A method and system for processing a service procurement plan report

Abstract

The application discloses a kind of overhaul service's purchase plan report processing method and system, first process supplier historical shipment data, construct multistage label system and historical purchase project service performance database;Then based on adjusted NLP model extraction demand entity, generate standardization feature vector;And demand characteristics are mapped with shipment label, and if not matched, then supplement record;Finally, four-dimensional score is calculated, and dynamically distributes weight in combination with historical service data, and total matching score is obtained;Set scene threshold to screen compliant supplier;Label non-compliance reason;The method effectively improves the accuracy of supplier screening, solves the problem that existing technology relies on fuzzy or single-dimensional matching, does not associate historical shipment key attributes with purchase demand, leading to non-compliant suppliers mixing in, enhances the relevance of matching results and actual service capabilities of supplier, solves the problem that dimension weight in existing technology is subjective and fixed, and is disconnected with actual service quality.

Description

Technical Field

[0001] This invention belongs to the field of maintenance service procurement technology, specifically relating to a method and system for processing maintenance service procurement plan reports. Background Technology

[0002] Existing historical shipment volume matching technologies for maintenance service procurement generally adopt keyword fuzzy matching or single-dimensional presence / absence judgment. These methods can only make a preliminary judgment on whether a supplier has a certain level of historical shipment volume, and cannot achieve accurate matching by combining the special characteristics of maintenance procurement scenarios. This approach does not consider the correlation between historical shipment volume level, shipment time range and procurement cycle, resulting in non-compliant suppliers with insufficient historical shipment volume level or shipment time that cannot cover the procurement cycle being included in the candidate range.

[0003] Furthermore, the existing solutions only set fixed values ​​for dimension weights based on the urgency of demand and risk level, lacking correlation analysis with the impact of dimensions on service quality in actual procurement scenarios. The weight allocation is highly subjective, which may lead to a disconnect between high-weight dimensions and key dimensions that actually affect service quality, further reducing the accuracy of supplier matching. Therefore, it is necessary to design a processing method and system for maintenance service procurement plan reports to solve the above problems. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for processing procurement plan reports for maintenance services, aiming to solve the problems of low matching accuracy and subjective weighting in existing technologies.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A method for processing a procurement plan report for maintenance services includes the following steps: S1. Identify and process historical shipment data, extract relevant fields, and then construct a historical shipment label system for maintenance procurement. Based on the historical shipment label system for maintenance procurement, assign structured labels to the fields and store them. Construct a historical procurement project service performance database that is associated with the supplier's historical shipment database and the system threshold database. S2, construct a corpus of historical shipments for maintenance procurement, train an NLP model based on the corpus of historical shipments for maintenance procurement; then input the procurement demand text into the trained NLP model, extract target entities through named entity recognition, label the urgency and risk level of the demand, and output a standardized demand feature vector containing the quantitative value of the entity. S3, based on the historical shipment volume tagging system, maps the target shipment volume level, target shipment service type, and procurement cycle in the demand feature vector to the structured tags in the supplier's historical shipment volume database. If there is no match, it marks the item to be supplemented and triggers a reminder to supplement the supplier's historical shipment volume data. S4 defines four dimensions: matching degree of shipment volume level, matching degree of shipment time, matching degree of shipment service type, and matching degree of shipment customer evaluation. The scores of each dimension are calculated, and then the weights of the dimensions are dynamically allocated by combining the urgency of the demand and the service performance data of historical procurement projects. The total matching score of the supplier is calculated. S5 sets corresponding scenario-based matching thresholds based on the urgency of demand and risk level, compares the total matching score of suppliers with the threshold, retains suppliers whose total matching score is not lower than the threshold, and generates a list of compliant suppliers in descending order of total score. S6: For suppliers whose total matching score is below the threshold, the non-compliance reasons are marked according to the matching degree of shipment volume level, the coverage of shipment time, the adaptability of shipment service type, and the matching degree of shipment customer evaluation, and a list of non-compliant suppliers is generated.

[0006] Preferably, step S1 further includes: S101 uses data collection technology to process the historical shipment data submitted by suppliers. The relevant data includes historical order records, transaction filing data, and purchase ledgers. Then, it extracts four fields from the data: shipment service type, shipment volume level, shipment time range, and shipment customer evaluation. If any fields are missing, it triggers a reminder for the supplier to submit supplementary data. S102, construct a multi-level maintenance procurement historical shipment volume labeling system, including four levels of labels: Primary tag: Shipping service category, divided according to maintenance service type; Secondary label: Shipment volume level, set according to the total shipment range; Level 3 tag: Shipping time parameter, which includes two sub-parameters: shipping start time and shipping end time; Level 4 tag: Customer reviews of shipped products, set according to preset review score ranges; S103 will extract the complete historical shipment volume field and match and assign values ​​with the maintenance procurement historical shipment volume tag system: the shipment service type corresponds to the first-level tag, the shipment volume level corresponds to the second-level tag, the shipment time range corresponds to the third-level tag, and the shipment customer evaluation corresponds to the fourth-level tag. All tags are bound to the supplier ID and stored in the relational database. A historical procurement project service performance database is constructed to store the performance records of past maintenance procurement projects. This database is linked to the supplier's historical shipment volume database and the system threshold database, and stores the following fields: Historical procurement project ID, corresponding supplier ID, demand urgency-risk level combination, supplier four-dimensional score, total supplier matching score, on-time performance rate, service process problem occurrence rate and project acceptance pass rate; The database is composed of performance reports, acceptance records, and service quality assessment documents from the company's past maintenance and procurement projects, and is updated in real time.

[0007] Preferably, step S2 further includes: S201, Collect historical maintenance service procurement requirement texts, and annotate each text segment in two dimensions: The first dimension is entity labeling, which includes three types of entities in the labeled text: target shipment volume level, target shipment service type, and procurement cycle. The second dimension is classification labeling, which includes two types of tags: demand urgency and risk level. The labeling is based on the demand characteristic rules in the maintenance procurement scenario. The labeled text is divided into training set and validation set according to a preset ratio to form a corpus of historical shipment volume in maintenance procurement. S202, a pre-trained NLP model was selected as the base model, and the model was adjusted using the training set of the maintenance and procurement corpus: The model training task is set as a dual task of named entity recognition and text classification. The named entity recognition task is trained to identify three types of entities: target shipment volume level, target shipment service type, and procurement cycle. The text classification task is trained to classify two types of tags: demand urgency and risk level. During the training process, the model performance is monitored in real time using a validation set. When the entity recognition accuracy and classification accuracy of the model on the validation set both reach the preset stable value, the training is stopped, the NLP model is adjusted, and an NLP model adapted to the maintenance and procurement scenario is obtained. S203: Input the procurement requirement text to be processed into the adjusted NLP model. The model first extracts the target shipment volume level, target shipment service type, and procurement cycle entities from the text through the named entity recognition task. Then, it labels the urgency and risk level corresponding to the requirement through the text classification task. The extracted entities are quantified. According to preset rules, the target shipment volume level is converted into a level quantification value, the procurement cycle is converted into a time range quantification value, and the target shipment service type is converted into a code value. The entity quantification values ​​and the labeled urgency and risk levels are combined to form a standardized requirement feature vector containing all information. The vector structure clearly corresponds to the quantification value of each entity and the requirement level label.

[0008] Preferably, step S3 further includes: S301, based on the maintenance history shipment volume labeling system constructed in S1, clarify the correspondence between each entity in the demand feature vector and the labeling system: the target shipment service type entity corresponds to the first-level label of the labeling system, the target shipment volume level entity corresponds to the second-level label of the labeling system, and the procurement cycle entity corresponds to the third-level label of the labeling system. S302: Retrieve supplier ID-structured tag data pairs from the supplier's historical shipment volume database. For each supplier, compare the target shipment volume level quantification value in the demand feature vector with the level parameter of the second-level tag in the supplier's tag system to confirm whether the shipment volume level matches. Then, compare the procurement cycle quantification value with the shipment start time and shipment end time sub-parameters of the third-level tag to confirm whether the validity period of the shipment time range covers the procurement cycle. Finally, compare the target shipment service type code value with the type parameter of the first-level tag to confirm whether the shipment service category is suitable. S303: For any supplier, if the target shipment volume level, target shipment service type, and procurement cycle entity in the demand feature vector do not have a matching tag in the supplier's structured tags, mark the unmatched demand feature as an item to be supplemented, and generate a supplementary reminder instruction by associating it with the corresponding supplier ID.

