A bid evaluation method, device and medium based on electronic trading platform

Abstract

The present invention discloses a bid evaluation method, device and medium based on an electronic transaction platform, which belongs to the field of electronic bid evaluation technology; monitors the project progress in real time, determines whether the bidding project has reached the current progress, and automatically obtains the current project information if the project progress reaches the current node, enters the bidding document preparation and bid evaluation method setting; otherwise, the system continues to monitor the project progress; by introducing a conflict detection mechanism, it effectively solves the conflict problem that may exist in the information block processing process. First, the NLP technology is used to segment the information block, remove stop words and punctuation marks, obtain the text after segmentation, and obtain sentence vectors through the BERT model, extract named entities, and use syntactic analysis technology to build a grammatical tree of the sentence, thereby realizing semantic analysis of the information block. By calculating the semantic similarity between information blocks and setting a threshold to determine whether there is repeated or similar content, the conflict between information blocks can be effectively identified and resolved.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic bid evaluation, and in particular to a bid evaluation method, device and medium based on an electronic transaction platform. Background Art

[0002] When preparing engineering bidding documents, it is crucial to select the bid evaluation method and formulate the bid evaluation rules. It greatly affects the selection of the winning bidder and ultimately affects the winning bid price and project quality. In the process of formulating the bid evaluation rules, the evaluation factors and the weight of each factor are generally determined by a small number of staff of the tenderer or the tendering agent through "experience", which is highly subjective. Introducing fuzzy mathematics can greatly reduce the subjectivity of bid evaluation. The hierarchical analysis method makes the weight of the evaluation factors more reasonable and objective. [1]

[0003] The patent document with publication number CN116720773B discloses a bidding method and device based on block evaluation to obtain the bid documents submitted by the bidders and generate a bid number for the bid documents; perform text matching on the name mark and the end mark, and when the name mark and the end mark of each information block can be identified and matched from the bid document, execute the block step; if not, the bid document is invalidated; the block step: according to the part between the name mark and the end mark as an information block, the bid document is divided according to the information block, and the information block is sent to the audit platform corresponding to the type of information block according to the information block number to generate audit information; the audit information is sent to the tenderer for centralization; the final review step is executed; this application ensures that each information block of the bid document can be finally evaluated, and ensures the integrity of the information block content during the final review process, thereby achieving the reliability of the bid evaluation process.

[0004] However, the above reference documents have the following problems:

[0005] (1) The information block processing method mentioned in the above reference document may not be able to effectively resolve these conflicts when faced with mutually exclusive content or requirements. When faced with complex conflict situations, the current system can often only perform simple exclusion or manual intervention, and cannot automatically coordinate and resolve complex conflicts.

[0006] (2) The audit process usually relies on independent audits by third-party organizations. Although this model ensures the fairness and independence of the audit to a certain extent, it also has some significant problems and limitations. First, the involvement of third-party audit organizations increases the time and economic costs of the bid evaluation process. It may also introduce interference from external factors, affecting the efficiency and transparency of the audit.

[0007] (3)AHP [1]The assumption that each indicator is independent does not take into account the possible interactions and interdependencies between indicators in real situations. This limitation makes AHP unable to accurately reflect the true relationship between indicators in practical applications, especially in complex bid evaluation processes.

[0008] References

[0009] [1] Wei Jinchang, Research on Bid Evaluation Method Based on Fuzzy Mathematics and Hierarchy Analysis, Zhangzhou, Fujian 363105, China. Summary of the Invention

[0010] The purpose of the present invention is to provide a bid evaluation method, device and medium based on an electronic trading platform to solve the technical problems in existing solutions.

[0011] The purpose of the present invention can be achieved through the following technical solutions:

[0012] A bid evaluation method based on an electronic trading platform comprises the following steps:

[0013] Monitor the project progress in real time to determine whether the bidding project has reached the current progress. If the project progress reaches the current node, the system will automatically obtain the current project information and enter the bidding document preparation and bid evaluation method settings; otherwise, the system will continue to monitor the project progress.

[0014] Add preliminary review steps based on the project bidding documents, set review content and review criteria according to qualification review, formal review, and responsiveness review, and determine whether the content and review criteria are correct after completing the settings and submitting them;

[0015] Add detailed review steps based on the project bidding documents, set the review content, review criteria, score range and summary method according to the commercial review and technical review respectively, and after completing the settings and submitting them, check whether the set scores, weights, summary scores and review criteria are correct;

[0016] Set the price score calculation range and rules, benchmark value setting rules, and price calculation algorithm formula. After completing the settings, the system will automatically detect and provide feedback;

[0017] After the bid evaluation method is set up, the system will package, sign, and seal the entire bidding document. After the self-check content format is qualified, it will be encrypted with a digital certificate and transmitted to the cloud server in a unified format.

[0018] After the bid opening, the bidder decrypts the bid documents remotely. After all documents are successfully decrypted, the server automatically issues the bid documents for evaluation according to the bid evaluation method in the bidding documents.

[0019] Optionally, the preliminary review step includes the following steps:

[0020] Set review content and standards according to qualification review, form review, and responsiveness review;

[0021] The system automatically checks whether the content and evaluation criteria are correct. If the detection fails, go back to modify and submit again until it succeeds.

[0022] Optionally, the quotation review step includes the following steps:

[0023] According to the bidding documents, add quotation review steps, set the price score calculation range and rules, benchmark value setting rules, and price calculation algorithm formula for this project;

[0024] The calculation range sets the bidders who have passed the preliminary review as valid bidders, and the bid evaluation benchmark price is the arithmetic mean of the valid bid quotations. When the number of valid bidders is greater than five, the highest and lowest bids are removed when calculating the bid evaluation benchmark price; when the number of valid bidders is less than or equal to five, the highest and lowest bids are not removed when calculating the bid evaluation benchmark price.

[0025] The calculation formula is:

[0026] Evaluation benchmark price A = (A1 + A2 + ... + An) / n, where n is the valid number and An is a valid bid price;

[0027] The formula for calculating the deviation rate of bid quotations is: Deviation rate = 100% × (bid quotations of valid bidders - bid evaluation benchmark price) / bid evaluation benchmark price;

[0028] The formula for calculating the bid quotation score is: bid quotation score = F - absolute value of deviation rate × 100 × E, where F is the weight of the bid quotation, F = 70 points, and E is the deduction value for each percentage point above or below the bid evaluation benchmark price, E = 2 points, and the deduction shall not exceed 10 points;

[0029] After the settings are completed and submitted, determine whether the set score, calculation algorithm, and evaluation criteria are correct. If the system setting detection is successful, proceed to the next step; otherwise, return to modify and submit again for verification until success.

