Hardware supply chain intelligent matching recommendation system and method

Abstract

The application discloses a hardware supply chain intelligent matching recommendation system and method, and particularly relates to the technical field of intelligent matching recommendation, which parses original demand information submitted by a purchaser, extracts key fields and generates a feature vector; by comparing current and historical procurement records, difference fields are identified, and a two-dimensional vector is constructed according to a change amplitude and a business weight, and a clustering algorithm is used to divide difference risks into four levels. The system further calculates the functional dependence between materials, historical co-occurrence and procurement quantity correlation, and quantifies strong correlation. Combined with the difference level and the strong correlation index, the system calculates the comprehensive urgency, and determines the supplier selection strategy accordingly: when the urgency is low, the global supplier is screened through similarity matching; when the urgency is high, the historical cooperative supplier is preferentially recommended, so that fine classification of procurement difference risks and intelligent supplier decision-making are realized, and the efficiency of supply chain management is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent matching recommendation, more specifically, the present application relates to a hardware supply chain intelligent matching recommendation system and method. BACKGROUND

[0002] With the development of intelligent manufacturing and digital supply chain, the production link of hardware products puts forward higher requirements on the response speed, accuracy and flexibility of the supply chain. Most of the existing supply chain management systems focus on static supplier information management or simple recommendation based on material matching, and lack the ability of multi-dimensional semantic understanding and quantitative matching between procurement demand and supplier capability. At the same time, the existing technology is generally difficult to identify the difference risk between procurement demand and historical procurement data, and lacks in-depth analysis of the correlation between procurement materials, and cannot effectively judge whether the materials should be bundled for procurement and how to manage the difference risk. Therefore, when the procurement demand changes or there is a potential risk, it often relies on manual experience to determine, resulting in insufficient accuracy of supplier recommendation, making it difficult to respond to sudden changes in market or production demand, affecting the continuity of production plan and the overall efficiency of the supply chain. In view of the above problems, there is an urgent need for a method and system that can realize intelligent analysis of procurement demand, quantitative classification of difference risk, and intelligent supplier recommendation based on historical data, to improve the intelligent level and decision efficiency of supply chain management. SUMMARY

[0003] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present application provide a hardware supply chain intelligent matching recommendation system and method to solve the problems raised in the background art.

[0004] To achieve the above object, the present application provides the following technical scheme:

[0005] The hardware supply chain intelligent matching recommendation method comprises the following steps:

[0006] The original procurement demand information submitted by the procurement party is analyzed, the key fields are extracted and structured, and the procurement demand feature vector is generated. The current procurement list and the historical procurement record list are compared, the difference fields are identified, and a two-dimensional vector is constructed based on the change amplitude and business weight of the difference fields. The clustering algorithm is used to map the difference fields to different difference field levels;

[0007] According to the support degree, the promotion degree and the correlation coefficient of the time series of the procurement quantity between the procurement materials, the strong correlation index between the materials is generated. According to the difference field level and the strong correlation index between the materials, the comprehensive urgency is calculated, and according to the comprehensive urgency, different supplier selection strategies are adopted: when the comprehensive urgency is lower than the system threshold, the optimal supplier is selected among all suppliers through similarity matching; when the comprehensive urgency is higher than the system threshold, the supplier is selected from the suppliers that have cooperated with the procurement party in the past.

[0008] In a preferred embodiment, the identification of the difference field, in combination with the historical procurement record list and the current procurement list, first extracts all material items from the current procurement list, aligns and matches them with the historical procurement records according to their unique identifiers; after completing the pairing, semantic mapping and format standardization are performed through the field dictionary to unify the field naming, data type and unit of measurement, and then, the current-historical record pairs are compared field by field to identify the difference field.

[0009] In a preferred embodiment, the business weight is obtained through offline learning based on historical material shortages and rework losses.

[0010] In a preferred embodiment, K-Means clustering algorithm is used to cluster the two-dimensional vectors formed by the change amplitude of the difference field and the business weight, and the difference field is mapped to different difference field levels.