[0009] Preferably, step S4 further includes: S40 defines four matching dimensions: shipment volume level matching degree, shipment volume level quantitative value, shipment time coverage, and shipment service type adaptability. The scoring calculation logic for each dimension is clearly defined: the shipment volume level matching degree is based on the supplier's shipment volume level quantitative value and the demand target shipment volume level quantitative value. The ratio between the two is calculated. If the supplier's quantitative value is higher than the demand quantitative value, the dimension gets full marks. If the supplier has no corresponding shipment volume tag, it gets zero marks. The delivery time coverage is based on the supplier's delivery time range and the procurement cycle time range. The intersection duration of the two time ranges is calculated, and the ratio of the intersection duration to the total procurement cycle duration is used as the score for this dimension. The service type matching score is calculated based on the supplier's service type code value and the target service type code value. The ratio between the two is calculated. If the supplier's service type code value and the target service type code value are the same, the score is full; if they are not the same, the score is zero. The matching degree of the customer evaluation of the shipment is based on the evaluation score and shipment frequency of the supplier's shipment customers within the supplier's historical preset period. The shipment frequency is converted into a weight coefficient according to the preset rules, and the product of the evaluation score and the weight coefficient is used as the score of that dimension. All dimension scores are unified to the same score range. S402, based on the urgency and risk level of the demand as labeled by the demand feature vector, clarifies that all combinations of the two are categorized into four types: high urgency - high risk level, high urgency - low risk level, low urgency - high risk level, and low urgency - low risk level. Historical project data matching the current maintenance demand urgency-risk level combination is selected from the historical procurement project service performance database. This database stores the company's past maintenance procurement project performance records, with fields including historical procurement project ID, corresponding supplier ID, demand urgency-risk level combination, supplier four-dimensional score, total matching score, on-time performance rate, service process problem incidence rate, and project acceptance pass rate. A multiple linear regression model is constructed using the supplier's overall actual service performance score from historical projects as the dependent variable Y, and the four-dimensional scores as independent variables X1, X2, X3, and X4. X1 corresponds to the shipment volume matching score, X2 to the shipment time coverage score, X3 to the shipment service type suitability score, and X4 to the shipment customer evaluation matching score. Then, the multi-dimensional actual service performance indicators are normalized into a unified score, as shown in the following formula: Y = (On-time performance rate × a + Acceptance rate × b + (1 - Service process problem occurrence rate) × c) × 100; Among them, a, b, and c are preset weight coefficients, which are set according to the enterprise's focus on timely delivery, acceptance, and process compliance in the maintenance and procurement scenario, and satisfy a+b+c=1; The dimensional contribution coefficients are calculated using the least squares method of a multiple linear regression model. The calculation process is as follows: using the independent variable data X1, X2, X3, X4 and the dependent variable data Y of historical projects as samples, minimize the model error term. The formula is as follows: ,in, The constant term in the model is used to fit the historical data using a formula, resulting in the final model. These represent the magnitude of change in the dependent variable Y when X1, X2, X3, and X4 change by 1 unit, respectively. Finally, the contribution coefficient is standardized to the dimensional weight percentage. The absolute value of is converted into a percentage form, as shown in the following formula: ; For each combination, a dimensional priority strategy is adopted, which prioritizes high-urgency needs to ensure delivery time coverage, high-risk needs to ensure delivery volume level matching and delivery service type adaptability, and low-urgency, low-risk needs to ensure delivery customer evaluation matching. Then, the dimensional weight ratios are adjusted based on the calculated dimensional weight ratios: for the highest priority dimension, a preset percentage increase is added to its corresponding base, and the same increase is deducted from the base of the lowest priority dimension. Finally, the weight ratios of delivery volume level matching W1, delivery time coverage W2, delivery service type adaptability W3, and delivery customer evaluation matching W4 are set. S403 uses a weighted summation formula to calculate the total matching score for each supplier. The formula is as follows: ; Where S total S1 represents the supplier's total matching score, S2 represents the shipping grade matching score, S3 represents the shipping time coverage score, S4 represents the shipping service type suitability score, and S5 represents the shipping customer evaluation matching score.

[0010] Preferably, step S5 further includes: S501: Based on the urgency and risk level of the demand marked by the demand feature vector, a unique scenario-based matching threshold is set for each combination type, and the thresholds corresponding to different combination types do not overlap; the correspondence between combination type and scenario-based matching threshold is stored in the system threshold database. S502, retrieve the total matching score S of all suppliers from the calculation results of S403. total Simultaneously, the system retrieves the scenario-based matching threshold corresponding to the current maintenance requirement from the system threshold database; for each supplier, the total matching score is compared with the scenario-based matching threshold, and two types of results are recorded: one type is suppliers with a total matching score higher than the scenario-based matching threshold, and the other type is suppliers with a total matching score lower than the scenario-based matching threshold. The two types of results are associated with the corresponding supplier IDs. S503: Suppliers with total matching scores higher than the scenario-based matching threshold are sorted in descending order of total matching score. In the sorted supplier list, basic information and matching scores for each dimension of each supplier are added to form a structured list of compliant suppliers. The basic information of the suppliers includes the supplier name and contact information. All data in the list is bound to the supplier ID.

[0011] Preferably, S6 further includes: S601 sets a passing score for each matching dimension, which is consistent with the score range for each dimension. The scores for each non-compliant supplier are evaluated sequentially based on the following order: shipment volume matching degree, shipment time coverage, shipment service type suitability, and shipment customer evaluation matching degree. If the shipment volume matching degree score is below the passing score for that dimension, the non-compliance is marked as insufficient shipment volume. If the shipment volume matching degree score meets the standard, but the shipment time coverage score is below the passing score, the non-compliance is marked as insufficient shipment time coverage. If the first two scores meet the standard, but the shipment service type suitability score is below the passing score, the non-compliance is marked as a mismatch in shipment service type. If the first three scores meet the standard, but the total matching score is still below the scenario-based matching threshold, the non-compliance is marked as insufficient comprehensive score across multiple dimensions. S602 integrates the supplier ID, supplier name, contact information, scores for each dimension, and the marked reasons for non-compliance for each non-compliant supplier to form a structured list of non-compliant suppliers. The list must clearly define the correspondence between each field, and the non-compliance reason field must be directly related to the dimension score used for judgment. Finally, the list is stored in the system results database.

[0012] Preferably, a system for processing procurement plan reports for maintenance services is provided, which is used to execute the processing method for the procurement plan reports for maintenance services. The system includes a supplier historical shipment processing and storage module, a procurement demand parsing module, a feature-tag mapping and supplementary entry reminder module, a multi-dimensional matching score calculation module, a compliant supplier screening module, and a non-compliant supplier labeling module. The supplier historical shipment volume processing and storage module is used to process the historical shipment volume data submitted by suppliers, extract the shipment service type, shipment volume level, shipment time range, and shipment customer evaluation fields, and assign and store them according to the historical shipment volume tag system. The procurement demand parsing module is used to extract the target shipment volume level, target shipment service type, and procurement cycle entity from the procurement demand text based on the adjusted NLP model, mark the urgency and risk level of the demand, and generate a standardized demand feature vector containing the quantitative value of the entity. The feature-tag mapping and supplementary data reminder module is responsible for matching entities in the demand feature vector with structured tags in the supplier's historical shipment volume tag library, marking demand features that do not find matching tags as items to be supplemented, and triggering a supplementary data reminder for the supplier's historical shipment volume. The multi-dimensional matching score calculation module is used to define four-dimensional matching standards and calculate the scores of each dimension. It calls the data of the historical procurement project service performance database, calculates the contribution coefficient of each dimension to the actual service performance through a multiple linear regression model and standardizes it into the weight ratio base. It adjusts the weight ratio by combining the urgency of the demand, the risk level and the preset dimension priority strategy, and finally calculates the total matching score of the supplier through a weighted summation formula. The compliant supplier screening module is responsible for setting scenario-based matching thresholds based on the urgency of the need and the risk level, comparing the total matching score of suppliers with the thresholds, retaining qualified suppliers and generating a list of compliant suppliers in descending order of scores. The non-compliant supplier labeling module is used to prioritize non-compliance based on shipment volume and combine the scores of each dimension to label the reasons for non-compliance, integrate supplier information to generate a list of non-compliant suppliers and store it.