[0030] Optionally, the detailed review step includes the following steps:

[0031] Set review content, review standards, score range and summary method according to business review and technical review respectively;

[0032] The system automatically checks whether the set scores, weights, summary scores and evaluation criteria are correct. If the detection fails, return to modify and submit again until success.

[0033] Optionally, after the bid evaluation method is set up, the system will package, sign, and seal the entire content of the bidding documents. After self-checking and confirming that the content format is qualified, it will be encrypted with a digital certificate to form a unified format file and transmitted to the cloud server.

[0034] Optionally, the following steps need to be performed after the bid opening:

[0035] After passing the identity verification of the multi-factor user authentication system, the login control client automatically obtains the projects to be reviewed;

[0036] After selecting and entering a project, preliminary and detailed reviews will be conducted on the mobile review page. The review results of each review step will be stored in the storage space allocated by the server, which will be used for data storage of this review project and summary calculation according to the rules set in the bid evaluation method.

[0037] The quotation review automatically collects data based on the previous review results, calculates scores based on the quotation algorithm, and generates a quotation score.

[0038] Optionally, after all review steps are completed, the system will summarize the review data according to the rules and generate an evaluation report, which will be electronically signed using the provided signature device.

[0039] A bid evaluation device based on an electronic trading platform, comprising:

[0040] Monitor the project progress in real time to determine whether the bidding project has reached the current progress. If the project progress reaches the current node, the system will automatically obtain the current project information and enter the bidding document preparation and bid evaluation method settings; otherwise, the system will continue to monitor the project progress.

[0041] The preliminary review module adds preliminary review steps based on the project bidding documents, sets review content and review criteria according to qualification review, form review, and responsiveness review, and determines whether the content and review criteria are correct after the settings are completed and submitted;

[0042] The detailed review module adds detailed review steps based on the project bidding documents, sets the review content, review criteria, score range and summary method according to the commercial review and technical review respectively, and determines whether the set scores, weights, summary scores and review criteria are correct after the settings are submitted;

[0043] Setting module, set the price score calculation range and rules, benchmark value setting rules, and price calculation algorithm formula. After the setting is completed, the system will automatically detect and feedback;

[0044] Transmission module: After the bid evaluation method is set up, the system will package, sign, and seal the entire bidding document. After self-checking and confirming that the content format is qualified, it will be encrypted with a digital certificate and transmitted to the cloud server in a unified format;

[0045] In the review module, after the bid opening, the bidder decrypts the bid documents remotely. After all documents are successfully decrypted, the server automatically issues the bid documents for review according to the bid evaluation method in the bidding documents.

[0046] A storage medium comprising at least one processor;

[0047] and a memory communicatively coupled to the at least one processor;

[0048] The memory stores a computer program that can be executed by at least one processor, and the computer program is executed by at least one processor so that the at least one processor can execute the above-mentioned bid evaluation method and device based on the electronic trading platform.

[0049] Compared with the existing solutions, the present invention achieves the following beneficial effects:

[0050] One aspect of the present invention is to effectively solve the conflict problems that may exist in the process of information block processing by introducing a conflict detection mechanism. First, NLP technology is used to segment the information block, remove stop words and punctuation marks, obtain the text after segmentation, and obtain sentence vectors through the BERT model, extract named entities, and use syntactic analysis technology to build a grammatical tree of the sentence, thereby realizing semantic analysis of the information block. By calculating the semantic similarity between information blocks and setting a threshold to determine whether there is duplicate or similar content, the conflicts between information blocks can be effectively identified and resolved. Secondly, automatic adjustment and correction are performed according to preset rules, and the processed information blocks are re-verified in real time to ensure that conflicts are correctly handled, thereby improving the accuracy of information block processing.

[0051] Another aspect of the present invention is that, by employing an automated audit system, independent audits by third-party organizations are no longer required. This significantly reduces the time and financial costs of the bid evaluation process, improving its efficiency and economic benefits. Furthermore, by eliminating reliance on external organizations, interference from external factors can be avoided, thereby enhancing the transparency and fairness of the audit process. Furthermore, the automated audit system can monitor and address various issues encountered during the bid evaluation process in real time, ensuring the accuracy and reliability of the audit and further improving the quality and credibility of the bid evaluation process.

[0052] Other aspects of the present invention utilize a comprehensive bid evaluation method, including AHP, ANP, and multi-level regression analysis, to comprehensively consider the potential interactions and interdependencies between bid evaluation indicators, thereby improving the scientific nature and reliability of the bid evaluation results. First, by constructing a hierarchical structure using the analytic hierarchy process (AHP), classifying bid evaluation objectives, criteria, and specific bid evaluation schemes, and establishing a judgment matrix, the relative importance of each indicator is effectively captured, reflecting the interactions between indicators. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The present invention will be further described below with reference to the accompanying drawings.

[0054] Figure 1 This is a flowchart of a bid evaluation method, device, and medium based on an electronic trading platform of the present invention.

[0055] Figure 2 A schematic diagram of the structure of a computer device for implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0057] like Figure 1 As shown, the present invention is a bid evaluation method based on an electronic trading platform, comprising the following steps:

[0058] Monitor the project progress in real time to determine whether the bidding project has reached the current progress. If the project progress reaches the current node, the system will automatically obtain the current project information and enter the bidding document preparation and bid evaluation method settings; otherwise, the system will continue to monitor the project progress.

[0059] By monitoring project progress in real time, the system automatically retrieves relevant information based on the current project stage and takes subsequent actions, such as preparing bidding documents and setting bid evaluation methods. This saves human resources and time, while also reducing the potential for human error. Furthermore, the system continuously monitors project progress, ensuring a timely and consistent process, making the bidding process smoother and more efficient. In summary, this feature helps improve the efficiency and quality of bidding management, providing participants with more convenient and reliable bidding services.