[0011] In a preferred embodiment, the support degree between the procurement materials is obtained by counting the frequency ratio of the simultaneous appearance of material A and material B in the procurement order. ; where n(A∩B) is the number of co-occurrence of material A and material B, n(A), n(B) are their respective procurement frequencies, and N is the total number of procurement orders; the correlation coefficient of the procurement quantity time series ; where qA(t), qB(t) represent the procurement quantity sequences of material A and material B at time t, respectively, capturing the consistency of their demand changes.

[0012] In a preferred embodiment, the support degree, lift degree and correlation coefficient of the procurement quantity time series are weighted and summed to obtain the strong correlation index between materials.

[0013] In a preferred embodiment, the comprehensive urgency is calculated by weighted summation according to the difference field level and the strong correlation index between materials.

[0014] In a preferred embodiment, when the comprehensive urgency is higher than the system threshold, the supplier is selected from the historical cooperative suppliers first, the coverage, capability score and supplier proportion are weighted and summed to calculate the supplier comprehensive score, and the candidate supplier with high supplier comprehensive score is selected as the procurement supplier. The candidate supplier is the set of all historical cooperative suppliers.

[0015] ​In a preferred embodiment, the supplier proportion is the ratio of the number of supply chains containing the supplier to the total number of supply chains; the coverage is 1 if all coupling components can be supplied at one time; if only p% of the coupling components are covered, the coverage is taken as the value; the capability score is obtained by weighted summation of the technical similarity, the delivery reliability and the historical performance after multi-dimensional normalization.

[0016] In a preferred embodiment, the following modules are included:

[0017] The procurement demand analysis and vectorization module is used to analyze the original procurement demand information submitted by the procurement party, extract and structure the key fields, and generate a procurement demand feature vector;

[0018] The difference analysis and risk assessment module is used to compare the current procurement list with the historical procurement record list, identify the difference fields, and construct a two-dimensional vector based on the change amplitude of the difference fields and the business weight; the clustering algorithm is used to map the difference fields to different difference field levels;

[0019] The comprehensive urgency and supplier strategy selection module is used to generate a strong correlation index between materials according to the support degree, the promotion degree and the correlation coefficient of the time series of the procurement quantity; the comprehensive urgency is calculated according to the difference field level and the strong correlation index between materials, and different supplier selection strategies are adopted according to the comprehensive urgency: when the comprehensive urgency is lower than the system threshold, the optimal supplier is selected from all suppliers through similarity matching; when the comprehensive urgency is higher than the system threshold, the supplier is selected from the suppliers that have cooperated with the procurement party in the past;

[0020] The supplier comprehensive score and recommendation module is used to select the supplier from the suppliers that have cooperated with the procurement party in the past when the comprehensive urgency is higher than the system threshold, to calculate the supplier comprehensive score by weighted summation of the coverage, the capability score and the supplier proportion, and to select the candidate supplier with a high supplier comprehensive score as the procurement supplier, wherein the candidate supplier is the set of all suppliers that have cooperated with the procurement party in the past.

[0021] The technical effects and advantages of the present application are as follows:

[0022] The present application realizes accurate identification of complex free text and structured data by intelligent analysis and vectorization processing of procurement demand information, significantly improves the understanding ability of the system to diversified procurement demand. By combining historical procurement records, using the difference field change amplitude and business weight to build a two-dimensional vector, and using clustering algorithm for risk grading, the quantitative management of procurement difference risk is effectively realized, and the problem of insufficient understanding of multi-dimensional difference and supply chain coupling relationship in the prior art is overcome. At the same time, the proposed Must-Buy Index (MBI) accurately describes the strong correlation between materials through multi-dimensional indexes such as support, promotion and time series correlation of procurement quantity, provides a scientific basis for bundled procurement, and improves the intelligent level of procurement decision.