[0013] Preferably, a computer device includes a memory and a processor, which are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the processing method for the maintenance service procurement plan report.

[0014] Preferably, a computer-readable storage medium stores computer instructions that cause a computer to execute the processing method for the procurement plan report of the maintenance service.

[0015] The beneficial effects of the method and system for processing procurement plan reports for maintenance services provided by this invention are as follows: 1. Effectively improves the accuracy of supplier selection, solving the problem of non-compliant suppliers being included due to existing technologies relying on vague or single-dimensional matching and failing to link key attributes of historical shipments with procurement needs. This invention constructs a four-level historical shipment tagging system and a database of historical procurement project service performance to structure supplier shipment data; achieves accurate mapping between demand characteristics and shipment tags and automatically triggers supplementary data reminders; and defines a four-dimensional scoring logic based on shipment level, time coverage, service type, and customer evaluation, setting scenario-based thresholds for screening according to demand urgency and risk level. Multi-stage collaboration ensures that only suppliers meeting multiple dimensions are retained, significantly improving the fit between suppliers and procurement needs. 2. Enhance the correlation between matching results and the supplier's actual service capabilities, addressing the problem of subjectively fixed weights for existing technical dimensions that are disconnected from actual service quality. This invention extracts similar demand data from a historical procurement project service performance database, constructs a comprehensive score for actual service performance based on metrics such as on-time delivery rate and acceptance rate, calculates the contribution coefficient of each dimension to service quality using a multiple linear regression model, standardizes the coefficients into weight bases, and fine-tunes them based on demand priority. This ensures that weight allocation is based on objective data rather than subjective experience, aligning high-weight dimensions with key factors affecting service quality, avoiding situations where high scores indicate poor service, and making matching results more closely reflect the supplier's true service capabilities. Attached Figure Description

[0016] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a schematic diagram of the computer device structure in an embodiment of the present invention. Detailed Implementation

[0017] Example 1: like Figure 1 As shown, a method for processing a procurement plan report for maintenance services includes the following steps: S1. Identify and process historical shipment data, extract relevant fields, and then construct a historical shipment label system for maintenance procurement. Based on the historical shipment label system for maintenance procurement, assign structured labels to the fields and store them. Construct a historical procurement project service performance database that is associated with the supplier's historical shipment database and the system threshold database. S2, construct a corpus of historical shipments for maintenance procurement, train an NLP model based on the corpus of historical shipments for maintenance procurement; then input the procurement demand text into the trained NLP model, extract target entities through named entity recognition, label the urgency and risk level of the demand, and output a standardized demand feature vector containing the quantitative value of the entity. S3, based on the historical shipment volume tagging system, maps the target shipment volume level, target shipment service type, and procurement cycle in the demand feature vector to the structured tags in the supplier's historical shipment volume database. If there is no match, it marks the item to be supplemented and triggers a reminder to supplement the supplier's historical shipment volume data. S4 defines four dimensions: matching degree of shipment volume level, matching degree of shipment time, matching degree of shipment service type, and matching degree of shipment customer evaluation. The scores of each dimension are calculated, and then the weights of the dimensions are dynamically allocated by combining the urgency of the demand and the service performance data of historical procurement projects. The total matching score of the supplier is calculated. S5 sets corresponding scenario-based matching thresholds based on the urgency of demand and risk level, compares the total matching score of suppliers with the threshold, retains suppliers whose total matching score is not lower than the threshold, and generates a list of compliant suppliers in descending order of total score. S6: For suppliers whose total matching score is below the threshold, the non-compliance reasons are marked according to the matching degree of shipment volume level, the coverage of shipment time, the adaptability of shipment service type, and the matching degree of shipment customer evaluation, and a list of non-compliant suppliers is generated.

[0018] Preferably, step S1 further includes: S101 uses data collection technology to process the historical shipment data submitted by suppliers. The relevant data includes historical order records, transaction filing data, and purchase ledgers. Then, it extracts four fields from the data: shipment service type, shipment volume level, shipment time range, and shipment customer evaluation. If any fields are missing, it triggers a reminder for the supplier to submit supplementary data. S102, construct a multi-level maintenance procurement historical shipment volume labeling system, including four levels of labels: Primary tag: Shipping service category, divided according to maintenance service type; Secondary label: Shipment volume level, set according to the total shipment range; Level 3 tag: Shipping time parameter, which includes two sub-parameters: shipping start time and shipping end time; Level 4 tag: Customer reviews of shipped products, set according to preset review score ranges; S103 will extract the complete historical shipment volume field and match and assign values ​​with the maintenance procurement historical shipment volume tag system: the shipment service type corresponds to the first-level tag, the shipment volume level corresponds to the second-level tag, the shipment time range corresponds to the third-level tag, and the shipment customer evaluation corresponds to the fourth-level tag. All tags are bound to the supplier ID and stored in the relational database. A historical procurement project service performance database is constructed to store the performance records of past maintenance procurement projects. This database is linked to the supplier's historical shipment volume database and the system threshold database, and stores the following fields: Historical procurement project ID, corresponding supplier ID, demand urgency-risk level combination, supplier four-dimensional score, total supplier matching score, on-time performance rate, service process problem occurrence rate and project acceptance pass rate; The database is composed of performance reports, acceptance records, and service quality assessment documents from the company's past maintenance and procurement projects, and is updated in real time.

[0019] Preferably, step S2 further includes: S201, Collect historical maintenance service procurement requirement texts, and annotate each text segment in two dimensions: The first dimension is entity labeling, which includes three types of entities in the labeled text: target shipment volume level, target shipment service type, and procurement cycle. The second dimension is classification labeling, which includes two types of tags: demand urgency and risk level. The labeling is based on the demand characteristic rules in the maintenance procurement scenario. The labeled text is divided into training set and validation set according to a preset ratio to form a corpus of historical shipment volume in maintenance procurement. S202, a pre-trained NLP model was selected as the base model, and the model was adjusted using the training set of the maintenance and procurement corpus: The model training task is set as a dual task of named entity recognition and text classification. The named entity recognition task is trained to identify three types of entities: target shipment volume level, target shipment service type, and procurement cycle. The text classification task is trained to classify two types of tags: demand urgency and risk level. During the training process, the model performance is monitored in real time using a validation set. When the entity recognition accuracy and classification accuracy of the model on the validation set both reach the preset stable value, the training is stopped, the NLP model is adjusted, and an NLP model adapted to the maintenance and procurement scenario is obtained. S203: Input the procurement requirement text to be processed into the adjusted NLP model. The model first extracts the target shipment volume level, target shipment service type, and procurement cycle entities from the text through the named entity recognition task. Then, it labels the urgency and risk level corresponding to the requirement through the text classification task. The extracted entities are quantified. According to preset rules, the target shipment volume level is converted into a level quantification value, the procurement cycle is converted into a time range quantification value, and the target shipment service type is converted into a code value. The entity quantification values ​​and the labeled urgency and risk levels are combined to form a standardized requirement feature vector containing all information. The vector structure clearly corresponds to the quantification value of each entity and the requirement level label.

