery time, and qualifications. When A cannot cooperate, B or C can be used directly without restarting the recommendation process, thereby improving the resilience of the supply chain.

[0048] In some embodiments, see Figure 3 , Figure 3 This is a flowchart illustrating steps S301-S303 provided in the embodiments of this application. The specific process of pushing customized training courses to suppliers based on capability gap tags includes steps S301-S303, which will be explained in conjunction with each step.

[0049] In step S301, when the dynamic capability profile generates new capability deficiency labels or identifies persistent shortcomings, a training recommendation task is triggered in real time. In step S302, matching is completed based on the built-in course knowledge base. Each course in the course knowledge base is labeled with a corresponding resolvable capability gap tag. The course most relevant to the supplier's gap is selected by the recommendation algorithm. In step S303, matching courses are pushed through the enterprise portal, SMS or email. After the supplier completes the training, business data such as the quality pass rate and on-time delivery rate of subsequent orders are continuously monitored to determine the training effect.

[0050] The triggering conditions are divided into two categories: one is the generation of new capability deficiency labels (such as suppliers adding labels for low quality pass rate), and the other is persistent shortcomings (such as labels for delivery capability that need improvement that have existed for more than 2 months without improvement). Both types of conditions trigger training recommendation tasks in real time. The logic for real-time triggering is as follows: every time the dynamic profile engine updates the supplier label, it sends a trigger signal to the AI ​​empowerment center, and the AI ​​empowerment center immediately starts the training recommendation process to avoid training delays.

[0051] The system's built-in course knowledge base adopts a tag-course mapping structure. Each course is labeled with a tag indicating the capability gaps it can address (e.g., lean production management is labeled with tags indicating quality control needs improvement and cost control needs improvement, and supply chain collaboration planning is labeled with tags indicating delivery capability needs improvement). The recommendation algorithm selects courses based on tag matching. For example, if a supplier has both low quality pass rate and delivery delay tags, the algorithm will prioritize matching courses that cover both tags, or recommend corresponding courses according to tag priority (e.g., quality tags have higher priority than delivery tags), ensuring that courses are accurately matched to the gaps.

[0052] Based on supplier outreach habits, the system utilizes multiple channels—including the enterprise portal (a pop-up notification on the supplier's homepage after login), SMS (sending course links to contacts' mobile phones), and email (attaching course details and access points)—to improve reach. The core of effectiveness monitoring is tracking business data after training. For example, after a supplier completes a quality control course, the system continuously monitors their product quality pass rate for the following three months. If the pass rate increases from 85% to 95%, the training is deemed effective. If the pass rate shows no significant change (remaining below 90%), the training is considered ineffective, triggering a new round of recommendations (such as recommending more basic quality inspection courses or adding practical training resources) to ensure the training truly addresses shortcomings.

[0053] In some embodiments, the feedback loop is specifically implemented as follows: If the purchasing personnel do not select the top 1 supplier in the recommendation list or mark the recommendation results as inaccurate, the behavioral data will be asynchronously sent back to the Kafka message queue to drive incremental training of the recommendation algorithm. If the business data does not improve after the supplier completes the training, a new round of differentiated learning resource recommendations will be triggered; if the business data improves significantly, the empowerment success will be recorded and the correlation weight of the corresponding courses and capability gap tags will be strengthened.

[0054] When procurement personnel do not select the top-1 supplier in the recommendation list, or manually mark the recommendation results as inaccurate, the system asynchronously sends this behavior data (e.g., reason for not selecting top-1: price too high, inaccurate marking: risk level assessment deviation) back to the Kafka message queue. Asynchronous sending is to avoid affecting the real-time operation of procurement business. After the data is sent back, the Flink real-time computing engine extracts key information from the feedback (e.g., price too high reflects insufficient weight of value benefit dimension) and uses it as incremental training data for the AI ​​recommendation algorithm to adjust the weight of each scoring criterion in the algorithm (e.g., increase the weight of procurement cost in the overall score) so that subsequent recommendations are more in line with the decision preferences of procurement personnel.