[0023] The present application also adaptively selects a supplier strategy according to the comprehensive urgency, when the urgency is low, the optimal supplier is screened globally through similarity matching, when the urgency is high, the historical cooperative supplier is preferentially recommended, ensuring the stability and rapid response of the supply chain. By calculating the coverage, ability score and historical proportion of the supplier, a multi-dimensional comprehensive scoring model is formed, and the scientific selection of the supplier is realized. Overall, the present application significantly improves the intelligence and automation of the procurement process, enhances the flexibility and decision efficiency of the supply chain, and has important practical application value and broad popularization prospect. BRIEF DESCRIPTION OF DRAWINGS

[0024] In order to facilitate the understanding of those skilled in the art, the present application will be further described below in conjunction with the drawings;

[0025] Figure 1 The flowchart of the hardware supply chain intelligent matching recommendation method of the present application is shown in the figure;

[0026] Figure 2 The flowchart of the difference risk quantification and urgency modeling is shown in the figure;

[0027] Figure 3 The structure diagram of the hardware supply chain intelligent matching recommendation system of the present application is shown in the figure. DETAILED DESCRIPTION

[0028] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0029] Embodiment 1: The hardware supply chain intelligent matching recommendation method of the present application, as shown in the figure, includes the following steps: Figure 1

[0030] ​Step S1: Analysis of Procurement Requirement Information

[0031] This step is used to perform semantic understanding and structured processing on the original procurement requirements submitted by the purchaser, laying the data foundation for subsequent matching and recommendation.

[0032] First, the system receives the original procurement requirements submitted by the purchaser. This information may originate from standardized forms, free text descriptions, or attachments containing technical parameters. To ensure universality, the system uniformly converts all input content into a processable text stream, and uses this as the basis for requirement parsing.

[0033] During the text processing phase, the system first preprocesses the natural language content, employing technologies such as Chinese word segmentation, part-of-speech tagging, and named entity recognition to identify key material names, technical processes, specifications, units of measurement, delivery time requirements, and relevant certification requirements. For content involving numerical information, the system automatically identifies and standardizes the processing units. If the original text contains referential, ambiguous, or missing expressions, such as "delivery as soon as possible," the system will perform conventional normalization based on the contextual semantics or platform default rules; for example, "as soon as possible" is defaulted to a 14-day delivery period.

[0034] After completing keyword extraction and numerical standardization, the system maps the original procurement requirements information into structured procurement fields, which mainly include: product category, target application scenario, material type, process flow, dimensional accuracy, functional indicators, delivery time, batch size, cost constraints, quality certification requirements, etc.

[0035] Step S2: Construct the procurement demand feature vector

[0036] The purpose of this step is to transform the structured procurement fields (i.e., the original procurement requirement information) into a numerical feature vector representation, which will serve as the input for subsequent steps.

[0037] First, based on the structured procurement fields obtained in step S1, each field is encoded and quantified. For categorical fields (such as target product category, industry, certification requirements, etc.), the system uses one-hot encoding or embedded vector encoding to represent each value, transforming each category value into a sparse or dense vector of a fixed dimension.

[0038] For numerical fields (such as batch quantity, delivery cycle, wall thickness size, budget unit price, etc.), the system will normalize them. Common methods include max-min normalization, Z-score standardization, or piecewise mapping, etc. to eliminate the scale difference between different numerical dimensions. For example, the field of "delivery time of 15 days" can be normalized to the interval [0, 1] to be compared with the supplier's deliverable capacity field.

[0039] For text requirements extracted from natural language descriptions (such as "high appearance requirements", "long-term cooperation intention"), the system will call word vector models or semantic embedding models (such as Word2Vec, GloVe or BERT, etc.) to encode them into fixed-dimensional semantic vectors. Semantic vectors usually represent the semantic information of a sentence in a 100-768 dimensional continuous space, so that subsequent sentence-level or phrase-level similarity calculations can be performed. For material descriptions with hierarchical structure (such as "plastic > engineering plastic > polycarbonate (PC)"), the system uses hierarchical encoding to retain multiple levels of semantic labels in the vector to enhance the discriminability of the representation.