[0020] Preferably, step S3 further includes: S301, based on the maintenance history shipment volume labeling system constructed in S1, clarify the correspondence between each entity in the demand feature vector and the labeling system: the target shipment service type entity corresponds to the first-level label of the labeling system, the target shipment volume level entity corresponds to the second-level label of the labeling system, and the procurement cycle entity corresponds to the third-level label of the labeling system. S302: Retrieve supplier ID-structured tag data pairs from the supplier's historical shipment volume database. For each supplier, compare the target shipment volume level quantification value in the demand feature vector with the level parameter of the second-level tag in the supplier's tag system to confirm whether the shipment volume level matches. Then, compare the procurement cycle quantification value with the shipment start time and shipment end time sub-parameters of the third-level tag to confirm whether the validity period of the shipment time range covers the procurement cycle. Finally, compare the target shipment service type code value with the type parameter of the first-level tag to confirm whether the shipment service category is suitable. S303: For any supplier, if the target shipment volume level, target shipment service type, and procurement cycle entity in the demand feature vector do not have a matching tag in the supplier's structured tags, mark the unmatched demand feature as an item to be supplemented, and generate a supplementary reminder instruction by associating it with the corresponding supplier ID.

[0021] Preferably, step S4 further includes: S40 defines four matching dimensions: shipment volume level matching degree, shipment volume level quantitative value, shipment time coverage, and shipment service type adaptability. The scoring calculation logic for each dimension is clearly defined: the shipment volume level matching degree is based on the supplier's shipment volume level quantitative value and the demand target shipment volume level quantitative value. The ratio between the two is calculated. If the supplier's quantitative value is higher than the demand quantitative value, the dimension gets full marks. If the supplier has no corresponding shipment volume tag, it gets zero marks. The delivery time coverage is based on the supplier's delivery time range and the procurement cycle time range. The intersection duration of the two time ranges is calculated, and the ratio of the intersection duration to the total procurement cycle duration is used as the score for this dimension. The service type matching score is calculated based on the supplier's service type code value and the target service type code value. The ratio between the two is calculated. If the supplier's service type code value and the target service type code value are the same, the score is full; if they are not the same, the score is zero. The matching degree of the customer evaluation of the shipment is based on the evaluation score and shipment frequency of the supplier's shipment customers within the supplier's historical preset period. The shipment frequency is converted into a weight coefficient according to the preset rules, and the product of the evaluation score and the weight coefficient is used as the score of that dimension. All dimension scores are unified to the same score range. S402, based on the urgency and risk level of the demand as labeled by the demand feature vector, clarifies that all combinations of the two are categorized into four types: high urgency - high risk level, high urgency - low risk level, low urgency - high risk level, and low urgency - low risk level. Historical project data matching the current maintenance demand urgency-risk level combination is selected from the historical procurement project service performance database. This database stores the company's past maintenance procurement project performance records, with fields including historical procurement project ID, corresponding supplier ID, demand urgency-risk level combination, supplier four-dimensional score, total matching score, on-time performance rate, service process problem incidence rate, and project acceptance pass rate. A multiple linear regression model is constructed using the supplier's overall actual service performance score from historical projects as the dependent variable Y, and the four-dimensional scores as independent variables X1, X2, X3, and X4. X1 corresponds to the shipment volume matching score, X2 to the shipment time coverage score, X3 to the shipment service type suitability score, and X4 to the shipment customer evaluation matching score. Then, the multi-dimensional actual service performance indicators are normalized into a unified score, as shown in the following formula: Y = (On-time performance rate × a + Acceptance rate × b + (1 - Service process problem occurrence rate) × c) × 100; Among them, a, b, and c are preset weight coefficients, which are set according to the enterprise's focus on timely delivery, acceptance, and process compliance in the maintenance and procurement scenario, and satisfy a+b+c=1; The dimensional contribution coefficients are calculated using the least squares method of a multiple linear regression model. The calculation process is as follows: using the independent variable data X1, X2, X3, X4 and the dependent variable data Y of historical projects as samples, minimize the model error term. The formula is as follows: ,in, The constant term in the model is used to fit the historical data using a formula, resulting in the final model. These represent the magnitude of change in the dependent variable Y when X1, X2, X3, and X4 change by 1 unit, respectively. Finally, the contribution coefficient is standardized to the dimensional weight percentage. The absolute value of is converted into a percentage form, as shown in the following formula: ; For each combination, a dimensional priority strategy is adopted, which prioritizes high-urgency needs to ensure delivery time coverage, high-risk needs to ensure delivery volume level matching and delivery service type adaptability, and low-urgency, low-risk needs to ensure delivery customer evaluation matching. Then, the dimensional weight ratios are adjusted based on the calculated dimensional weight ratios: for the highest priority dimension, a preset percentage increase is added to its corresponding base, and the same increase is deducted from the base of the lowest priority dimension. Finally, the weight ratios of delivery volume level matching W1, delivery time coverage W2, delivery service type adaptability W3, and delivery customer evaluation matching W4 are set. S403 uses a weighted summation formula to calculate the total matching score for each supplier. The formula is as follows: ; Where S total S1 represents the supplier's total matching score, S2 represents the shipping grade matching score, S3 represents the shipping time coverage score, S4 represents the shipping service type suitability score, and S5 represents the shipping customer evaluation matching score.

[0022] Preferably, step S5 further includes: S501: Based on the urgency and risk level of the demand marked by the demand feature vector, a unique scenario-based matching threshold is set for each combination type, and the thresholds corresponding to different combination types do not overlap; the correspondence between combination type and scenario-based matching threshold is stored in the system threshold database. S502, retrieve the total matching score S of all suppliers from the calculation results of S403. total Simultaneously, the system retrieves the scenario-based matching threshold corresponding to the current maintenance requirement from the system threshold database; for each supplier, the total matching score is compared with the scenario-based matching threshold, and two types of results are recorded: one type is suppliers with a total matching score higher than the scenario-based matching threshold, and the other type is suppliers with a total matching score lower than the scenario-based matching threshold. The two types of results are associated with the corresponding supplier IDs. S503: Suppliers with total matching scores higher than the scenario-based matching threshold are sorted in descending order of total matching score. In the sorted supplier list, basic information and matching scores for each dimension of each supplier are added to form a structured list of compliant suppliers. The basic information of the suppliers includes the supplier name and contact information. All data in the list is bound to the supplier ID.

[0023] Preferably, S6 further includes: S601 sets a passing score for each matching dimension, which is consistent with the score range for each dimension. The scores for each non-compliant supplier are evaluated sequentially based on the following order: shipment volume matching degree, shipment time coverage, shipment service type suitability, and shipment customer evaluation matching degree. If the shipment volume matching degree score is below the passing score for that dimension, the non-compliance is marked as insufficient shipment volume. If the shipment volume matching degree score meets the standard, but the shipment time coverage score is below the passing score, the non-compliance is marked as insufficient shipment time coverage. If the first two scores meet the standard, but the shipment service type suitability score is below the passing score, the non-compliance is marked as a mismatch in shipment service type. If the first three scores meet the standard, but the total matching score is still below the scenario-based matching threshold, the non-compliance is marked as insufficient comprehensive score across multiple dimensions. S602 integrates the supplier ID, supplier name, contact information, scores for each dimension, and the marked reasons for non-compliance for each non-compliant supplier to form a structured list of non-compliant suppliers. The list must clearly define the correspondence between each field, and the non-compliance reason field must be directly related to the dimension score used for judgment. Finally, the list is stored in the system results database.