[0055] If the supplier's business data does not improve after completing the training (e.g., the on-time delivery rate is still below the threshold after learning the delivery capability course), the system will automatically trigger a new round of differentiated recommendations and adjust the recommendation strategy (e.g., change the course type: from theoretical courses to practical courses; or supplement the supporting resources: such as inviting industry experts for one-on-one tutoring) to avoid repeatedly recommending ineffective courses; If business data shows significant improvement (e.g., the quality pass rate increases from 80% to 95%), the system records the empowerment as successful and strengthens the association weight between the corresponding course and the weakness tag (e.g., the association weight between the quality control course and the low quality pass rate tag increases from 0.8 to 0.9). When encountering suppliers with the same tag in the future, the system will prioritize recommending that course to improve recommendation efficiency. At the same time, the successful case will be included in the best practice library to provide reference for other suppliers.

[0056] Through the processing of the two types of feedback mentioned above, the system forms a closed loop of recommendation / training-feedback-model optimization-re-recommendation / retraining, realizing continuous iteration of technical solutions and ensuring that supplier management capabilities continue to improve as business progresses.

[0057] In summary, the embodiments of this application have the following beneficial effects: By real-time acquisition and integrated batch processing of multi-source heterogeneous data, data silos in traditional supplier management are broken down, and the problem of lagging static assessments is solved. Based on the multi-dimensional dynamic capability profiles built by GuassDWS, precise quantification and intelligent identification of weaknesses in core dimensions such as supplier quality control, delivery capabilities, and cost control are achieved, effectively improving the timeliness of risk warnings and the comprehensiveness of assessments. Leveraging a two-way AI empowerment mechanism, on the one hand, through NLP demand analysis and multi-objective optimization ranking algorithms, optimal supplier recommendations are provided to buyers, taking into account business matching, historical performance, risk level, and supply chain resilience, avoiding order allocation imbalances caused by crude resource matching. On the other hand, this provides support for high-quality small and medium-sized enterprises. On the one hand, it creates equal exposure opportunities for suppliers; on the other hand, it pushes customized training based on capability gap tags and forms an effect tracking closed loop, promoting targeted improvement of supplier capabilities and solving the problems of passive inefficiency and disconnected content in traditional training. In addition, the full-link closed-loop design of procurement decision feedback and training effect data to feed back into the model optimization enables the system to have the ability to continuously iterate and self-optimize. Ultimately, it transforms supplier management from a passive, experience-driven model to a proactive, AI-driven collaborative model, significantly improving the efficiency of supply chain resource allocation, reducing cooperation risks and operating costs, and can be widely applied to multiple industry scenarios such as retail e-commerce, manufacturing, and financial services, with strong practical application value and promotion significance.

[0058] Based on the same inventive concept, this application also provides a supplier lifecycle management device based on dynamic profiling and AI, which corresponds to the supplier lifecycle management method based on dynamic profiling and AI in the first embodiment. Since the principle of the device in this application is similar to the supplier lifecycle management method based on dynamic profiling and AI, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be described again.

[0059] like Figure 4 As shown, Figure 4 This is a schematic diagram of the supplier lifecycle management device 400 based on dynamic profiling and AI empowerment provided in this application embodiment. The supplier lifecycle management device 400 based on dynamic profiling and AI empowerment includes: The data processing module 401 is used to use Kafka message queue as a unified data bus, and to access the supplier's basic static data, business dynamic data, external public opinion data and supply chain relationship data through API interface, log collection, IoT devices and web crawler. It uses Flink real-time computing engine to clean and transform high-speed data streams, and at the same time uses MRS big data platform to batch import and process large-scale historical data. The tag generation module 402 is used to build a multi-dimensional capability assessment model based on GuassDWS. The model includes core dimensions such as quality control, delivery capability, cost control, technical strength, and response speed. The supplier capability score is calculated through quantifiable indicators of each dimension, and the score is compared with industry benchmarks or preset thresholds to automatically identify capability shortcomings and generate corresponding capability shortcoming tags. The filtering and push module 403 is used to analyze the procurement needs of buyers and filter, sort and recommend suppliers; and to push customized training courses to suppliers based on their capability gap tags. The feedback collection module 404 is used to collect procurement decision feedback data and training effect data through human-computer interaction via the buyer portal and supplier portal, and to send the data back to the data processing link to drive incremental training and optimization of the capability assessment model and recommendation algorithm.