[0040] After all fields are encoded, the system will combine these numerical, categorical and semantic vectors into a complete feature vector, which can be directly used for matching score calculation with the supplier capability vector. The feature vector dimension is fixed, and the arrangement order is organized according to the field category to ensure the consistency of the input structure. In order to improve the discrimination ability of the vector, this step can also introduce feature weight initialization strategy, that is, according to historical matching data or business expert experience to set the relative importance of some fields, record as initial weight parameters for subsequent scoring function. The final output of the procurement demand feature vector represents the "position" of the current task in the multi-dimensional feature space, which can be used to support content matching, vector search, collaborative filtering and other recommendation mechanisms.

[0041] Step S3: Standardization of supplier capability information

[0042] This step is used to structure and standardize the information of candidate suppliers for subsequent effective matching and scoring with the feature vector of procurement demand. Supplier capability data usually comes from platform registration information, enterprise archives, third-party qualification database or historical transaction records, with complex data structure, including structured fields (such as registered capital, main industry, geographic location), semi-structured fields (such as product catalog, technology list) and unstructured fields (such as enterprise introduction text, customer feedback, etc.). To ensure the uniformity of the matching process, the system needs to convert the supplier information into the same expression as the feature vector of the procurement demand.

[0043] Step S4: Difference risk quantification and urgency modeling

[0044] As shown in Figure 2 , in combination with the historical purchase record list, which integrates the purchase list in the past period (for example, one month, three months), all material items are first extracted from the current purchase list, and matched with the historical purchase record according to its unique identifier (such as item_id, spec_code or composite key item_id+version); the matching logic preferentially selects the historical order with the latest time and the status of "completed" or "warehoused" as the comparison baseline, to ensure the business comparability of the comparison basis.

[0045] After pairing is completed, the system performs semantic mapping and format standardization through the field dictionary, unifies the field naming, data type and measurement unit, for example, converts the price field into a unified currency, the quantity field into a standard piece or standard packaging quantity, the delivery date format into an ISO date, and the enumeration type field is mapped and synonym merged according to the standard dictionary of the current version, to eliminate false differences caused by inconsistent expressions in historical data. Subsequently, each pair of current-historical records is compared field by field.

[0046] After the difference field is identified, two normalized indicators are calculated for each field, one is the change amplitude of the difference field: ; wherein index i corresponds to the i-th difference field in the difference field, is the value of the field in the historical purchase record list, is the value of the field in the current purchase list, and are the maximum and minimum values of the same field in the past about six months (or any selected historical window), and the difference between them gives the historical upper and lower limits, so that whether the field is price, threshold or embedded enumeration, it is linearly converted into relative amplitude; when is close to 0, the change can be almost ignored, and close to 1 means far exceeding the past fluctuations.

[0047] The second is the business weight , which is obtained by offline learning through logistic regression based on historical material shortages and rework losses.

[0048] The change amplitude of each difference field and the business weight are combined into a two-dimensional vector: ;

[0049] to form a sample set ; for clustering modeling. Set the number of risk levels K to be divided, usually choose K = 4: corresponding to {Trivial, Minor, Major, Critical}; can select adaptive strategy: use Elbow Method or Silhouette Coefficient to determine the optimal K.

[0050] Modeling using standard K-Means algorithm, finally get K clustering clusters; give each cluster a risk level label, which can be sorted according to the position of its centroid (mean vector), corresponding to the difference field level named Critical, Major, Minor, Trivial in turn. Further map each difference field to four levels, this embodiment takes four levels as an example. Finally, the difference field level is obtained.

[0051] A continuous and computable strong correlation index, namely Must-Buy Index (MBI), is constructed. The core idea is to quantify multiple dimensions that affect the coupling relationship of procurement, including functional dependence, historical co-occurrence, cost synergy, and supply risk factors, and unify and normalize them after weighting and fusion, output a real number score in the interval [0, 1] as the basis for whether to bundle procurement.