[0024] Preferably, a system for processing procurement plan reports for maintenance services is provided, which is used to execute the processing method for the procurement plan reports for maintenance services. The system includes a supplier historical shipment processing and storage module, a procurement demand parsing module, a feature-tag mapping and supplementary entry reminder module, a multi-dimensional matching score calculation module, a compliant supplier screening module, and a non-compliant supplier labeling module. The supplier historical shipment volume processing and storage module is used to process the historical shipment volume data submitted by suppliers, extract the shipment service type, shipment volume level, shipment time range, and shipment customer evaluation fields, and assign and store them according to the historical shipment volume tag system. The procurement demand parsing module is used to extract the target shipment volume level, target shipment service type, and procurement cycle entity from the procurement demand text based on the adjusted NLP model, mark the urgency and risk level of the demand, and generate a standardized demand feature vector containing the quantitative value of the entity. The feature-tag mapping and supplementary data reminder module is responsible for matching entities in the demand feature vector with structured tags in the supplier's historical shipment volume tag library, marking demand features that do not find matching tags as items to be supplemented, and triggering a supplementary data reminder for the supplier's historical shipment volume. The multi-dimensional matching score calculation module is used to define four-dimensional matching standards and calculate the scores of each dimension. It calls the data of the historical procurement project service performance database, calculates the contribution coefficient of each dimension to the actual service performance through a multiple linear regression model and standardizes it into the weight ratio base. It adjusts the weight ratio by combining the urgency of the demand, the risk level and the preset dimension priority strategy, and finally calculates the total matching score of the supplier through a weighted summation formula. The compliant supplier screening module is responsible for setting scenario-based matching thresholds based on the urgency of the need and the risk level, comparing the total matching score of suppliers with the thresholds, retaining qualified suppliers and generating a list of compliant suppliers in descending order of scores. The non-compliant supplier labeling module is used to prioritize non-compliance based on shipment volume and combine the scores of each dimension to label the reasons for non-compliance, integrate supplier information to generate a list of non-compliant suppliers and store it.

[0025] Example 2: This embodiment takes a boiler maintenance service procurement project of a hydropower station as an example to fully implement the procurement plan report processing method and system for maintenance services described in this invention. All data and parameters are based on the assumptions of this procurement scenario. The specific content is as follows: I. Processing and Database Construction of Supplier Historical Shipment Data: This procurement process first processed the historical shipment data submitted by the four participating suppliers (Suppliers A, B, C, and D). Data sources included historical order records from 2022-2023 provided by each supplier, data from the industry transaction filing platform, and past procurement ledgers. Data collection technology was used to extract four core fields for each supplier: "shipment service type, shipment volume level, shipment time range, and customer reviews." For suppliers with missing fields, such as Supplier D which did not provide customer reviews, a reminder to submit supplementary data was triggered until the data was complete.

[0026] Based on the extracted data, a multi-level maintenance procurement historical shipment volume labeling system is constructed: Level 1 labels are divided into "boiler maintenance shipment (code 001), pressure vessel maintenance shipment (code 002), and pipeline maintenance shipment (code 003)" according to the maintenance service type; Level 2 labels are set with grades and quantitative values ​​according to the total shipment volume range, specifically "Level 1 (1-10 times, quantitative value 1), Level 2 (11-20 times, quantitative value 2), Level 3 (21-30 times, quantitative value 3)"; Level 3 labels are shipment time parameters, including two sub-parameters: "shipment start time" and "shipment end time"; Level 4 labels are set according to customer evaluation score ranges, specifically "Excellent (85-100 points, quantitative value 3), Good (70-84 points, quantitative value 2), Average (60-69 points, quantitative value 1), Poor (<60 points, quantitative value 0)".

[0027] The complete fields of each supplier are matched and assigned values ​​according to the tag system: Supplier A's shipping service type is "boiler maintenance", corresponding to the first-level tag 001; the total number of shipments is 18, corresponding to the second-level tag level 2 (quantitative value 2); the shipping time range is from May 1, 2022 to May 31, 2024, corresponding to the third-level tag "start 2022.05.01-end 2024.05.31"; the average customer evaluation score is 85 points, corresponding to the fourth-level tag "excellent (quantitative value 3)". All tags are bound to Supplier A's ID (ID-A) and stored in the relational database. Similarly, Supplier B's secondary label is Grade 1 (quantitative value 1), the tertiary label is "2023.01.01-2024.03.31", and the quaternary label is "Good (quantitative value 2)"; Supplier C's secondary label is Grade 3 (quantitative value 3), the tertiary label is "2022.10.01-2024.10.31", and the quaternary label is "Excellent (quantitative value 3)"; after supplementing customer evaluation, Supplier D's secondary label is Grade 2 (quantitative value 2), the tertiary label is "2022.08.01-2024.04.30", and the quaternary label is "Medium (quantitative value 1)".

[0028] Simultaneously, a historical procurement project service performance database is constructed, storing boiler overhaul procurement project data from the past 3 years. Core fields are as follows: Project ID (XM2022001), corresponding supplier ID (ID-A), demand urgency-risk level combination (high-high), supplier four-dimensional score (shipment volume matching 100, delivery time coverage 90, service type suitability 100, customer evaluation matching 85), total matching score 92, on-time performance rate 95%, service process problem occurrence rate 2%, and project acceptance pass rate 98%. All data originates from past project performance reports and acceptance records and is updated in real time.

[0029] II. Procurement Demand Analysis and Feature Vector Generation: We collected 500 texts of procurement requirements for maintenance services of hydropower station boilers and pressure vessels from 2021 to 2023. Each text was annotated in two dimensions: entity annotation extracted "target shipment volume level, target shipment service type, and procurement cycle", such as annotating "15 boiler maintenance shipments required" with "15 shipments (target shipment volume level) and boiler maintenance (service type)"; category annotation included "urgency and risk level", such as annotating "pre-flood season maintenance" with "high urgency and high risk level". The annotated texts were divided into a training set (350 texts) and a validation set (150 texts) in a 7:3 ratio to form a corpus of historical shipment volumes for maintenance procurement.

[0030] The BERT pre-trained NLP model was selected as the base model, and the model was adjusted using the training set: the named entity recognition task was set to be trained for "target shipment volume level, target shipment service type, and procurement cycle", and the text classification task was trained for "urgency and risk level". The performance was monitored through the validation set during training. When the entity recognition accuracy reached 96% and the classification accuracy reached 94%, the training was stopped, and a special NLP model adapted to the maintenance and procurement scenario was obtained.

[0031] The procurement requirement text, "Procurement of boiler maintenance services for hydropower stations in 2024, requiring suppliers to have provided similar services in the past two years, with at least 15 shipments of similar services in the past, and the procurement period is from June 1, 2024 to August 31, 2024. Due to the approaching flood season, the demand is highly urgent, and the equipment safety associated risk level is high," is input into the adjusted model. Named entity recognition is used to extract "target shipment volume level (≥15 times), target shipment service type (boiler maintenance), and procurement period (June 1, 2024 - August 31, 2024)". Text classification is used to label it as "high urgency" and "high risk level". The extracted entities are quantified: the target shipment volume level ≥15 times corresponds to a secondary label level 2, with a quantified value of 2; the procurement period is converted into a time range quantified value of "starting June 1, 2024 - ending August 31, 2024"; the target shipment service type "boiler maintenance" corresponds to the primary label code 001. The entity quantification values ​​(2, 001, 2024.06.01-2024.08.31) are integrated with "high urgency and high risk level" to form a standardized demand feature vector: [Target shipment volume level quantification value: 2, Target shipment service type code: 001, Procurement cycle quantification value: 2024.06.01-2024.08.31, Demand urgency: high, Demand risk level: high].

[0032] III. Demand Characteristics and Supplier Tag Mapping: Based on the historical shipment volume labeling system constructed by S1, the correspondence between demand feature vectors and labels is clarified: the target shipment service type code (001) corresponds to the first-level label, the target shipment volume level quantification value (2) corresponds to the second-level label, and the procurement cycle quantification value (2024.06.01-2024.08.31) corresponds to the third-level label.

