An electronic bidding system and method based on XML structured data

Abstract

The application relates to the technical field of electronic bidding systems, in particular to an electronic bidding system and method based on XML structured data, which comprises the following units: a structured content division unit which divides bidding content information into multiple segments and converts the segments into corresponding XML segments; a data binding unit which establishes a mapping relationship between structured label variables and associated data in a system database or user input project information; a template storage unit which combines the screened XML segments into complete XML bidding templates and stores the templates according to industry categories; a bidding file generation unit which calls the bidding templates matched with the user input project information, automatically fills the associated data into the label variables, and generates complete bidding XML files; a bidding file generation unit which classifies the project data filled by bidders according to requirements and generates bidding XML file data; and an evaluation assisting unit which extracts the project data bound by the structured label variables and generates a horizontal comparison analysis table.

Description

Technical Field

[0001] This invention relates to the field of electronic bidding system technology, and in particular to an electronic bidding system and method based on XML structured data. Background Technology

[0002] With the continuous advancement of e-government and information technology construction, electronic bidding systems have been widely used in engineering construction, government procurement, and enterprise procurement. Traditional electronic bidding systems mostly organize bidding content and response documents in text document format, which not only results in unclear and inconsistent data structures but also makes it difficult to achieve efficient intelligent analysis and automated processing, seriously affecting the standardization of the bidding process and the efficiency and fairness of the bid evaluation.

[0003] Especially in the process of generating bidding documents and tender documents, existing technologies generally rely on manual filling and template editing, which leads to problems such as data redundancy, inconsistent formats, and frequent filling errors. In addition, the evaluation process lacks structured data support, and evaluation experts often need to manually review and compare a large number of documents, which is inefficient and easily influenced by human factors, making it difficult to achieve objective, quantitative, and automated evaluation decision support.

[0004] Therefore, there is an urgent need for an electronic bidding system that can structure bidding content, intelligently generate bidding data, and efficiently compare evaluation information, in order to improve the accuracy of bidding document generation, the standardization of bid responses, and the scientific nature and efficiency of the evaluation process. Summary of the Invention

[0005] In view of the above-mentioned shortcomings and deficiencies of the prior art, the present invention provides an electronic bidding system and method based on XML structured data.

[0006] To achieve the above objectives, the main technical solutions adopted by the present invention include:

[0007] In a first aspect, embodiments of the present invention provide an electronic bidding system based on XML structured data, comprising:

[0008] The structured content segmentation unit is used to divide the bidding content information entered by the bidding administrator into multiple content fragments and convert each content fragment into a corresponding XML fragment;

[0009] A tag variable definition unit is used to define structured tag variables in the XML fragment, wherein the structured tag variables are XML placeholder tags with semantic identifiers and unique bookmark-ids;

[0010] The data binding unit is used to establish a mapping relationship between the structured tag variables and the associated data in the system database or the project information input by the user;

[0011] The template storage unit is used to combine the XML fragments selected by the bidding administrator into a complete XML bidding template and store the complete XML bidding template according to industry category.

[0012] The tender document generation unit receives project information input by the user, and calls a complete XML tender template that matches the project information input by the user. It is also used to match the project information input by the user with the structured tag variables, and automatically fill the project information input by the user and the associated data in the system database into the corresponding structured tag variables in the matched complete XML tender template according to the mapping relationship established by the data binding unit, so as to obtain a complete tender XML file.

[0013] The tender document generation unit is used to classify the various project data filled in by the bidder according to the project requirements in the complete tender XML document, and obtain the tender XML document data.

[0014] The bid evaluation assistance unit is used to extract the project data bound to the structured tag variables in each bid XML file based on the path of the project data in the bid XML file data that is pre-bound to the structured tag variables, and generate a horizontal comparison analysis table containing the project data bound to the structured tag variables in each bid XML file data.

[0015] Preferably, the system further includes:

[0016] The bid evaluation and screening unit is used to calculate the score of each bid XML file based on the project data bound to the structured tag variables in the bid XML file data, and select the bid file data with the highest score as the winning candidate.

[0017] Preferably,

[0018] The bidding information includes basic information about the proposed project, the scope of the bidding, technical specifications, qualification requirements, scoring criteria, bid document format requirements, and draft contract terms.

[0019] The project information entered by the user includes the specific project name, project budget, industry category, implementation location, planned duration, project manager, and information about the bidding unit.