[0052] Specifically, in the historical co-occurrence layer, the proportion of the frequency of material A and material B appearing in the procurement order at the same time is calculated to obtain the support ; and calculate the lift ; where n(A∩B) is the number of co-occurrence of material A and material B, n(A), n(B) is their respective procurement frequency, and N is the total number of procurement orders. If Lift > 1, it means that material A and material B are more related than random occurrence.

[0053] Further introduce the correlation coefficient of the purchase quantity time series ; where qA(t), qB(t) represent the purchase quantity sequence of material A and material B at time t, capturing the consistency of their demand changes.

[0054] The strong correlation index is obtained by integrating the lift, support, and correlation coefficient of the purchase quantity time series. According to the following formula: ; where QG represents the strong correlation index; 、 、 Adjusting parameters can be designed as needed.

[0055] The comprehensive urgency is calculated by summing up the increment level and the strong correlation index. First, the difference field levels Critical, Major, Minor and Trivial are assigned values in sequence; for example, Critical is assigned a value of 1, Major is assigned a value of 0.75, Minor is assigned a value of 0.5, and Trivial is assigned a value of 0.25.

[0056] The comprehensive urgency is calculated by summing up the increment level and the strong correlation index. First, the difference field levels Critical, Major, Minor and Trivial are assigned values in sequence; for example, Critical is assigned a value of 1, Major is assigned a value of 0.75, Minor is assigned a value of 0.5, and Trivial is assigned a value of 0.25. ; wherein Z represents the comprehensive urgency, DG represents the value after the increment level is assigned, and , respectively, are the weight coefficients of the value after the strong correlation index and the increment level are assigned.

[0057] Step S5: Supplier comprehensive score and preferred decision

[0058] If the comprehensive urgency is less than the system threshold, it means that the difference field has a large difference with the historical procurement record list and is relatively unimportant; the supplier is determined according to the similarity matching in the prior art. The following is a brief description.

[0059] A similarity calculation relationship is established between the difference field feature vector and the supplier capability vector, and the difference field feature vector is processed in the same way as the procurement demand feature vector; the matching degree is evaluated according to the closeness in the multi-dimensional feature space, thereby providing a basic score for subsequent recommendation sorting. First, the difference field feature vector and the supplier capability vector set are input into the matching engine, and a one-to-many matching degree calculation process is entered. The core of the matching process is to construct a scoring function to measure the similarity between any supplier capability vector and the difference field feature vector. The scoring function usually adopts multiple calculation strategies for fusion according to the different types of features. The supplier with high similarity is selected.

[0060] If the comprehensive urgency is greater than the system threshold, it means that the difference field has a strong connection with the historical procurement record list and is relatively important; the supplier is selected from the historical cooperating suppliers; and it is more stable.

[0061] For each candidate supplier, the candidate supplier represents a historical cooperating supplier.

[0062] There are multiple original existing supply chains, and the supplier proportion in each supply chain is counted; specifically, the supplier proportion is the number of supply chains containing the supplier divided by the total number of supply chains.

[0063] Two items can be directly queried from the existing vector library: coverage cs∈[0, 1]: if all coupled parts can be supplied at one time, 1 is taken; if only p% is covered, the coverage value is taken, for example, if 10% is covered, the value is 0.1.

[0064] The capability score as∈[0, 1]: the weighted sum of the technical similarity, the delivery reliability, the historical compliance, the unit price and the like after multi-dimensional normalization.

[0065] The comprehensive coverage, the capability score and the supplier proportion are comprehensively considered to calculate a supplier comprehensive score, and a candidate supplier with a high supplier comprehensive score is taken as a procurement supplier.