[0033] The system retrieves ID-structured tag data pairs from the supplier's historical shipment volume database for four suppliers and compares them one by one: For supplier A, the target shipment volume level quantification value 2 is compared with the secondary tag quantification value 2, and the condition "demand quantification value ≤ supplier quantification value" is met, thus determining "shipment volume level match"; the procurement period 2024.06.01-2024.08.31 is compared with the tertiary tag "2022.05.01-2024.05.31", and the supplier's shipment time does not cover the procurement period, thus determining "shipment time range not covered"; the target shipment service type code 001 is compared with the primary tag code 001, thus determining "shipment service category match". For Supplier B, the target shipment volume level quantification value 2 is compared with the secondary label quantification value 1, and the requirement of "demand ≤ supplier" is not met, so it is determined that "shipment volume level does not match". The procurement cycle is compared with the tertiary label "2023.01.01-2024.03.31", and it is not covered, so it is determined that "time is not covered". The service type code 001 matches, so it is determined that "service category is suitable". For Supplier C, the target shipment volume level quantification value 2 is compared with the secondary label quantification value 3, and they match. The procurement cycle is compared with the tertiary label "2022.10.01-2024.10.31", and it covers, so it is determined that "time covers". The service type code 001 matches, so it is determined that "service category is suitable". For supplier D, the target shipment volume level quantification value 2 is compared with the secondary label quantification value 2 and they match; the procurement cycle is compared with the tertiary label "2022.08.01-2024.04.30" and is not covered, so it is determined that "the time is not covered"; the service type code 002 is compared with the demand 001 and it is determined that "the service category is not suitable".

[0034] For unmatched demand characteristics, mark them as items to be supplemented: Supplier B "shipment volume level mismatch", triggering a supplementary record reminder, requiring the supplementation of boiler maintenance shipment records from 2022 to 2024 to improve the level; Supplier D "service category mismatch", triggering a supplementary record reminder, requiring the supplementation of boiler maintenance related shipment data; Suppliers A and D "time not covered", triggering a supplementary record reminder, requiring confirmation of whether there is proof of boiler maintenance service capability from June to August 2024.

[0035] IV. Multi-dimensional score calculation and total matching score: (1) Calculation of scores for each dimension: We define four dimensions: matching degree of shipment volume level, coverage of shipment time, matching degree of shipment service type, and matching degree of shipment customer evaluation. The scores for each dimension are uniformly set from 0 to 100. Shipment volume level matching degree: quantified by supplier shipment volume level (S) quant ) and the quantitative value of demand target (D) quant The ratio of ) is calculated, and a ratio ≥ 1 earns 100 points, otherwise (S) is not awarded. quant / D quant(100 points) No label gets 0 points. Supplier A: 2 / 2 = 1 → 100 points; Supplier B: 1 / 2 = 0.5 → 50 points; Supplier C: 3 / 2 = 1.5 → 100 points; Supplier D: 2 / 2 = 1 → 100 points.

[0036] Delivery time coverage: Calculated as the intersection of the supplier's delivery time range and the procurement cycle / total procurement cycle length × 100. The total procurement cycle length is 92 days (June 1, 2024 - August 31, 2024). Supplier A: 0 days → 0 minutes; Supplier B: 0 days → 0 minutes; Supplier C: 92 days → 92 minutes; Supplier D: 0 days → 0 minutes.

[0037] Service type matching: 100 points for supplier service type code matching demand code, otherwise 0 points. Suppliers A, B, and C are 001 → 100 points; Supplier D is 002 → 0 points.

[0038] Customer evaluation matching degree: calculated by multiplying the supplier's customer evaluation quantification value by the shipping frequency weighting coefficient. The coefficient is 0.8 for shipping frequency 1-10 times, 0.9 for 11-20 times, and 1.0 for more than 21 times. Supplier A's evaluation quantification value is 3, frequency is 18 times → 3 × 0.9 × (100 / 3) = 90 points (normalized to a maximum score of 100); Supplier B's evaluation quantification value is 2, frequency is 8 times → 2 × 0.8 × (100 / 3) ≈ 53.3 points; Supplier C's evaluation quantification value is 3, frequency is 25 times → 3 × 1.0 × (100 / 3) = 100 points; Supplier D's evaluation quantification value is 1, frequency is 16 times → 1 × 0.9 × (100 / 3) = 30 points.

[0039] (2) Dynamic allocation of dimensional weights: Based on the demand combination of "high urgency - high risk level", 10 past project data for this combination were selected from the historical procurement project service performance database. A multiple linear regression model was constructed with "actual service performance comprehensive score Y" as the dependent variable and the four-dimensional scores (X1-X4) as independent variables: Y calculation: Y = (on-time performance rate × 0.4 + acceptance pass rate × 0.4 + (1 - service process problem occurrence rate) × 0.2) × 100. For example, if a project has an on-time performance rate of 95%, an acceptance pass rate of 98%, and a problem occurrence rate of 2%, then Y = (95 × 0.4 + 98 × 0.4 + 98 × 0.2) × 100 = 96.4 points.

[0040] Regression calculation: Minimize the error term using the least squares method: ; get .

[0041] Weight cardinality standardization: Calculated .

[0042] Weighting adjustments: Prioritize ensuring delivery time coverage for "high urgency - high risk level" and add 3% to W2'. Subtract 3% from W4', which has the lowest priority. Final weights: W1=25%, W2=45%, W3=33%, W4=7%.

[0043] (3) Calculation of total matching score: According to formula S total Calculate S1×W1+S2×W2+S3×W3+S4×W4: Supplier A: 100×0.25+0×0.45+100×0.33+90×0.07=25+0+33+6.3=64.3 points; Supplier B: 50×0.25+0×0.45+100×0.33+53.3×0.07≈12.5+0+33+3.73=49.23 points; Supplier C: 100×0.25+92×0.45+100×0.33+100×0.07=25+41.4+33+7=106.4 points; Supplier D: 100×0.25+0×0.45+0×0.33+30×0.07=25+0+0+2.1=27.1 points.

[0044] V. Selection of Compliant Suppliers: Based on the demand for "high urgency - high risk level" combination, a scenario-based matching threshold of 70 points is set (the threshold varies for different combinations, such as 50 points for the "low-low" combination), and the combination-threshold correspondence (high-high: 70) is stored in the system threshold database.

[0045] The total scores of four suppliers were retrieved from the S4 results and compared with the threshold of 70: Supplier C (106.4 points) ≥ 70, while Suppliers A (64.3 points), B (49.23 points), and D (27.1 points) < 70. Supplier C, which met the criteria, was sorted in descending order of score (since only one supplier met the criteria, it was listed first). Its basic information was supplemented, including name and contact information, and scores for each dimension (S1=100, S2=92, S3=100, S4=100), forming a list of compliant suppliers. All data in the list was stored bound to Supplier C's ID (ID-C).

[0046] VI. Non-compliant supplier labeling: Preset pass / fail scores for each dimension: Shipment volume level matching score ≥ 60 points, shipment time coverage score ≥ 50 points, shipment service type suitability score ≥ 80 points, and shipment customer evaluation matching score ≥ 40 points. Non-compliance reasons are determined based on the priority of "shipment volume level matching score → shipment time coverage score → shipment service type suitability score → shipment customer evaluation matching score". Supplier A: Shipment volume matching score 100 (meets the standard), shipment time coverage score 0 (<50, does not meet the standard) → marked "Insufficient shipment time coverage"; Supplier B: Shipment volume level matching score 50 (<60, not up to standard) → marked "Insufficient shipment volume level"; Supplier D: Shipment volume matching score 100 (meets the standard), shipment time coverage score 0 (does not meet the standard), but the preceding dimension is judged first. Because the shipment service type matching score is 0 (<80, judged after the time dimension) → marked "shipment service type mismatch".

[0047] The IDs, names, contact information, scores for each dimension, and reasons for non-compliance of the three non-compliant suppliers are integrated to form a list of non-compliant suppliers. For example, supplier A (ID-A, a certain repair shop, telephone number, S1=100, S2=0, S3=100, S4=90, reason: insufficient delivery time coverage) is associated with the "reason for non-compliance" in the list and the corresponding dimension score (S2=0), and stored in the system results database.

[0048] VII. System Module Application: In this procurement, the supplier historical shipment processing and storage module processed the shipment data of four suppliers, constructing a tagging system and a historical service performance database; the procurement demand parsing module generated standardized demand feature vectors through a dedicated NLP model; the feature-tag mapping and supplementary entry reminder module compared the demand with the tags, triggering three supplementary entry reminders; the multi-dimensional matching score calculation module calculated the scores for each dimension and the total score; the compliant supplier screening module generated a compliant list containing only supplier C; and the non-compliant supplier labeling module generated a non-compliant list of three suppliers. The entire process was achieved through the collaborative operation of system modules, requiring no manual intervention, thus improving procurement efficiency and accuracy.