[0060] Automatically obtain detailed information of the current project and automatically generate bidding documents based on project characteristics;

[0061] Add preliminary review steps based on the project bidding documents, set review content and review criteria according to qualification review, formal review, and responsiveness review, and determine whether the content and review criteria are correct after completing the settings and submitting them;

[0062] Set review content and standards according to qualification review, form review, and responsiveness review;

[0063] The system automatically checks whether the content and review criteria are correct. If the check fails, go back to modify and submit again until it succeeds;

[0064] Add detailed review steps based on the project bidding documents, set the review content, review criteria, score range and summary method according to the commercial review and technical review respectively, and after completing the settings and submitting them, check whether the set scores, weights, summary scores and review criteria are correct;

[0065] Obtain the bid document submitted by the bidder and generate a bid document number. Split the bid document into a specific number of information blocks, set a name tag at the beginning of each information block and an end tag at the end, and perform text matching on the name tags and end tags. If the name tags and end tags of each information block can be identified and matched, the block division step is executed; otherwise, the bid document is invalidated.

[0066] The information block is sent to the corresponding review platform according to the information block number, and the review end generates review information, which includes the score of the information block and corresponds to the information block one by one through the identifier;

[0067] The basic weights were determined using the AHP, and the weights were adjusted using the ANP. Multi-level regression analysis was then used to quantify the mutual influences among the indicators. First, the basic weights were determined using the AHP, a commonly used weight determination method that allows for a hierarchical assessment of the relative importance of each indicator based on opinions and criteria, ensuring that all factors are considered in the decision. Second, the weights were adjusted using the ANP, which considers the interdependencies among indicators, enabling a more comprehensive assessment of their importance and improving the accuracy and reliability of the decision weights. Finally, the mutual influences among the indicators were quantified using multi-level regression analysis. This provides a deeper understanding of the relationships between indicators, helping decision makers more accurately predict the impact of changes in each indicator on the decision outcome, thereby optimizing the decision-making process.

[0068] Set the price score calculation range and rules, benchmark value setting rules, and price calculation algorithm formula. After completing the settings, the system will automatically detect and provide feedback;

[0069] After the bid evaluation method is set up, the system will package, sign, and seal the entire bidding document. After the self-check content format is qualified, it will be encrypted with a digital certificate and transmitted to the cloud server in a unified format.

[0070] After the bid opening, the bidder decrypts the bid documents remotely. After all documents are successfully decrypted, the server automatically issues the bid documents for evaluation according to the bid evaluation method in the bidding documents.

[0071] The system receives the electronic bids submitted by bidders and generates a unique bid number for each bid document;

[0072] The bid document is divided into several information blocks according to a predefined structure. Each information block contains specific content. A name tag is set at the beginning of each information block and an end tag is set at the end of each information block.

[0073] During the information block segmentation and processing process, a conflict detection mechanism is introduced to automatically identify and resolve conflicts between information blocks. This conflict detection mechanism is crucial in bid processing. This mechanism allows the system to automatically identify and resolve potential conflicts between information blocks, ensuring the consistency and integrity of bid content. This feature not only improves processing efficiency and reduces human error, but also enhances the user experience, providing reliable and efficient service for both tenderers and bidders. During the tender processing process, conflicts can lead to omissions or duplications of information, reducing the quality and credibility of tender documents. The introduction of a conflict detection mechanism allows for timely identification and resolution of these issues, ensuring the accuracy and completeness of tender content. This not only improves the efficiency and quality of tender processing but also reduces potential disputes and controversies in subsequent processing, laying a more solid foundation for cooperation between tenderers and bidders. In summary, the introduction of a conflict detection mechanism plays an important role in tender processing, positively promoting the efficiency, accuracy, and reliability of the bidding process.

[0074] Scan the bid document content, identify and match the name tag and end tag of each information block, and proceed to the next step when the name tag and end tag of each information block can be identified and matched; otherwise, the bid document will be invalidated;

[0075] Successfully matched bids will be segmented into information blocks, each of which will be assigned a unique block number. The blocks will then be sent to the corresponding review platform based on the block number, and the review end will generate review information.

[0076] If the name tag and the end tag fail to match, the bid will be marked as invalid and the bidder will be automatically notified of the reason for the invalidation of the bid, allowing it to modify and resubmit.

[0077] After the bid evaluation method is set up, the system will package, sign, and seal the entire content of the bidding documents. After self-inspection to ensure that the content format is qualified, it will be encrypted with a digital certificate to form a unified format file and transmitted to the cloud server.

[0078] The following steps need to be performed after the bid opening:

[0079] After passing the identity verification of the multi-factor user authentication system, the login control client automatically obtains the projects to be reviewed;

[0080] After selecting and entering a project, preliminary and detailed reviews will be conducted on the mobile review page. The review results of each review step will be stored in the storage space allocated by the server, which will be used for data storage of this review project and summary calculation according to the rules set in the bid evaluation method.

[0081] The quotation review automatically collects data based on the previous review results, calculates scores based on the quotation algorithm, and generates a quotation score.

[0082] During the information block segmentation and processing, a conflict detection mechanism is introduced to automatically identify and handle conflicts between information blocks. The steps are as follows:

[0083] Use NLP technology to segment the bid text S, remove stop words and punctuation marks, and obtain the segmented text T, T = {w1, w2, ..., w n}, input information block text S, obtain sentence vector v(S) through BERT model, extract named entity E from sentence S, use syntactic analysis technology to build sentence syntax tree Tree(S), extract the main body and modification information of sentence S, calculate the semantic similarity between information blocks, set threshold to determine whether there is repeated or similar content, and calculate the semantic similarity Sim(S1,S2) of two sentences S1 and S2. Its expression is: According to the semantic similarity Sim(S1, S2), it is determined whether the two sentences have repeated or similar content. If Sim(S1, S2) ≥ the threshold, it is considered that S1 and S2 have repeated or similar content.

[0084] Automatic adjustments and corrections are made according to preset rules, and the processed information blocks are re-verified in real time to ensure that conflicts have been corrected.