[0020] Preferably,

[0021] The structured content segmentation unit divides the bidding information entered by the bidding administrator into multiple content fragments, and converts each content fragment into a corresponding XML fragment, specifically including:

[0022] The structured content segmentation unit is used to initially segment the bidding content information input by the bidding administrator at the sentence level, resulting in multiple sentence-level content fragments;

[0023] Based on a preset objective function, the multiple sentence-level content fragments are combined to obtain multiple structurally sound bidding content fragments, each of which serves as the basic unit for generating XML fragments; wherein:

[0024] The objective function is used to optimize the selection among multiple sentence-level content fragment combination methods to obtain the content segmentation result with the highest total relevance score and the lowest redundancy.

[0025] The objective function is:

[0026]

[0027] C * This represents the optimal set of tender content fragments obtained through optimization, including a set of finally selected tender content fragments;

[0028] C = {c1, ..., c} i c n};c i It is the i-th segment of the tender content selected during the optimization process of the objective function;

[0029] u i This represents the i-th candidate tender content segment;

[0030] R(u i ) is the semantic relevance score of the i-th candidate tender content fragment;

[0031] D(u i ) represents the redundancy score of the i-th candidate tender content segment;

[0032] ε represents the semantic weight;

[0033] θ represents the weight of the redundancy penalty term;

[0034] n represents the total number of selected tender content segments.

[0035] Preferably,

[0036] Among them, the semantic relevance score R(u) of the i-th candidate bidding content fragment i The result was obtained using the first method.

[0037] The first method is to use a pre-trained language model to vectorize the i-th candidate tender content fragment and its preceding and following tender content fragments;

[0038] The semantic similarity between the i-th candidate tender content segment and its adjacent tender content segments is calculated based on cosine similarity. The semantic similarity between the i-th tender content segment and its adjacent tender content segments is added to obtain the first similarity value. Then, the reciprocal of the first similarity value is used as the semantic relevance score of the i-th candidate tender content segment.

[0039] Among them, the redundancy score D(u) of the i-th candidate bidding content segment i The result was obtained using the second method.

[0040] The second method is:

[0041] Construct sentence vector sets for each candidate tender content fragment;

[0042] Calculate the average semantic similarity between the i-th candidate tender content fragment and all other candidate tender content fragments, and use this average similarity as the redundancy score of the selected i-th tender content fragment.

[0043] Preferably,

[0044] The tender document generation unit receives project information input by the user and retrieves a complete XML tender template that matches the user's input project information.

[0045] The tender document generation unit uses formula (1) to obtain the fitness score of the complete XML tender template for the user-input project information;

[0046]

[0047] Among them, F T A score is given to the fit of the user-input project information in the complete XML tender template;

[0048] m represents the total number of specified fields involved in the complete XML tender template;

[0049] w i The industry weight coefficient for the i-th specified field in the complete XML tender template;

[0050] D i The i-th requirement description in the project information entered by the user;

[0051] T i The structure field definition for the i-th requirement description in the user-input project information within the complete XML tender template;

[0052] Sim(D i T i ) is D i With T iVector cosine similarity;

[0053] The complete XML tender template with the highest fit score to the user-input project information is selected as the matching complete XML tender template.

[0054] Preferably,

[0055] The tender document generation unit, based on the mapping relationship established by the data binding unit, automatically fills the user-input project information and related data from the system database into the corresponding structured tag variables in the matching complete XML tender template, resulting in a complete tender XML file, specifically including:

[0056] The tender document generation unit automatically fills the user-input project information and related data in the system database into the corresponding structured tag variables in the matching complete XML tender template according to the priority score from high to low.

[0057] The fill priority score is calculated using formula (2).

[0058] The formula (2) is:

[0059]

[0060] F freq (t k The field t corresponds to the project information entered by the user and the associated data in the system database. k Frequency of occurrence in the matching complete XML tender template;

[0061] β1 is the weighting parameter;

[0062] σ is an adjustable threshold;

[0063] S fill (t k The field t corresponds to the project information entered by the user and the associated data in the system database. k Fill priority score.

[0064] Preferably,

[0065] The various project data filled in by the bidders include: the composition of the bidding company's technical personnel, the number of successful cases of the bidding company, the total price of the bidding company, the warranty period of the bidding company, the qualification level of the bidding company, and the registered capital of the bidding company.