[0060] Those skilled in the art should understand that Figure 4 The functions of each unit in the supplier lifecycle management device 400 based on dynamic profiling and AI can be understood by referring to the relevant description of the supplier lifecycle management method based on dynamic profiling and AI. Figure 4 The functions of each unit in the supplier lifecycle management device 400 based on dynamic profiling and AI empowerment shown can be implemented through a program running on a processor or through specific logic circuits.

[0061] In one possible implementation, the basic static data includes supplier business information, qualification certifications, and registered capital, which are derived from ERP or SCRM systems. The business dynamic data includes order fulfillment rate, on-time delivery rate, product quality pass rate, transaction frequency and scale, which are collected in real time through a log collector; The external public opinion data is obtained by crawling industry association announcements, penalty notices, news public opinion and social media evaluations, and processed using NLP technology to perceive external risks to suppliers. The supply chain relationship data is extracted from the enterprise database and includes supplier equity structure, investment relationships, and information on senior management connections.

[0062] In one possible implementation, the quantitative calculation of the delivery capability dimension is based on the supplier's on-time delivery rate and order fulfillment rate. When the score of either dimension is consistently lower than a preset threshold, the system automatically generates a corresponding capability deficiency label and updates it to the supplier's dynamic profile in real time.

[0063] In one possible implementation, the specific process of analyzing the buyer's procurement needs and screening, ranking, and recommending suppliers is as follows: By using NLP technology to identify key information in procurement requirement texts, constraints such as product category, process, delivery date, and budget can be automatically extracted. Based on the extracted constraints, suppliers that meet the qualification and category requirements are retrieved from the supplier pool of the Elasticsearch index to form a candidate supplier set; A multi-objective optimization ranking algorithm is used to score and rank suppliers in the candidate set to generate a recommendation list.

[0064] In one possible implementation, the scoring criteria of the multi-objective optimization ranking algorithm include: business matching degree, historical performance, risk level, value return and supply chain resilience, and a list of alternative suppliers is generated simultaneously during the ranking process to reduce the risk of dependence on a single supplier.

[0065] In one possible implementation, the specific process of pushing customized training courses to suppliers based on capability gap tags is as follows: When the dynamic competency profile generates new competency gap labels or identifies persistent gaps, training recommendation tasks are triggered in real time. Matching is performed based on a built-in course knowledge base, in which each course is labeled with a corresponding resolvable capability gap tag, and a recommendation algorithm is used to select the courses most relevant to the supplier's gaps. By sending matching courses through the enterprise portal, SMS or email, and after the supplier completes the training, the enterprise continuously monitors business data such as the quality pass rate and on-time delivery rate of subsequent orders to determine the training effectiveness.

[0066] In one possible implementation, the feedback loop is specifically implemented as follows: If the purchasing personnel do not select the top 1 supplier in the recommendation list or mark the recommendation results as inaccurate, the behavioral data will be asynchronously sent back to the Kafka message queue to drive incremental training of the recommendation algorithm. If the business data does not improve after the supplier completes the training, a new round of differentiated learning resource recommendations will be triggered; if the business data improves significantly, the empowerment success will be recorded and the correlation weight of the corresponding courses and capability gap tags will be strengthened.