[0085] Before identifying duplicate or similar content, we also use graph neural networks to model the relationship between information blocks, perform conflict detection through the graph structure, and treat each information block as a node in the graph. Finally, the final embedding h of each node is calculated i The expression is:

[0086]

[0087] Where, represents the feature vector of node i in the lth layer, N(i) represents the set of neighbor nodes of node i, c ij is the normalization constant, W (l) is the weight matrix of the lth layer, σ is the activation function;

[0088] According to the final embedding h of each node i, dynamically adjust the weights of various conflict detections, prioritize high-frequency and high-priority conflicts, and use reinforcement learning to optimize conflict handling strategies, automatically select the best handling method, build a reinforcement learning environment, define state, action, and reward functions, and optimize conflict handling strategies through experiments and learning. The expression is:

[0089]

[0090] Among them, π θt+1 (a t ∣s t ) means in state s t Next select action a t strategy, θ t are the parameters of the strategy, the state s t Indicates the current conflict handling environment information and processing progress, r t =R(s t ,a t ) represents the reward function, which means that in state s t Take action t After the reward, V φt (s t ) represents state s t The value function of t The expected cumulative reward under , φt is the parameter of the value function, γ is the discount factor, which indicates the discount rate of future rewards, α is the learning rate of policy update, η is the learning rate of weight adjustment, is the policy gradient, which represents the gradient of the policy parameters, αf i +βp i is an adaptive weight adjustment term, based on the conflict frequency f i and priority p i Adjust weight, action a t Indicates the specific operation selected to handle the conflict in the current state, and the reward r t Represents the reward value of the environment feedback after taking a certain action.

[0091] By using NLP technology to segment and syntactically analyze information blocks, and using the BERT model to obtain sentence vectors and extract named entities, we can efficiently process large amounts of bid texts, accurately identify information content and structure, and help improve the accuracy and efficiency of bid processing.

[0092] Secondly, by calculating the semantic similarity between information blocks and modeling the relationship between information blocks, the system can promptly detect and process duplicate or similar content, avoiding redundancy and contradictions in the bid information and improving the quality and credibility of the bid.

[0093] Furthermore, by introducing reinforcement learning to optimize conflict handling strategies, the system can dynamically adjust weights according to conflict frequency and priority, and automatically select the best handling method, effectively reducing manual intervention and improving the efficiency and accuracy of conflict handling.

[0094] The steps of using AHP to determine the basic weights, using ANP to adjust the weights, and using multi-level regression analysis to quantify the mutual influence between indicators are as follows:

[0095] Construct a hierarchical structure of the analytic hierarchy process, placing the evaluation objectives at the top level, the evaluation criteria at the middle level, and the specific evaluation plan at the bottom level. Based on the hierarchical structure, construct a judgment matrix A, and determine the relative weights of the criteria by comparing their importance in pairs. The expression of the judgment matrix A is:

[0096]

[0097] By calculating the maximum eigenvalue λ of the judgment matrix A max And the consistency ratio CR, which is expressed as:

[0098]

[0099] Among them, CI is the consistency index, RI is the random consistency index, and when CR < 0.1, the consistency of the judgment matrix is ​​accepted;

[0100] The mutual influence relationship in the network structure is quantified, and an unweighted super matrix W is constructed. The unweighted super matrix is ​​normalized according to the weight of each criterion to obtain the weighted super matrix W*. The limit super matrix W is obtained through multiple iterative calculations. ∞ , to determine the final weight of each criterion, the expression is:

[0101]

[0102] W * =normalize(W)

[0103]

[0104] Among them, W ij is a submatrix that represents the influence relationship from node j to node i. To normalize the unweighted super matrix W, lim k→∞ (W * ) k For the limit operation, the weighted super matrix W * Perform infinite iterations;

[0105] Collect historical and current data of the project, including the specific scores of each evaluation criterion and the project results. Based on the collected data, build a multi-level regression model and use the target variable Y to quantify the mutual influence between the evaluation criteria. The expression is:

[0106] Y=β0+β1X1+β2X2+…+β n X n +ò

[0107] Among them, X i represents the score of the i-th evaluation criterion, β i is the regression coefficient, ò is the error term;

[0108] Use the training data to train the regression model and minimize the mean square error (MSE). After training, use the validation data to validate the model, analyze the regression coefficients of the regression model, determine the specific impact of each evaluation criterion on the project results, and adjust the weight of each criterion based on these impacts.

[0109]

[0110] Where m is the number of samples, is the target variable predicted by the model;

[0111] Using a trained machine learning model, the characteristic data of the current project is input to predict the weights of each evaluation criterion. Based on the predicted results, the weights in the AHP and ANP models are dynamically adjusted to better reflect the actual situation. By adopting a comprehensive bid evaluation method, including AHP, ANP, and multi-level regression analysis, it is possible to fully consider the potential interactions and interdependencies between evaluation indicators, thereby improving the scientific nature and reliability of the evaluation results. First, by constructing a hierarchical structure of the analytic hierarchy process, the evaluation objectives, criteria, and specific evaluation plans are categorized, and a judgment matrix is ​​established. This effectively captures the relative importance of each indicator and reflects the interactive influence between indicators.

[0112] Secondly, the consistency of the judgment matrix was tested by calculating its consistency ratio, which ensured the consistency and reliability of the bid evaluation process and further emphasized the importance of the interdependence between indicators.

[0113] Furthermore, by constructing an unweighted supermatrix and normalizing it according to the weights of each criterion, a weighted supermatrix is ​​obtained. This is then iterated multiple times to obtain the limit supermatrix, which determines the final weights of each criterion. This process not only fully considers the mutual influence of each indicator, but also fully reflects the interdependence between indicators, thereby enhancing the comprehensiveness and holistic nature of the bid evaluation process.

[0114] Use the trained machine learning model, input the characteristic data of the current project, predict the weight of each evaluation criterion, and dynamically adjust the weights in the AHP and ANP models based on the prediction results to make them more consistent with the actual situation. The steps are as follows:

[0115] Collect relevant data of the current project, extract the key features of the project, and use feature selection methods to screen out the most representative features;

[0116] Input the feature data of the current project into the machine learning model, ensuring that the feature data format is consistent with that used during model training;

[0117] Use the model's prediction method to predict the weight of each evaluation criterion based on the input feature data; apply the model's predicted weight results to the AHP and ANP models to adjust the weight of each evaluation criterion;

[0118] Apply the weighted results predicted by the model to the AHP and ANP models to adjust the weights of each evaluation criterion. Update the weight parameters in the AHP and ANP models to make them consistent with the predicted results.