[0066] Preferably,

[0067] The scoring model is as follows:

[0068] S total =ρ·S tech+γ·S price +δ·S business ;

[0069] Among them, S tech = 0.6·f1(x1) + 0.4·f2(x2);

[0070] f1(x1)=min(100,10·log(x1+1)); x1=a+0.5b+0.2c;

[0071] Where a represents the number of PhDs among the technical personnel of the bidding company; b represents the number of Master's degree holders among the technical personnel of the bidding company; and c represents the number of Bachelor's degree holders among the technical personnel of the bidding company.

[0072] f2(x2)=min(100,20·log(x2+1));

[0073] x2 represents the number of successful cases for the bidding company;

[0074] S price = 0.7·f3(x3) + 0.3·f4(x4);

[0075]

[0076] x3 represents the total bid price from the bidding company;

[0077] P min The lowest price in the price range; P max The highest price in the price range;

[0078] f4(x4) = min(100, 15·x4);

[0079] x4 represents the warranty period for the bidding company;

[0080] S business = 0.5·f5(x5) + 0.5·f6(x6);

[0081]

[0082] x5 represents the qualification level of the bidding company, where the qualification level of the bidding company is A, B, or C.

[0083] f6(x6)=min(100,20·log 10 (x6+1));

[0084] x6 represents the registered capital of the bidding company.

[0085] On the other hand, this embodiment also provides an electronic bidding method based on XML structured data, which is executed by the electronic bidding system based on XML structured data described in the first aspect.

[0086] The beneficial effects of this invention are:

[0087] This invention discloses an electronic bidding system based on XML structured data. By employing structured content segmentation units to divide bidding information into corresponding XML fragments, it achieves standardized processing and structured storage of bidding information compared to existing technologies, thereby improving information processing efficiency and data reusability. By using tag variable definition units to introduce structured tag variables with semantic identifiers and unique bookmark-ids into the XML fragments, it allows for flexible definition and positioning of data fields compared to traditional static template methods, enhancing system scalability and adaptability to various bidding paradigms. Furthermore, by using data binding units to map structured tag variables to database or user input data, it can automatically fill in the bidding document content compared to manual filling, reducing manual operations and improving accuracy.

[0088] By setting up template storage units to store complete XML bidding templates by industry category, compared to a unified template, matching templates can be called on demand, thereby improving template reusability and industry adaptability.

[0089] Because the tender document generation unit intelligently generates complete tender XML files based on user input, it automatically generates standardized documents, accelerating the document creation process and reducing error rates compared to manual editing. Because the bid document generation unit structures and categorizes the content submitted by bidders, it achieves unified data formatting, improving the efficiency of subsequent bid evaluation. Because the bid evaluation assistance unit extracts bid data based on structured paths and generates horizontal comparison tables, it significantly improves evaluation efficiency and the objectivity of data comparison compared to manual comparison, assisting experts in conducting efficient and accurate bid evaluations.

[0090] Furthermore, the system also includes a bid evaluation and screening unit that scores and ranks each bid document based on a scoring model. Compared to purely manual scoring, it can generate intelligent bid evaluation suggestions, thereby improving the scientific nature, fairness, and efficiency of bid evaluation. Attached Figure Description

[0091] Figure 1 This is a schematic diagram of an electronic bidding system based on XML structured data according to the present invention;

[0092] Figure 2 This is a schematic diagram of an electronic bidding system based on XML structured data, as shown in Embodiment 2 of the present invention. Detailed Implementation

[0093] To better explain and facilitate understanding of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0094] To better understand the above technical solutions, exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that the present invention can be understood more clearly and thoroughly, and that the scope of the present invention can be fully conveyed to those skilled in the art.

[0095] Example 1

[0096] In traditional bidding processes, the bidding entity manually drafts the bidding documents, bidders manually fill out forms, and experts manually evaluate the bids. This method is time-consuming, labor-intensive, and prone to errors. This system, through structured data and automated processing, makes this process intelligent, fast, and secure. (See also...) Figure 1 This embodiment provides an electronic bidding system based on XML structured data, including:

[0097] The structured content segmentation unit is used to divide the bidding content information entered by the bidding administrator into multiple content fragments and convert each content fragment into a corresponding XML fragment;

[0098] In this embodiment, a large amount of text information input by the tenderer is automatically broken down into structured "small segments," and each segment is converted into XML format that the computer can understand.

[0099] The bidding information includes basic information about the proposed project, the scope of the bidding, technical specifications, qualification requirements, scoring criteria, bid document format requirements, and draft contract terms.