[0067] The aforementioned supplier lifecycle management device, based on dynamic profiling and AI empowerment, breaks down data silos in traditional supplier management and solves the problem of lagging static assessments through real-time collection and integrated batch processing of multi-source heterogeneous data. Relying on the multi-dimensional dynamic capability profiles built by GuassDWS, it achieves precise quantification and intelligent identification of weaknesses in core dimensions such as supplier quality control, delivery capabilities, and cost control, effectively improving the timeliness of risk warnings and the comprehensiveness of assessments. Through a two-way AI empowerment mechanism, on the one hand, it provides buyers with optimal supplier recommendations that balance business matching, historical performance, risk level, and supply chain resilience through NLP demand analysis and multi-objective optimization ranking algorithms, avoiding order placement errors caused by inefficient resource matching. While addressing the imbalance in single-source allocation, the system creates fair exposure opportunities for high-quality small and medium-sized suppliers. On the other hand, it pushes customized training based on capability gap tags and forms a closed loop for effect tracking, promoting targeted improvement of supplier capabilities and solving the problems of passive inefficiency and disconnected content in traditional training. In addition, the full-link closed-loop design of procurement decision feedback and training effect data to feed back into the model optimization enables the system to have the ability to continuously iterate and self-optimize. Ultimately, it transforms supplier management from a passive, experience-driven model to a proactive, AI-driven collaborative model, significantly improving the efficiency of supply chain resource allocation, reducing cooperation risks and operating costs. Moreover, it can be widely applied to multiple industry scenarios such as retail e-commerce, manufacturing, and financial services, and has strong practical application value and promotion significance.

[0068] like Figure 5 As shown, Figure 5 This is a schematic diagram of the composition structure of the electronic device 500 provided in the embodiments of this application. The electronic device 500 includes: The device includes a processor 501, a storage medium 502, and a bus 503. The storage medium 502 stores machine-readable instructions that can be executed by the processor 501. When the electronic device 500 is running, the processor 501 communicates with the storage medium 502 via the bus 503. The processor 501 executes the machine-readable instructions to perform the steps of the supplier lifecycle management method based on dynamic profiling and AI empowerment described in the embodiments of this application.

[0069] In practical applications, the various components in the electronic device 500 are coupled together via a bus 503. It is understood that the bus 503 is used to achieve communication between these components. In addition to a data bus, the bus 503 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in... Figure 5 The general designated all buses as Bus 503.

[0070] The aforementioned electronic equipment, through real-time acquisition and integrated batch processing of multi-source heterogeneous data, breaks down data silos in traditional supplier management and solves the problem of lagging static assessments. Relying on the multi-dimensional dynamic capability profiles built by GuassDWS, it achieves precise quantification and intelligent identification of weaknesses in core dimensions such as supplier quality control, delivery capabilities, and cost control, effectively improving the timeliness of risk warnings and the comprehensiveness of assessments. Leveraging an AI-powered two-way empowerment mechanism, on the one hand, through NLP demand analysis and multi-objective optimization ranking algorithms, it provides buyers with optimal supplier recommendations that balance business matching, historical performance, risk level, and supply chain resilience, avoiding order allocation imbalances caused by crude resource matching. Simultaneously, it provides... On the one hand, it creates fair exposure opportunities for small, high-quality suppliers. On the other hand, it pushes customized training based on capability gap tags and forms an effect tracking closed loop, promoting targeted improvement of supplier capabilities and solving the problems of passive inefficiency and disconnected content in traditional training. In addition, the full-link closed-loop design of procurement decision feedback and training effect data to feed back into the model optimization enables the system to have the ability to continuously iterate and optimize itself. Ultimately, it transforms supplier management from a passive, experience-driven model to a proactive, AI-driven collaborative model, significantly improving the efficiency of supply chain resource allocation, reducing cooperation risks and operating costs. Moreover, it can be widely applied to multiple industry scenarios such as retail e-commerce, manufacturing, and financial services, and has extremely strong practical application value and promotion significance.

[0071] This application also provides a computer-readable storage medium storing executable instructions. When the executable instructions are executed by at least one processor 501, the supplier lifecycle management method based on dynamic profiling and AI empowerment described in this application is implemented.