[0049] Example 3: like Figure 2 As shown, embodiments of the present invention also provide a computer device. Figure 2Taking a single processor 10 as an example, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other using different buses and can be mounted on a common motherboard or otherwise installed as needed. The processor can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices, such as display devices coupled to the interface. In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations, for example, as a server array, a group of blade servers, or a multiprocessor system.

[0050] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0051] The memory 20 stores instructions executable by at least one processor 10 to cause the at least one processor 10 to perform the method shown in the above embodiments.

[0052] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0053] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0054] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0055] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

Claims

1. A method for processing procurement plan reports for maintenance services, characterized in that, Includes the following steps: S1. Identify and process historical shipment data, extract relevant fields, and then construct a historical shipment label system for maintenance procurement. Based on the historical shipment label system for maintenance procurement, assign structured labels to the fields and store them. Construct a historical procurement project service performance database that is associated with the supplier's historical shipment database and the system threshold database. S2, construct a corpus of historical shipments for maintenance procurement, train an NLP model based on the corpus of historical shipments for maintenance procurement; then input the procurement demand text into the trained NLP model, extract target entities through named entity recognition, label the urgency and risk level of the demand, and output a standardized demand feature vector containing the quantitative value of the entity. S3, based on the historical shipment volume tagging system, maps the target shipment volume level, target shipment service type, and procurement cycle in the demand feature vector to the structured tags in the supplier's historical shipment volume database. If there is no match, it marks the item to be supplemented and triggers a reminder to supplement the supplier's historical shipment volume data. S4 defines four dimensions: matching degree of shipment volume level, matching degree of shipment time, matching degree of shipment service type, and matching degree of shipment customer evaluation. The scores of each dimension are calculated, and then the weights of the dimensions are dynamically allocated by combining the urgency of the demand and the service performance data of historical procurement projects. The total matching score of the supplier is calculated. S5 sets corresponding scenario-based matching thresholds based on the urgency of demand and risk level, compares the total matching score of suppliers with the threshold, retains suppliers whose total matching score is not lower than the threshold, and generates a list of compliant suppliers in descending order of total score. S6: For suppliers whose total matching score is below the threshold, the non-compliance reasons are marked according to the matching degree of shipment volume level, the coverage of shipment time, the adaptability of shipment service type, and the matching degree of shipment customer evaluation, and a list of non-compliant suppliers is generated.

2. The method for processing a maintenance service procurement plan report according to claim 1, characterized in that, Step S1 further includes: S101 uses data collection technology to process the historical shipment data submitted by suppliers. The relevant data includes historical order records, transaction filing data, and purchase ledgers. Then, it extracts four fields from the data: shipment service type, shipment volume level, shipment time range, and shipment customer evaluation. If any fields are missing, it triggers a reminder for the supplier to submit supplementary data. S102, construct a multi-level maintenance procurement historical shipment volume labeling system, including four levels of labels: Primary tag: Shipping service category, divided according to maintenance service type; Secondary label: Shipment volume level, set according to the total shipment range; Level 3 tag: Shipping time parameter, which includes two sub-parameters: shipping start time and shipping end time; Level 4 tag: Customer reviews of shipped products, set according to preset review score ranges; S103 will extract the complete historical shipment volume field and match and assign values ​​with the maintenance procurement historical shipment volume tag system: the shipment service type corresponds to the first-level tag, the shipment volume level corresponds to the second-level tag, the shipment time range corresponds to the third-level tag, and the shipment customer evaluation corresponds to the fourth-level tag. All tags are bound to the supplier ID and stored in the relational database. A historical procurement project service performance database is constructed to store the performance records of past maintenance procurement projects. This database is linked to the supplier's historical shipment volume database and the system threshold database, and stores the following fields: Historical procurement project ID, corresponding supplier ID, demand urgency-risk level combination, supplier four-dimensional score, total supplier matching score, on-time performance rate, service process problem occurrence rate and project acceptance pass rate; The database is composed of performance reports, acceptance records, and service quality assessment documents from the company's past maintenance and procurement projects, and is updated in real time.

3. The method for processing a maintenance service procurement plan report according to claim 1, characterized in that, Step S2 further includes: S201, Collect historical maintenance service procurement requirement texts, and annotate each text segment in two dimensions: The first dimension is entity labeling, which includes three types of entities in the labeled text: target shipment volume level, target shipment service type, and procurement cycle. The second dimension is classification labeling, which includes two types of tags: demand urgency and risk level. The labeling is based on the demand characteristic rules in the maintenance procurement scenario. The labeled text is divided into training set and validation set according to a preset ratio to form a corpus of historical shipment volume in maintenance procurement. S202, a pre-trained NLP model was selected as the base model, and the model was adjusted using the training set of the maintenance and procurement corpus: The model training task is set as a dual task of named entity recognition and text classification. The named entity recognition task is trained to identify three types of entities: target shipment volume level, target shipment service type, and procurement cycle. The text classification task is trained to classify two types of tags: demand urgency and risk level. During the training process, the model performance is monitored in real time using a validation set. When the entity recognition accuracy and classification accuracy of the model on the validation set both reach the preset stable value, the training is stopped, the NLP model is adjusted, and an NLP model adapted to the maintenance and procurement scenario is obtained. S203: Input the procurement requirement text to be processed into the adjusted NLP model. The model first extracts the target shipment volume level, target shipment service type, and procurement cycle entities from the text through the named entity recognition task. Then, it labels the urgency and risk level corresponding to the requirement through the text classification task. The extracted entities are quantified. According to preset rules, the target shipment volume level is converted into a level quantification value, the procurement cycle is converted into a time range quantification value, and the target shipment service type is converted into a code value. The entity quantification values ​​and the labeled urgency and risk levels are combined to form a standardized requirement feature vector containing all information. The vector structure clearly corresponds to the quantification value of each entity and the requirement level label.

4. The method for processing a maintenance service procurement plan report according to claim 1, characterized in that, Step S3 further includes: S301, based on the maintenance history shipment volume labeling system constructed in S1, clarify the correspondence between each entity in the demand feature vector and the labeling system: the target shipment service type entity corresponds to the first-level label of the labeling system, the target shipment volume level entity corresponds to the second-level label of the labeling system, and the procurement cycle entity corresponds to the third-level label of the labeling system. S302: Retrieve supplier ID-structured tag data pairs from the supplier's historical shipment volume database. For each supplier, compare the target shipment volume level quantification value in the demand feature vector with the level parameter of the second-level tag in the supplier's tag system to confirm whether the shipment volume level matches. Then, compare the procurement cycle quantification value with the shipment start time and shipment end time sub-parameters of the third-level tag to confirm whether the validity period of the shipment time range covers the procurement cycle. Finally, compare the target shipment service type code value with the type parameter of the first-level tag to confirm whether the shipment service category is suitable. S303: For any supplier, if the target shipment volume level, target shipment service type, and procurement cycle entity in the demand feature vector do not have a matching tag in the supplier's structured tags, mark the unmatched demand feature as an item to be supplemented, and generate a supplementary reminder instruction by associating it with the corresponding supplier ID.