[0119] Update the weights of each criterion in the analytic hierarchy process (AHP), adjust the AHP judgment matrix, and recalculate the eigenvector and consistency based on the new weights;

[0120] In the Analytical Network Process (ANP) model, the weighted supermatrix and the limit supermatrix are recalculated to determine the final weight adjustment results. Relevant data for the current project is collected and key features are extracted. Feature selection methods are used to select the most representative features, which are then input into the machine learning model, ensuring that the feature data format is consistent with that used during model training. Subsequently, the model's prediction method is used to predict the weights of each evaluation criterion based on the input feature data. The predicted results are then applied to the AHP and ANP models to dynamically adjust the weights of each evaluation criterion. This process effectively combines the machine learning model with the traditional AHP and ANP models, leveraging the predictive power of machine learning and the weight adjustment capabilities of the AHP and ANP models to achieve automated adjustment of evaluation criterion weights. Subsequently, based on the model's predicted weights, the weight parameters in the AHP and ANP models are updated to align with the predicted results. This in turn updates the weights of each criterion in the Analytical Hierarchy Process (AHP), adjusts the AHP judgment matrix, and recalculates the eigenvectors and consistency based on the new weights. In the Analytical Network Process (ANP) model, the weighted supermatrix and the limit supermatrix are recalculated to determine the final weight adjustment results. The implementation of this series of steps can effectively improve the accuracy of the evaluation criteria weights, provide reliable data support for project decision-making, optimize resource allocation, and improve decision-making efficiency and quality.

[0121] In this embodiment, the detailed review step includes the following steps:

[0122] Set review content, review standards, score range and summary method according to business review and technical review respectively;

[0123] The system automatically checks whether the set scores, weights, summary scores and evaluation criteria are correct. If the check fails, return to modify and submit again until it succeeds.

[0124] In this embodiment, the quotation review step includes the following steps:

[0125] According to the bidding documents, add quotation review steps, set the price score calculation range and rules, benchmark value setting rules, and price calculation algorithm formula for this project;

[0126] The calculation range sets the bidders who have passed the preliminary review as valid bidders, and the bid evaluation benchmark price is the arithmetic mean of the valid bid quotations. When the number of valid bidders is greater than five, the highest and lowest bids are removed when calculating the bid evaluation benchmark price; when the number of valid bidders is less than or equal to five, the highest and lowest bids are not removed when calculating the bid evaluation benchmark price.

[0127] The calculation formula is:

[0128] Evaluation benchmark price A = (A1 + A2 + ... + An) / n, where n is the valid number and An is a valid bid price;

[0129] The formula for calculating the deviation rate of bid quotations is: Deviation rate = 100% × (bid quotations of valid bidders - bid evaluation benchmark price) / bid evaluation benchmark price;

[0130] The formula for calculating the bid quotation score is: bid quotation score = F - absolute value of deviation rate × 100 × E, where F is the weight of the bid quotation, F = 70 points, and E is the deduction value for each percentage point above or below the bid evaluation benchmark price, E = 2 points, and the deduction shall not exceed 10 points;

[0131] After the settings are completed and submitted, determine whether the set score, calculation algorithm, and evaluation criteria are correct. If the system setting detection is successful, proceed to the next step; otherwise, return to modify and submit again for verification until success.

[0132] Automatically obtain the projects to be evaluated, and automatically obtain the evaluation method, evaluation steps, and evaluation item standard settings based on the project information. Compare the bidders' documents according to the evaluation standards and conduct the evaluation according to the preset step-by-step process. Each step is automatically consolidated and summarized according to the evaluation data, and preliminary and detailed evaluations are completed in sequence. According to the preset calculation algorithm, the extracted input quotation data is automatically extracted, analyzed, corrected, and calculated to form the final price score and provide it for review. Finally, the summary results, evaluation report, and bid evaluation report are generated;

[0133] After all review steps are completed, the system will summarize the review data according to the rules and generate an evaluation report, which will be electronically signed using the equipped signature device.

[0134] In this embodiment, after the bid evaluation method is set up, the following steps are also included

[0135] The overall content of the bidding documents is packaged, signed, and sealed, and the content format of the documents is self-checked. After passing the self-check, the documents are encrypted with a digital certificate to form a unified format file, which is then transmitted to the cloud server and provided to bidders for download;

[0136] Bidders prepare bidding documents in accordance with the bidding documents and evaluation methods, and in accordance with the bidding document format. Finally, the bidding documents are packaged, signed, and encrypted with digital certificates to form a unified format file, which is then transmitted to the cloud server before the deadline for bidding document submission.

[0137] After the bid opening, the bidder decrypts the bid documents remotely. After all documents are successfully decrypted, the server automatically issues the bid documents for evaluation according to the bid evaluation method in the bidding documents.

[0138] A bid evaluation device based on an electronic bid evaluation device, comprising:

[0139] The real-time project progress monitoring module monitors the project progress in real time and determines whether the bidding project has reached the current progress. If the project progress reaches the current node, the system automatically obtains the current project information and enters the bidding document preparation and bid evaluation method setting; otherwise, the system continues to monitor the project progress;

[0140] Automatic bidding document generation module, which automatically obtains detailed information of the current project and automatically generates bidding documents based on project characteristics;

[0141] Review process setting and management module, setting preliminary review, detailed review and quotation review steps and defining the review criteria and score calculation method;

[0142] The bid processing and segmentation module receives bids submitted by bidders and generates bid numbers. The bids are segmented into a specific number of information blocks, with a name tag at the beginning and an end tag at the end of each block. The name tags and end tags are then text-matched. When the name tags and end tags of each block are recognized and matched, the segmentation step is executed; otherwise, the bid is discarded.

[0143] The audit information generation and management module sends the information block to the corresponding audit platform according to the information block number. The audit end generates audit information, which includes the score of the information block and corresponds to the information block through the identifier.

[0144] Multi-level review weight adjustment module uses AHP to determine basic weights, ANP to adjust weights, and multi-level regression analysis to quantify the mutual influence between indicators;

[0145] The automatic generation and encryption module of the bid evaluation report automatically generates the bid evaluation report during the data transmission and storage process, including the scoring and comprehensive analysis results of each evaluation step using multiple encryption technology. The automatic generation and encryption module of the bid evaluation report automatically generates the bid evaluation report, including the scoring and comprehensive analysis results of each evaluation step using multiple encryption technology.