[0072] In some embodiments, the storage medium may be a magnetic random access memory (FRAM), a read-only memory (ROM), or a programmable read-only memory (PROM). Erasable Programmable Read-Only Memory (EPROM) Electrically Erasable Programmable Read-Only Memory (EEPROM) Read-only memory, flash memory, magnetic surface storage, optical disc, or CD-ROM ROM, Compact Disc Read It can be a memory such as a memory only; or it can be a device that includes one or any combination of the above-mentioned memories.

[0073] In some embodiments, executable instructions may take the form of a program, software, software module, script, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a standalone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0074] As an example, executable instructions may, but do not necessarily, correspond to files in the file system. They may be stored as part of a file that holds other programs or data, for example, in one or more scripts in a HyperText Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple collaborating files (e.g., a file that stores one or more modules, subroutines, or code sections).

[0075] As an example, executable instructions can be deployed to execute on a single computing device, or on multiple computing devices located in one location, or on multiple computing devices distributed across multiple locations and interconnected via a communication network.

[0076] The aforementioned computer-readable storage media, through real-time acquisition and integrated batch processing of multi-source heterogeneous data, breaks down data silos in traditional supplier management and solves the problem of lag in static assessment. Relying on the multi-dimensional dynamic capability profile built by GuassDWS, it achieves precise quantification and intelligent identification of weaknesses in core dimensions such as supplier quality control, delivery capabilities, and cost control, effectively improving the timeliness of risk warnings and the comprehensiveness of assessments. Leveraging an AI-powered two-way empowerment mechanism, on the one hand, through NLP demand analysis and multi-objective optimization ranking algorithms, it provides buyers with optimal supplier recommendations that balance business matching, historical performance, risk level, and supply chain resilience, avoiding order allocation imbalances caused by crude resource matching. On the one hand, it creates fair exposure opportunities for high-quality small and medium-sized suppliers. On the other hand, it pushes customized training based on capability gap tags and forms an effect tracking closed loop, promoting targeted improvement of supplier capabilities and solving the problems of passive inefficiency and disconnected content in traditional training. In addition, the full-link closed-loop design of procurement decision feedback and training effect data to feed back into the model optimization enables the system to have the ability to continuously iterate and optimize itself. Ultimately, it transforms supplier management from a passive, experience-driven model to a proactive, AI-driven collaborative model, significantly improving the efficiency of supply chain resource allocation, reducing cooperation risks and operating costs. Moreover, it can be widely applied to multiple industry scenarios such as retail e-commerce, manufacturing, and financial services, and has extremely strong practical application value and promotion significance.

[0077] In the several embodiments provided in this application, it should be understood that the disclosed methods and electronic devices can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components may be combined, or integrated into another system, or some features may be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed may be through some interfaces, and the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0078] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0079] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0080] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a platform server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0081] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A supplier lifecycle management method based on dynamic profiling and AI empowerment, characterized in that, Includes the following steps: Using Kafka message queue as a unified data bus, we access basic static data, business dynamic data, external public opinion data and supply chain relationship data from suppliers through API interfaces, log collection, IoT devices and web crawlers. We use the Flink real-time computing engine to clean and transform high-speed data streams, and at the same time, we use the MRS big data platform to batch import and process large-scale historical data. Based on GuassDWS, a multi-dimensional capability assessment model is constructed. The model includes core dimensions such as quality control, delivery capability, cost control, technical strength, and response speed. The supplier capability score is calculated through quantifiable indicators of each dimension, and the score is compared with industry benchmarks or preset thresholds to automatically identify capability shortcomings and generate corresponding capability shortcoming labels. For buyers, we analyze their procurement needs and filter, rank, and recommend suppliers; for suppliers, we push customized training courses based on their capability gap tags. Human-computer interaction is achieved through the buyer portal and supplier portal, and procurement decision feedback data and training effect data are collected. The data is then sent back to the data processing link to drive incremental training and optimization of the capability assessment model and recommendation algorithm.

2. The method according to claim 1, characterized in that, The basic static data includes supplier business information, qualification certifications, and registered capital, which are derived from ERP or SCRM systems. The business dynamic data includes order fulfillment rate, on-time delivery rate, product quality pass rate, transaction frequency and scale, which are collected in real time through a log collector; The external public opinion data is obtained by crawling industry association announcements, penalty notices, news public opinion and social media evaluations, and processed using NLP technology to perceive external risks to suppliers. The supply chain relationship data is extracted from the enterprise database and includes supplier equity structure, investment relationships, and information on senior management connections.