[0100] A tag variable definition unit is used to define structured tag variables in the XML fragment, wherein the structured tag variables are XML placeholder tags with semantic identifiers and unique bookmark-ids;

[0101] The data binding unit is used to establish a mapping relationship between the structured tag variables and the associated data in the system database or the project information input by the user;

[0102] The template storage unit is used to combine the XML fragments selected by the bidding administrator into a complete XML bidding template and store the complete XML bidding template according to industry category.

[0103] In this embodiment, the template storage unit combines the various parts of the bidding documents into a complete bidding document template (XML format) and saves them according to industry type, such as "construction" and "software development". Different users can then choose the most suitable template from these available options.

[0104] The tender document generation unit receives project information input by the user, and calls a complete XML tender template that matches the project information input by the user. It is also used to match the project information input by the user with the structured tag variables, and automatically fill the project information input by the user and the associated data in the system database into the corresponding structured tag variables in the matched complete XML tender template according to the mapping relationship established by the data binding unit, so as to obtain a complete tender XML file.

[0105] The project information entered by the user includes the specific project name, project budget, industry category, implementation location, planned duration, project manager, and information about the bidding unit.

[0106] The tender document generation unit is used to classify the various project data filled in by the bidder according to the project requirements in the complete tender XML document, and obtain the tender XML document data.

[0107] The bid evaluation assistance unit is used to extract the project data bound to the structured tag variables in each bid XML file based on the path of the project data in the bid XML file data that is pre-bound to the structured tag variables, and generate a horizontal comparison analysis table containing the project data bound to the structured tag variables in each bid XML file data.

[0108] In the practical application of this embodiment, the structured content segmentation unit divides the bidding content information input by the bidding administrator into multiple content fragments and converts each content fragment into a corresponding XML fragment, specifically including:

[0109] The structured content segmentation unit is used to initially segment the bidding content information input by the bidding administrator at the sentence level, resulting in multiple sentence-level content fragments;

[0110] Based on a preset objective function, the multiple sentence-level content fragments are combined to obtain multiple structurally sound bidding content fragments, each of which serves as the basic unit for generating XML fragments; wherein:

[0111] The objective function is used to optimize the selection among multiple sentence-level content fragment combination methods to obtain the content segmentation result with the highest total relevance score and the lowest redundancy.

[0112] The objective function is:

[0113]

[0114] C * C represents the optimal set of tender content fragments obtained through optimization, including a set of finally selected tender content fragments; C = {c1, ..., c2} i c n};c i It is the i-th fragment of the tender content selected during the optimization process of the objective function; u i R(u) represents the i-th fragment of the tender content of the candidate; i ) represents the semantic relevance score of the i-th candidate tender content fragment; D(u i ) represents the redundancy score of the i-th candidate tender content segment; ε is the semantic weight; θ is the weight of the redundancy penalty term. n represents the total number of selected tender content segments.

[0115] The structured content segmentation unit first categorizes the tender content at the sentence level. For example, suppose the tender administrator inputs the content: "This project requires the procurement of server equipment with a CPU frequency of no less than 2.4GHz, a memory capacity of no less than 32GB, and a storage space of no less than 1TB. The equipment must support RAID5 configuration and provide a three-year warranty. Bidders must have ISO9001 certification." The sentence-level segmentation result would be:

[0116] 1: "This project involves the procurement of server equipment."

[0117] 2: "The CPU frequency must be no less than 2.4GHz."

[0118] 3: "Memory capacity is no less than 32GB."

[0119] 4: "Storage space is no less than 1TB."

[0120] 5: "The device must support RAID5 configuration."

[0121] 6: "And provides a three-year warranty service."

[0122] 7: "Bidders must have ISO9001 certification."

[0123] The system then uses a preset objective function to optimize and combine these sentence-level fragments to form structurally sound tender content fragments. The system may evaluate the following combination methods:

[0124] Candidate combination 1:

[0125] "This project requires the procurement of server equipment with a CPU frequency of no less than 2.4GHz, a memory capacity of no less than 32GB, and a storage space of no less than 1TB." -------- Excerpt from the first tender item in Candidate Combination 1;

[0126] "The equipment must support RAID5 configuration and provide a three-year warranty." -------- Excerpt from the second tender item in Candidate Combination 1;

[0127] "Bidders must have ISO9001 certification." — Excerpt from the third tender item in Candidate Combination 1;

[0128] Candidate combination 2:

[0129] "This project is procuring server equipment," — excerpt from the first tender item in candidate combination 2;

[0130] "The CPU clock speed must be no less than 2.4GHz, the memory capacity no less than 32GB, and the storage space no less than 1TB. The device must support RAID5 configuration." — Excerpt from the second tender content in Candidate Combination 2;

[0131] "A three-year warranty service shall be provided. Bidders must have ISO9001 certification." -------- Excerpt from the third tender item in Candidate Combination 2;

[0132] The system calculates the total score for each combination to obtain the content segmentation result with the highest total relevance score and the lowest redundancy. Finally, each selected tender content segment is converted into a corresponding XML segment. This structured processing method makes the tender information more standardized, easier to process, and facilitates subsequent automatic analysis and comparison.