[0146] like Figure 2 1 is a schematic diagram of the structure of a computer device for implementing a bid evaluation medium based on an electronic transaction platform provided by an embodiment of the present invention.

[0147] The computer device may include a processor, a memory and a bus, and may also include a computer program stored in the memory and run on the processor, such as a management program of an electronic bidding evaluation device.

[0148] Among them, the memory includes at least one type of readable storage medium, and the readable storage medium includes flash memory, mobile hard disk, multimedia card, card-type memory (for example: SD or DX memory, etc.), magnetic memory, disk, optical disk, etc. In some embodiments, the memory can be an internal storage unit of a computer device, such as a mobile hard disk of the computer device. In other embodiments, the memory can also be an external storage device of a computer device, such as a plug-in mobile hard disk, a smart memory card (SmartMediaCard, SMC), a secure digital (SecureDigital, SD) card, a flash card (FlashCard), etc. equipped on the computer device. Furthermore, the memory can also include both an internal storage unit and an external storage device of the computer device. The memory can not only be used to store application software and various types of data installed in the computer device, such as the code of a management program of an electronic bidding evaluation device, but can also be used to temporarily store data that has been output or is to be output.

[0149] In some embodiments, the processor may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and a combination of various control chips. The processor is the control core (Control Unit) of the computer device, connecting the various components of the entire computer device using various interfaces and lines, and executing programs or modules stored in the memory (such as a management program for an electronic bid evaluation device, etc.), as well as calling data stored in the memory, to perform various functions of the computer device and process data.

[0150] The bus may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is configured to enable communication between the memory and at least one processor, etc.

[0151] Figure 2 Only a computer device with components is shown, and those skilled in the art will understand that Figure 2 The illustrated structure does not constitute a limitation on the computer device, and may include fewer or more components than shown, or combine certain components, or arrange the components differently.

[0152] For example, although not shown, the computer device may further include a power source (such as a battery) for supplying power to the various components. Preferably, the power source may be logically connected to at least one processor via a power management device, thereby enabling functions such as charge management, discharge management, and power consumption management via the power management device. The power source may further include any of one or more DC or AC power sources, a recharging device, a power failure detection circuit, a power converter or inverter, a power status indicator, and the like. The computer device may further include various sensors, Bluetooth modules, Wi-Fi modules, and the like, which are not further described herein.

[0153] Furthermore, the computer device may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the computer device and other computer devices.

[0154] Optionally, the computer device may further include a user interface, which may be a display or an input unit (such as a keyboard). Optionally, the user interface may also be a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the computer device and to display a visual user interface.

[0155] It should be understood that the above embodiment is for illustration purposes only and the scope of the patent application is not limited to this structure.

[0156] The management program of an electronic bidding evaluation device stored in the memory of a computer device is a combination of multiple instructions.

[0157] Specifically, the specific implementation method of the processor for the above instructions can refer to Figure 1 The description of the relevant steps in the corresponding embodiments will not be repeated here.

[0158] Furthermore, if the modules / units integrated into a computer device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium may include any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0159] The present invention also provides a computer-readable storage medium, which stores a computer program. The computer program is executed by a processor of a computer device.

[0160] In the several embodiments provided by the present invention, it should be understood that the disclosed methods can be implemented in other ways. For example, the above-described embodiments of the invention are merely illustrative. For example, the division of modules is only a logical function division, and other division methods may be used in actual implementation.

[0161] Modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical units, that is, they may be located in one place or distributed across multiple network elements. Some or all of these modules may be selected to achieve the purpose of this embodiment based on actual needs.

[0162] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or hardware plus software functional modules.