3. The method according to claim 1, characterized in that, The quantitative calculation of delivery capability is based on the supplier's on-time delivery rate and order fulfillment rate. When the score of either dimension is consistently lower than the preset threshold, the system automatically generates a corresponding capability deficiency label and updates it to the supplier's dynamic profile in real time.

4. The method according to claim 1, characterized in that, The specific process for analyzing the purchasing needs of buyers and selecting, ranking, and recommending suppliers is as follows: By using NLP technology to identify key information in procurement requirement texts, constraints such as product category, process, delivery date, and budget can be automatically extracted. Based on the extracted constraints, suppliers that meet the qualification and category requirements are retrieved from the supplier pool of the Elasticsearch index to form a candidate supplier set; A multi-objective optimization ranking algorithm is used to score and rank suppliers in the candidate set to generate a recommendation list.

5. The method according to claim 4, characterized in that, The scoring criteria of the multi-objective optimization ranking algorithm include: business matching degree, historical performance, risk level, value return and supply chain resilience. In addition, a list of alternative suppliers is generated simultaneously during the ranking process to reduce the risk of dependence on a single supplier.

6. The method according to claim 1, characterized in that, The specific process for pushing customized training courses to suppliers based on capability gap tags is as follows: When the dynamic competency profile generates new competency gap labels or identifies persistent gaps, training recommendation tasks are triggered in real time. Matching is performed based on a built-in course knowledge base, in which each course is labeled with a corresponding resolvable capability gap tag, and a recommendation algorithm is used to select the courses most relevant to the supplier's gaps. By sending matching courses through the enterprise portal, SMS or email, and after the supplier completes the training, the enterprise continuously monitors business data such as the quality pass rate and on-time delivery rate of subsequent orders to determine the training effectiveness.

7. The method according to claim 1, characterized in that, The specific implementation method of feedback loop is as follows: If the purchasing personnel do not select the top 1 supplier in the recommendation list or mark the recommendation results as inaccurate, the behavioral data will be asynchronously sent back to the Kafka message queue to drive incremental training of the recommendation algorithm. If the business data does not improve after the supplier completes the training, a new round of differentiated learning resource recommendations will be triggered; if the business data improves significantly, the empowerment success will be recorded and the correlation weight of the corresponding courses and capability gap tags will be strengthened.

8. A supplier lifecycle management device based on dynamic profiling and AI empowerment, characterized in that, The device includes: The data processing module uses Kafka message queue as a unified data bus to access basic static data, business dynamic data, external public opinion data and supply chain relationship data from suppliers through API interfaces, log collection, IoT devices and web crawlers. It uses Flink real-time computing engine to clean and transform high-speed data streams, and at the same time uses the MRS big data platform to batch import and process large-scale historical data. The tag generation module is used to build a multi-dimensional capability assessment model based on GuassDWS. The model includes core dimensions such as quality control, delivery capability, cost control, technical strength, and response speed. The module calculates the supplier's capability score through quantifiable indicators of each dimension and compares the score with industry benchmarks or preset thresholds to automatically identify capability shortcomings and generate corresponding capability shortcoming tags. The filtering and recommendation module is used to analyze the procurement needs of buyers and filter, sort and recommend suppliers; and to push customized training courses to suppliers based on their capability gap tags. The feedback collection module is used to collect procurement decision feedback data and training effect data through human-computer interaction via the buyer portal and supplier portal, and to send the data back to the data processing link to drive incremental training and optimization of the capability assessment model and recommendation algorithm.

9. An electronic device, characterized in that, include: The device includes a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the supplier lifecycle management method based on dynamic profiling and AI empowerment as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, performs the supplier lifecycle management method based on dynamic profiling and AI empowerment as described in any one of claims 1 to 7.

Images