[0066] The specific formula is as follows: ; wherein ZH represents the supplier comprehensive score; and zb represents the supplier proportion; the higher the supplier proportion, the more popular the supplier is, and the more preferred the procurement party is, and the higher the supplier comprehensive score is; 、 、 are weight coefficients of the coverage, the capability score and the supplier proportion respectively.

[0067] Embodiment 2: The design of the hardware supply chain intelligent matching recommendation system of the application is based on the method in embodiment 1, and specifically as shown in Figure 3 includes the following modules: a procurement demand analysis and vectorization module, a supplier information standardization module, a difference analysis and risk assessment module, a comprehensive urgency and supplier matching module, a supplier comprehensive score and recommendation module;

[0068] The procurement demand analysis and vectorization module is used to receive and analyze the original procurement demand information submitted by the procurement party, extract the material name, technical parameter, specification, delivery period, quantity, authentication requirement and the like through natural language processing technology, and convert the analyzed information into a multi-dimensional procurement demand feature vector through one-hot encoding, normalization or semantic embedding and the like, as the basis data for subsequent matching.

[0069] The supplier information standardization module is used to collect and process multi-dimensional data such as the registration information, product catalog, technical capability and historical transaction record of the supplier, and standardize the data into a supplier capability vector consistent with the format of the procurement demand vector, so as to realize the vectorization input for subsequent matching calculation.

[0070] The difference analysis and risk assessment module is used to compare the current procurement demand with the historical procurement record, identify the difference field, and divide the difference field into four risk levels of Critical, Major, Minor and Trivial by calculating the change amplitude and the business weight trained based on the historical material shortage or rework loss; at the same time, a Must-Buy Index (MBI) is constructed to analyze the functional dependence, historical co-occurrence and procurement quantity time sequence correlation among materials, form a strong correlation index, and assist in decision-making whether to bundle procurement.

[0071] Comprehensive urgency and supplier matching module: based on the risk level and strong correlation indicators, the comprehensive urgency indicator is calculated to determine the urgency of the procurement demand and the historical dependence relationship. If the comprehensive urgency is low, the similarity matching algorithm is used to select the supplier with the highest matching degree from all candidate suppliers; if the urgency is high, the recommended suppliers are selected from the historical cooperation suppliers to ensure the stability and reliability of the supply.

[0072] Supplier comprehensive score and recommendation module: used for comprehensive evaluation of the coverage, capability score and proportion in the historical supply chain of the candidate supplier, the supplier comprehensive score is formed by weighted calculation, the candidate suppliers are sorted, the supplier with the highest comprehensive score is selected as the optimal cooperation object, and the recommendation result is fed back to the user to support subsequent model optimization and continuous learning.

[0073] Those skilled in the art can appreciate that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Professionals can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0074] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be realized in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0075] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0076] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit.