[0133] Among them, the semantic relevance score R(u) of the i-th candidate bidding content fragment i The result was obtained using the first method.

[0134] The first method is to use a pre-trained language model (such as BERT, RoBERTa, etc.) to vectorize the i-th candidate tender content fragment and its surrounding tender content fragments;

[0135] The semantic similarity between the i-th candidate tender content segment and its adjacent tender content segments is calculated based on cosine similarity. The semantic similarity between the i-th tender content segment and its adjacent tender content segments is added to obtain the first similarity value. Then, the reciprocal of the first similarity value is used as the semantic relevance score of the i-th candidate tender content segment.

[0136] For example, assuming the third candidate tender content fragment and its preceding and following tender content fragments are vectorized, a pre-trained language model (such as BERT, RoBERTa, etc.) is used to "vectorize" the second, third, and fourth tender content fragments. That is, these three text segments are converted into semantic vector representations that machines can understand. Then, the cosine similarity is calculated between the third tender content fragment and the two preceding and following tender content fragments. Finally, the two are added together to obtain the first similarity value.

[0137] In this embodiment, by semantic vectorizing the preceding and following text and calculating cosine similarity, the algorithm no longer relies on keyword matching but instead utilizes a language model to understand the closeness of meaning, thus more accurately measuring the rationality of a segment. When using a language model for vectorization, the model considers the context; even if a segment lacks keywords, it can judge its rationality based on the surrounding semantics. This is more intelligent than traditional word segmentation or keyword algorithms. By calculating scores, segments that are semantically disconnected from the preceding and following content can be filtered out, such as off-topic content, incomplete formatting, or irrelevant explanations, improving the clarity and professionalism of the tender content.

[0138] Among them, the redundancy score D(u) of the i-th candidate bidding content segment i The result was obtained using the second method.

[0139] The second method is:

[0140] Construct sentence vector sets for each candidate tender content fragment;

[0141] Calculate the average semantic similarity between the i-th candidate tender content fragment and all other candidate tender content fragments, and use this average similarity as the redundancy score of the selected i-th tender content fragment.

[0142] In this embodiment, by calculating the average similarity between each segment and all other segments, semantically repetitive content can be automatically identified, even if the expressions are different. Traditional methods may rely on keyword matching, but this is easily misled by synonyms. However, by calculating semantic similarity based on a language model, similar meanings can be identified even if the words appear different on the surface.

[0143] The tender document generation unit receives project information input by the user and retrieves a complete XML tender template that matches the user's input project information.

[0144] The tender document generation unit uses formula (1) to obtain the fitness score of the complete XML tender template for the user-input project information;

[0145]

[0146] Among them, F T The score given to the complete XML tender template reflects the suitability of the user-inputted project information; m is the total number of specified fields involved in the complete XML tender template; w i The industry weight coefficient for the i-th specified field in the complete XML tender template; D i The i-th requirement description in the project information entered by the user; T i The complete XML tender template defines the structure field definition for the i-th requirement description in the user-input project information; Sim(D i T i ) is D i With T i The vector cosine similarity.

[0147] The complete XML tender template with the highest fit score to the user-input project information is selected as the matching complete XML tender template.

[0148] In this embodiment, after the user inputs project information, the system filters candidate templates from the template library and calculates the matching degree between each template and the user's needs. The matching degree calculation uses a weighted similarity algorithm, comprehensively considering the importance of each field and content similarity. The system calculates a fitness score for each template. The scoring formula considers both content matching degree (calculated through text vector similarity) and the industry weight of different fields (e.g., qualification requirements have higher weight than labor options). Finally, the system automatically selects the template with the highest score. The specific parameters input by the user are automatically filled into the selected standardized XML template to generate a standardized tender document, which is the matched complete XML tender template.