[0163] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0164] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A bid evaluation method based on an electronic trading platform, characterized in that: The following steps are involved: Monitor the project progress in real time to determine whether the bidding project has reached the current progress. If the project progress reaches the current node, the system will automatically obtain the current project information and enter the bidding document preparation and bid evaluation method settings; otherwise, the system will continue to monitor the project progress. Add preliminary review steps based on the project bidding documents, set review content and review criteria according to qualification review, formal review, and responsiveness review, and determine whether the content and review criteria are correct after completing the settings and submitting them; Add detailed review steps based on the project bidding documents, set the review content, review criteria, score range and summary method according to the commercial review and technical review respectively, and after completing the settings and submitting them, check whether the set scores, weights, summary scores and review criteria are correct; Set the price score calculation range and rules, benchmark value setting rules, and price calculation algorithm formula. After completing the settings, the system will automatically detect and provide feedback; After the bid evaluation method is set up, the system will package, sign, and seal the entire bidding document. After the self-check content format is qualified, it will be encrypted with a digital certificate and transmitted to the cloud server in a unified format. After the bid opening, the bidder decrypts the bid documents remotely. Once all documents are successfully decrypted, the server automatically issues the bid documents for evaluation according to the bid evaluation method in the bidding documents. During the information block segmentation and processing, a conflict detection mechanism is introduced to automatically identify and handle conflicts between information blocks. The steps are as follows: Use NLP technology to convert information blocks into tender text Perform word segmentation, remove stop words and punctuation, and obtain the text after word segmentation , , enter the information block text , get the sentence vector through the BERT model , from the sentence Extract named entities , using syntactic analysis techniques to construct a grammar tree for the sentence , extract sentences The main body and modification information of the sentence are calculated, the semantic similarity between the information blocks is calculated, and the threshold is set to determine whether there is repeated or similar content. and , calculate its semantic similarity , whose expression is: , according to semantic similarity Determine whether there are repeated or similar contents in two sentences. If ≥ threshold, it is considered and There is duplicate or similar content; Automatically adjust and correct according to preset rules, and re-verify the processed information blocks in real time to ensure that conflicts have been corrected; Before identifying duplicate or similar content, we also use graph neural networks to model the relationship between information blocks, perform conflict detection through the graph structure, and treat each information block as a node in the graph. , and finally calculate the final embedding of each node The expression is: ； Where, represents the feature vector of node i in layer l, represents the set of neighbor nodes of node i, is the normalization constant, is the weight matrix of layer l, is the activation function; According to the final embedding of each node , dynamically adjust the weights of various conflict detections, prioritize high-frequency and high-priority conflicts, and use reinforcement learning to optimize conflict handling strategies, automatically select the best handling method, build a reinforcement learning environment, define state, action, and reward functions, and optimize conflict handling strategies through experiments and learning. The expression is: in, Indicates that the status Select Action strategy, are the parameters of the strategy, the state Indicates the current conflict handling environment information and processing progress, Represents the reward function, which means that in state Take action After receiving the reward, Indicates status The value function of The expected cumulative reward under are the parameters of the value function, is the discount factor, which represents the discount rate of future rewards, is the learning rate for policy updates, is the learning rate for weight adjustment, is the policy gradient, which represents the gradient of the policy parameters, Is an adaptive weight adjustment item, based on the conflict frequency and priority Adjust weights, actions Indicates the specific operation selected to handle the conflict in the current state, reward Represents the reward value of the environment feedback after taking a certain action; By using NLP technology to segment and parse information blocks, and using the BERT model to obtain sentence vectors and extract named entities, we can efficiently process large amounts of tender documents and accurately identify information content and structure, helping to improve the accuracy and efficiency of tender processing. Secondly, by calculating the semantic similarity between information blocks and modeling the relationships between information blocks, the system can promptly detect and process duplicate or similar content, avoiding redundancy and contradictions in the bid documents and improving the quality and credibility of the bid documents. Furthermore, by introducing reinforcement learning to optimize conflict handling strategies, the system can dynamically adjust weights based on conflict frequency and priority, automatically selecting the best handling method, effectively reducing manual intervention and improving the efficiency and accuracy of conflict handling. The steps of using AHP to determine the basic weights, using ANP to adjust the weights, and using multi-level regression analysis to quantify the mutual influence between indicators are as follows: Construct a hierarchical structure of the analytic hierarchy process, placing the evaluation objectives at the top level, the evaluation criteria at the middle level, and the specific evaluation plan at the bottom level. Based on the hierarchical structure, construct a judgment matrix A, and determine the relative weights of the criteria by comparing their importance in pairs. The expression of the judgment matrix A is: ； By calculating the judgment matrix The maximum eigenvalue of and consistency ratio , whose expression is: ; in, is a random consistency indicator, when When , the consistency of the judgment matrix is ​​accepted; Quantify the mutual influence relationships in the network structure and construct an unweighted super matrix , the unweighted super matrix is ​​normalized according to the weights of each criterion to obtain the weighted super matrix , the limit super matrix is ​​obtained through multiple iterative calculations , to determine the final weight of each criterion, the expression is: ； ； ； in, is a submatrix that represents the influence relationship from node j to node i. For the unweighted super matrix Normalization processing, For the limit operation, the weighted super matrix Perform infinite iterations; Collect historical and current data of the project, including the specific scores of each evaluation criteria and project results, and build a multi-level regression model based on the collected data. The mutual influence between the evaluation criteria is quantified as follows: ； in, represents the score of the i-th evaluation criterion, is the regression coefficient, is the error term; Use the training data to train the regression model and minimize the mean square error (MSE). After training, use the validation data to validate the model, analyze the regression coefficients of the regression model, determine the specific impact of each evaluation criterion on the project results, and adjust the weight of each criterion based on these impacts. ； in, is the sample size, is the target variable predicted by the model.

2. A bid evaluation method based on an electronic trading platform according to claim 1, characterized in that: Set review content and standards according to qualification review, form review, and responsiveness review; The system automatically checks whether the content and evaluation criteria are correct. If the detection fails, go back to modify and submit again until it succeeds.

3. The bid evaluation method based on an electronic trading platform according to claim 1, characterized in that: According to the bidding documents, add quotation review steps, set the price score calculation range and rules, benchmark value setting rules, and price calculation algorithm formula for this project; The calculation range sets the bidders who have passed the preliminary review as valid bidders, and the bid evaluation benchmark price is the arithmetic mean of the valid bid quotations. When the number of valid bidders is greater than five, the highest and lowest bids are removed when calculating the bid evaluation benchmark price; when the number of valid bidders is less than or equal to five, the highest and lowest bids are not removed when calculating the bid evaluation benchmark price. The calculation formula is: Evaluation benchmark price A=(A1+A2+...+An) / n, where n is the valid number and An is a valid bid price; The formula for calculating the deviation rate of bid quotations is: Deviation rate = 100% × (bid quotations of valid bidders - evaluation benchmark price) / evaluation benchmark price; The formula for calculating the bid quotation score is: bid quotation score = F - absolute value of deviation rate × 100 × E, where F is the weight of the bid quotation, F = 70 points, and E is the deduction value for each percentage point above or below the bid evaluation benchmark price, E = 2 points, and the deduction shall not exceed 10 points; After the settings are completed and submitted, determine whether the set score, calculation algorithm, and evaluation criteria are correct. If the system setting detection is successful, proceed to the next step; otherwise, return to modify and submit again for verification until success.

4. The bid evaluation method based on an electronic trading platform according to claim 1, characterized in that: The detailed review process includes the following steps: Set review content, review standards, score range and summary method according to business review and technical review respectively; The system automatically checks whether the set scores, weights, summary scores and evaluation criteria are correct. If the detection fails, return to modify and submit again until success.

5. The bid evaluation method based on an electronic trading platform according to claim 1, characterized in that: After the bid evaluation method is set up, the system will package, sign, and seal the entire content of the bidding documents. After self-inspection to ensure that the content format is qualified, it will be encrypted with a digital certificate to form a unified format file and transmitted to the cloud server.

6. The bid evaluation method based on an electronic trading platform according to claim 1, characterized in that: The following steps need to be performed after the bid opening: After passing the identity verification of the multi-factor user authentication system, the login control client automatically obtains the projects to be reviewed; After selecting and entering a project, preliminary and detailed reviews will be conducted on the mobile review page. The review results of each review step will be stored in the storage space allocated by the server, which will be used for data storage of this review project and summary calculation according to the rules set in the bid evaluation method. The quotation review automatically collects data based on the previous review results, calculates scores based on the quotation algorithm, and generates a quotation score.