[0077] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of the changes or replacements within the technical range disclosed by the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A hardware supply chain intelligent matching recommendation method, characterized by: The method comprises the following steps: The raw procurement demand information submitted by the procurement party is parsed, key fields are extracted and structured, and a procurement demand feature vector is generated; the current procurement list and the historical procurement record list are compared, difference fields are identified, and a two-dimensional vector is constructed based on the change amplitude of the difference fields and the business weight; a clustering algorithm is used to map the difference fields to different difference field levels; The proportion of the frequency of the statistical material A and the material B appearing in the purchase order at the same time is the support degree between the purchase materials Promotion degree Wherein n(A∩B) is the number of co-occurrence of material A and material B, n(A), n(B) is their respective purchase frequency, N is the total number of purchase orders; The correlation coefficient of the purchase quantity time series is p = corr(q A (t),q B (t)); Wherein q A (t), q B (t) respectively represent the purchase quantity sequence of material A and material B at time t, which captures the consistency of the demand change of the two. The support degree, the promotion degree, and the correlation coefficient of the time series of the procurement quantity between the comprehensive procurement materials are summed by weighting to obtain a strong correlation index between the materials; According to the difference field level and the strong correlation index between the materials, a comprehensive urgency is calculated, and according to the comprehensive urgency, different supplier selection strategies are adopted: when the comprehensive urgency is lower than a system threshold, the optimal supplier is selected from all suppliers through similarity matching; when the comprehensive urgency is higher than the system threshold, the supplier is preferentially selected from the candidate suppliers, a supplier comprehensive score is calculated by weighted summation of the comprehensive coverage, the capability score, and the supplier proportion, and the candidate supplier with the high supplier comprehensive score is selected as the procurement supplier. The candidate supplier is a set of all historical cooperating suppliers. 2.The hardware supply chain intelligent matching recommendation method of claim 1, wherein: The difference fields are identified, the historical procurement record list and the current procurement list are combined, all material items in the current procurement list are first extracted, and alignment and matching are performed according to the unique identifier and the historical procurement record; After pairing, semantic mapping and format standardization are performed through a field dictionary, the field naming, the data type, and the unit of measurement are unified, then, the current-historical record is compared field by field, and the difference fields are identified. 3.The hardware supply chain intelligent matching recommendation method of claim 1, wherein: The business weight is obtained by offline learning based on the historical material shortage and the rework loss through a logistic regression. 4.The hardware supply chain intelligent matching recommendation method of claim 1, wherein: A K-Means clustering algorithm is used to cluster the two-dimensional vector formed by the change amplitude of the difference fields and the business weight, and the difference fields are mapped to different difference field levels. 5.The hardware supply chain intelligent matching recommendation method of claim 1, wherein: The comprehensive urgency is calculated by weighted summation according to the difference field level and the strong correlation index between the materials. 6.The hardware supply chain intelligent matching recommendation method of claim 1, wherein: The supplier proportion is a ratio of the number of supply chains containing the supplier to the total number of supply chains; the coverage degree is 1 if all coupled parts can be supplied at one time; if only p% of the coupled parts are covered, the coverage rate is taken as the value; the capability score is obtained by weighted summation of the technical similarity, the delivery reliability, and the historical performance after multi-dimensional normalization.

7. A hardware supply chain intelligent matching recommendation system, characterized in that, The recommendation system comprises the method according to any one of claims 1-6, and comprises the following modules: A procurement demand analysis and vectorization module is used to parse the raw procurement demand information submitted by the procurement party, extract and structure key fields, and generate a procurement demand feature vector; A difference analysis and risk assessment module is used to compare the current procurement list and the historical procurement record list, identify difference fields, and construct a two-dimensional vector based on the change amplitude of the difference fields and the business weight; a clustering algorithm is used to map the difference fields to different difference field levels; The comprehensive urgency and supplier strategy selection module is used to count the frequency ratio of material A and material B appearing in the procurement order at the same time, and obtain the support degree between the procurement materials lift Wherein n(A∩B) is the number of co-occurrence of material A and material B, n(A), n(B) is their respective procurement frequency, N is the total number of procurement orders; The correlation coefficient of the time series of procurement quantity is p = corr(q A (t),q B (t)); Wherein q A (t), q B (t) respectively represent the procurement quantity sequence of material A and material B at time t, and capture the consistency of the demand change of the two. The support degree, the promotion degree, and the correlation coefficient of the time series of the procurement quantity between the comprehensive procurement materials are summed by weighting to obtain a strong correlation index between the materials; According to the strong correlation index between the difference field level and the material, the comprehensive urgency is calculated, and according to the comprehensive urgency, different supplier selection strategies are adopted: when the comprehensive urgency is lower than the system threshold, the optimal supplier is selected from all suppliers through similarity matching; The supplier comprehensive score and recommendation module is used to select the supplier from the candidate supplier when the comprehensive urgency is higher than the system threshold, the comprehensive coverage, the ability score and the supplier proportion are weighted and summed to calculate the supplier comprehensive score, and the candidate supplier with high supplier comprehensive score is selected as the procurement supplier. The candidate supplier is a set of all historical cooperating suppliers.

Images