5. The method for processing a maintenance service procurement plan report according to claim 1, characterized in that, Step S4 further includes: S40 defines four matching dimensions: shipment volume level matching degree, shipment volume level quantitative value, shipment time coverage, and shipment service type adaptability. The scoring calculation logic for each dimension is clearly defined: the shipment volume level matching degree is based on the supplier's shipment volume level quantitative value and the demand target shipment volume level quantitative value. The ratio between the two is calculated. If the supplier's quantitative value is higher than the demand quantitative value, the dimension gets full marks. If the supplier has no corresponding shipment volume tag, it gets zero marks. The delivery time coverage is based on the supplier's delivery time range and the procurement cycle time range. The intersection duration of the two time ranges is calculated, and the ratio of the intersection duration to the total procurement cycle duration is used as the score for this dimension. The service type matching score is calculated based on the supplier's service type code value and the target service type code value. The ratio between the two is calculated. If the supplier's service type code value and the target service type code value are the same, the score is full; if they are not the same, the score is zero. The matching degree of the customer evaluation of the shipment is based on the evaluation score and shipment frequency of the supplier's shipment customers within the supplier's historical preset period. The shipment frequency is converted into a weight coefficient according to the preset rules, and the product of the evaluation score and the weight coefficient is used as the score of that dimension. All dimension scores are unified to the same score range. S402, based on the urgency and risk level of the demand as labeled by the demand feature vector, clarifies that all combinations of the two are categorized into four types: high urgency - high risk level, high urgency - low risk level, low urgency - high risk level, and low urgency - low risk level. Historical project data matching the current maintenance demand urgency-risk level combination is selected from the historical procurement project service performance database. This database stores the company's past maintenance procurement project performance records, with fields including historical procurement project ID, corresponding supplier ID, demand urgency-risk level combination, supplier four-dimensional score, total matching score, on-time performance rate, service process problem incidence rate, and project acceptance pass rate. A multiple linear regression model is constructed using the supplier's overall actual service performance score from historical projects as the dependent variable Y, and the four-dimensional scores as independent variables X1, X2, X3, and X4. X1 corresponds to the shipment volume matching score, X2 to the shipment time coverage score, X3 to the shipment service type suitability score, and X4 to the shipment customer evaluation matching score. Then, the multi-dimensional actual service performance indicators are normalized into a unified score, as shown in the following formula: Y = (On-time performance rate × a + Acceptance rate × b + (1 - Service process problem occurrence rate) × c) × 100; Among them, a, b, and c are preset weight coefficients, which are set according to the enterprise's focus on timely delivery, acceptance, and process compliance in the maintenance and procurement scenario, and satisfy a+b+c=1; The dimensional contribution coefficients are calculated using the least squares method of a multiple linear regression model. The calculation process is as follows: using the independent variable data X1, X2, X3, X4 and the dependent variable data Y of historical projects as samples, minimize the model error term. The formula is as follows: ,in, The constant term in the model is used to fit the historical data using a formula, resulting in the final model. These represent the magnitude of change in the dependent variable Y when X1, X2, X3, and X4 change by 1 unit, respectively. Finally, the contribution coefficient is standardized to the dimensional weight percentage. The absolute value of is converted into a percentage form, as shown in the following formula: ； For each combination, a dimensional priority strategy is adopted, which prioritizes high-urgency needs to ensure delivery time coverage, high-risk needs to ensure delivery volume level matching and delivery service type adaptability, and low-urgency, low-risk needs to ensure delivery customer evaluation matching. Then, the dimensional weight ratios are adjusted based on the calculated dimensional weight ratios: for the highest priority dimension, a preset percentage increase is added to its corresponding base, and the same increase is deducted from the base of the lowest priority dimension. Finally, the weight ratios of delivery volume level matching W1, delivery time coverage W2, delivery service type adaptability W3, and delivery customer evaluation matching W4 are set. S403 uses a weighted summation formula to calculate the total matching score for each supplier. The formula is as follows: ； Where S total S1 represents the supplier's total matching score, S2 represents the shipping grade matching score, S3 represents the shipping time coverage score, S4 represents the shipping service type suitability score, and S5 represents the shipping customer evaluation matching score.

6. The method for processing a maintenance service procurement plan report according to claim 5, characterized in that, Step S5 further includes: S501: Based on the urgency and risk level of the demand marked by the demand feature vector, a unique scenario-based matching threshold is set for each combination type, and the thresholds corresponding to different combination types do not overlap; the correspondence between combination type and scenario-based matching threshold is stored in the system threshold database. S502, retrieve the total matching score S of all suppliers from the calculation results of S403. total At the same time, the system retrieves the scenario-based matching threshold corresponding to the current maintenance requirement from the system threshold database; for each supplier, the total matching score is compared with the scenario-based matching threshold, and two types of results are recorded: one type is suppliers with a total matching score higher than the scenario-based matching threshold, and the other type is suppliers with a total matching score lower than the scenario-based matching threshold. The two types of results are associated with the corresponding supplier IDs. S503: Suppliers with total matching scores higher than the scenario-based matching threshold are sorted in descending order of total matching score. In the sorted supplier list, basic information and matching scores for each dimension of each supplier are added to form a structured list of compliant suppliers. The basic information of the suppliers includes the supplier name and contact information. All data in the list is bound to the supplier ID.

7. The method for processing a maintenance service procurement plan report according to claim 6, characterized in that, S6 further includes: S601 sets a passing score for each matching dimension, which is consistent with the score range for each dimension. The scores for each non-compliant supplier are evaluated sequentially based on the following order: shipment volume matching degree, shipment time coverage, shipment service type suitability, and shipment customer evaluation matching degree. If the shipment volume matching degree score is below the passing score for that dimension, the non-compliance is marked as insufficient shipment volume. If the shipment volume matching degree score meets the standard, but the shipment time coverage score is below the passing score, the non-compliance is marked as insufficient shipment time coverage. If the first two scores meet the standard, but the shipment service type suitability score is below the passing score, the non-compliance is marked as a mismatch in shipment service type. If the first three scores meet the standard, but the total matching score is still below the scenario-based matching threshold, the non-compliance is marked as insufficient comprehensive score across multiple dimensions. S602 integrates the supplier ID, supplier name, contact information, scores for each dimension, and the marked reasons for non-compliance for each non-compliant supplier to form a structured list of non-compliant suppliers. The list must clearly define the correspondence between each field, and the non-compliance reason field must be directly related to the dimension score used for judgment. Finally, the list is stored in the system results database.

8. A system for processing procurement plan reports for maintenance services, used to execute the processing method for procurement plan reports for maintenance services as described in any one of claims 1-7, characterized in that, The system includes a module for processing and storing historical supplier shipments, a module for analyzing procurement needs, a module for feature-tag mapping and supplementary information reminders, a module for multi-dimensional matching score calculation, a module for screening compliant suppliers, and a module for labeling non-compliant suppliers. The supplier historical shipment volume processing and storage module is used to process the historical shipment volume data submitted by suppliers, extract the shipment service type, shipment volume level, shipment time range, and shipment customer evaluation fields, and assign and store them according to the historical shipment volume tag system. The procurement demand parsing module is used to extract the target shipment volume level, target shipment service type, and procurement cycle entity from the procurement demand text based on the adjusted NLP model, mark the urgency and risk level of the demand, and generate a standardized demand feature vector containing the quantitative value of the entity. The feature-tag mapping and supplementary data reminder module is responsible for matching entities in the demand feature vector with structured tags in the supplier's historical shipment volume tag library, marking demand features that do not find matching tags as items to be supplemented, and triggering a supplementary data reminder for the supplier's historical shipment volume. The multi-dimensional matching score calculation module is used to define four-dimensional matching standards and calculate the scores of each dimension. It calls the data of the historical procurement project service performance database, calculates the contribution coefficient of each dimension to the actual service performance through a multiple linear regression model and standardizes it into the weight ratio base. It adjusts the weight ratio by combining the urgency of the demand, the risk level and the preset dimension priority strategy, and finally calculates the total matching score of the supplier through a weighted summation formula. The compliant supplier screening module is responsible for setting scenario-based matching thresholds based on the urgency of the need and the risk level, comparing the total matching score of suppliers with the thresholds, retaining qualified suppliers and generating a list of compliant suppliers in descending order of scores. The non-compliant supplier labeling module is used to prioritize non-compliance based on shipment volume and combine the scores of each dimension to label the reasons for non-compliance, integrate supplier information to generate a list of non-compliant suppliers and store it.

9. A computer device, characterized in that, The system includes a memory and a processor, which are interconnected and communicate with each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the processing method for the maintenance service procurement plan report as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, A computer-readable storage medium stores computer instructions for causing a computer to execute a method for processing a procurement plan report for maintenance services as described in any one of claims 1 to 7.

Images