7. The bid evaluation method based on an electronic trading platform according to claim 1, characterized in that: After all review steps are completed, the system will summarize the review data according to the rules and generate an evaluation report, which will be electronically signed using the equipped signature device.

8. A bid evaluation device based on an electronic trading platform, characterized in that: include: Monitor the project progress in real time to determine whether the bidding project has reached the current progress. If the project progress reaches the current node, the system will automatically obtain the current project information and enter the bidding document preparation and bid evaluation method settings; otherwise, the system will continue to monitor the project progress. The preliminary review module adds preliminary review steps based on the project bidding documents, sets review content and review criteria according to qualification review, form review, and responsiveness review, and determines whether the content and review criteria are correct after the settings are completed and submitted; The detailed review module adds detailed review steps based on the project bidding documents, sets the review content, review criteria, score range and summary method according to the commercial review and technical review respectively, and determines whether the set scores, weights, summary scores and review criteria are correct after the settings are submitted; Setting module, set the price score calculation range and rules, benchmark value setting rules, and price calculation algorithm formula. After the setting is completed, the system will automatically detect and feedback; Transmission module: After the bid evaluation method is set up, the system will package, sign, and seal the entire bidding document. After self-checking and confirming that the content format is qualified, it will be encrypted with a digital certificate and transmitted to the cloud server in a unified format; Evaluation module: After the bid opening, the bidder decrypts the bid documents remotely. After all documents are successfully decrypted, the server automatically issues the bid documents for evaluation according to the evaluation method in the bidding documents. During the information block segmentation and processing, a conflict detection mechanism is introduced to automatically identify and handle conflicts between information blocks. The steps are as follows: Use NLP technology to convert information blocks into tender text Perform word segmentation, remove stop words and punctuation, and obtain the text after word segmentation , , enter the information block text , get the sentence vector through the BERT model , from the sentence Extract named entities , using syntactic analysis techniques to construct a grammar tree for the sentence , extract sentences The main body and modification information of the sentence are calculated, the semantic similarity between the information blocks is calculated, and the threshold is set to determine whether there is repeated or similar content. and , calculate its semantic similarity , whose expression is: , according to semantic similarity Determine whether there are repeated or similar contents in two sentences. If ≥ threshold, it is considered and There is duplicate or similar content; Automatically adjust and correct according to preset rules, and re-verify the processed information blocks in real time to ensure that conflicts have been corrected; Before identifying duplicate or similar content, we also use graph neural networks to model the relationship between information blocks, perform conflict detection through the graph structure, and treat each information block as a node in the graph. , and finally calculate the final embedding of each node The expression is: ； Where, represents the feature vector of node i in layer l, represents the set of neighbor nodes of node i, is the normalization constant, is the weight matrix of layer l, is the activation function; According to the final embedding of each node , dynamically adjust the weights of various conflict detections, prioritize high-frequency and high-priority conflicts, and use reinforcement learning to optimize conflict handling strategies, automatically select the best handling method, build a reinforcement learning environment, define state, action, and reward functions, and optimize conflict handling strategies through experiments and learning. The expression is: in, Indicates that the status Next select action strategy, are the parameters of the strategy, the state Indicates the current conflict handling environment information and processing progress, Represents the reward function, which means that in state Take action After receiving the reward, Indicates status The value function of The expected cumulative reward under are the parameters of the value function, is the discount factor, which represents the discount rate of future rewards, is the learning rate for policy updates, is the learning rate for weight adjustment, is the policy gradient, which represents the gradient of the policy parameters, Is an adaptive weight adjustment item, based on the conflict frequency and priority Adjust weights, actions Indicates the specific operation selected to handle the conflict in the current state, reward Represents the reward value of the environment feedback after taking a certain action; By using NLP technology to segment and parse information blocks, and using the BERT model to obtain sentence vectors and extract named entities, we can efficiently process large amounts of tender documents and accurately identify information content and structure, helping to improve the accuracy and efficiency of tender processing. Secondly, by calculating the semantic similarity between information blocks and modeling the relationships between information blocks, the system can promptly detect and process duplicate or similar content, avoiding redundancy and contradictions in the bid documents and improving the quality and credibility of the bid documents. Furthermore, by introducing reinforcement learning to optimize conflict handling strategies, the system can dynamically adjust weights based on conflict frequency and priority, automatically selecting the best handling method, effectively reducing manual intervention and improving the efficiency and accuracy of conflict handling. The steps of using AHP to determine the basic weights, using ANP to adjust the weights, and using multi-level regression analysis to quantify the mutual influence between indicators are as follows: Construct a hierarchical structure of the analytic hierarchy process, placing the evaluation objectives at the top level, the evaluation criteria at the middle level, and the specific evaluation plan at the bottom level. Based on the hierarchical structure, construct a judgment matrix A, and determine the relative weights of the criteria by comparing their importance in pairs. The expression of the judgment matrix A is: ； By calculating the judgment matrix The maximum eigenvalue of and consistency ratio , whose expression is: ; in, is a random consistency indicator, when When , the consistency of the judgment matrix is ​​accepted; Quantify the mutual influence relationships in the network structure and construct an unweighted super matrix , the unweighted super matrix is ​​normalized according to the weights of each criterion to obtain the weighted super matrix , the limit super matrix is ​​obtained through multiple iterative calculations , to determine the final weight of each criterion, the expression is: ； ； ； in, is a submatrix that represents the influence relationship from node j to node i. For the unweighted super matrix Normalization processing, For the limit operation, the weighted super matrix Perform infinite iterations; Collect historical and current data of the project, including the specific scores of each evaluation criteria and project results, and build a multi-level regression model based on the collected data. The mutual influence between the evaluation criteria is quantified as follows: ； in, represents the score of the i-th evaluation criterion, is the regression coefficient, is the error term; Use the training data to train the regression model and minimize the mean square error (MSE). After training, use the validation data to validate the model, analyze the regression coefficients of the regression model, determine the specific impact of each evaluation criterion on the project results, and adjust the weight of each criterion based on these impacts. ； in, is the sample size, is the target variable predicted by the model.

9. A storage medium, characterized in that: It includes at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the bid evaluation method based on the electronic trading platform as described in any one of claims 1 to 7 and the bid evaluation device based on the electronic trading platform as described in claim 8.

Images