[0149] Specifically, the tender document generation unit, based on the mapping relationship established by the data binding unit, automatically fills the user-input project information and related data from the system database into the corresponding structured tag variables in the matching complete XML tender template, thereby obtaining a complete tender XML file, specifically including:

[0150] The tender document generation unit automatically fills the user-input project information and related data in the system database into the corresponding structured tag variables in the matching complete XML tender template according to the priority score from high to low.

[0151] The fill priority score is calculated using formula (2).

[0152] The formula (2) is:

[0153]

[0154] F freq (tk The field t corresponds to the project information entered by the user and the associated data in the system database. k Frequency of occurrence in the matched complete XML tender template; β1 is the weight parameter; σ is an adjustable threshold; S fill (t k The field t corresponds to the project information entered by the user and the associated data in the system database. k Fill priority score.

[0155] F freq (t k This reflects the frequency of a field's appearance in the template. High-frequency fields automatically receive higher priority. For example, in engineering bidding, high-frequency core fields such as "qualification requirements" and "technical specifications" will be filled first, while low-frequency fields such as "remarks" will be processed later. Prioritizing high-frequency core fields avoids omitting key information.

[0156] In this embodiment, the various project data filled in by the bidder include: the composition of the bidding company's technical personnel, the number of successful cases of the bidding company, the total price of the bidding company, the warranty period of the bidding company, the qualification level of the bidding company, and the registered capital of the bidding company.

[0157] In one specific embodiment, the system further includes:

[0158] The bid evaluation and screening unit is used to calculate the score of each bid XML file based on the project data bound to the structured tag variables in the bid XML file data, and select the bid file data with the highest score as the winning candidate.

[0159] The scoring model is as follows:

[0160] S total =ρ·S tech +γ·S price +δ·S business ;

[0161] Among them, S tech = 0.6·f1(x1) + 0.4·f2(x2);

[0162] f1(x1)=min(100,10·log(x1+1)); x1=a+0.5b+0.2c;

[0163] Where a represents the number of PhDs among the technical personnel of the bidding company; b represents the number of Master's degree holders among the technical personnel of the bidding company; and c represents the number of Bachelor's degree holders among the technical personnel of the bidding company.

[0164] f2(x2)=min(100,20·log(x2+1));

[0165] x2 represents the number of successful cases for the bidding company;

[0166] S price = 0.7·f3(x3) + 0.3·f4(x4);

[0167]

[0168] x3 represents the total bid price from the bidding company;

[0169] P min The lowest price in the price range; P max The highest price in the price range;

[0170] f4(x4) = min(100, 15·x4);

[0171] x4 represents the warranty period for the bidding company;

[0172] S business = 0.5·f5(x5) + 0.5·f6(x6);

[0173]

[0174] x5 represents the qualification level of the bidding company, where the qualification level of the bidding company is A, B, or C.

[0175] f6(x6)=min(100,20·log 10 (x6+1));

[0176] x6 represents the registered capital of the bidding company.

[0177] On the other hand, this embodiment also provides an electronic bidding method based on XML structured data, which is executed by the electronic bidding system based on XML structured data described in Embodiment 1.

[0178] Example 2

[0179] See Figure 2 This embodiment also provides an electronic bidding system based on XML structured data. Building upon the electronic bidding system described in Embodiment 1, it further enhances the real-time adjustment capability of bidding content, the recommendation optimization effect, and the interactive feedback mechanism, aiming to achieve more efficient, intelligent, and personalized electronic bidding process management. The electronic bidding system based on XML structured data described in this embodiment (based on the electronic bidding system of Embodiment 1) also includes the following key functional units:

[0180] The user behavior data collection unit automatically and continuously collects user behavior data during the system's operation by users (including but not limited to bidding administrators, bidders, and evaluation experts). This data includes, but is not limited to: dwell time on structured bidding segments; frequency and location of modifications to bidding content; preferred terminology and templates; historically used combinations of structured segments; evaluation experts' scoring habits and biases across different dimensions (such as technical response and commercial pricing); and the level of attention to common tags across different industry sectors. Through continuous collection of this behavioral data, the system can generate fine-grained user behavior profiles, serving as a core input for recommendation and optimization models.

[0181] The semantic recommendation optimization unit combines historical data output by the user behavior collection unit, the system's preset XML tag variable set, industry standard structural paradigms, and a high-quality historical bidding case library to perform semantic-level recommendation and optimization of the structure and content of bidding documents.

[0182] The main functions of the semantic recommendation optimization unit in this embodiment include: using a deep learning model based on an attention mechanism (such as a combination of BERT and multi-head attention mechanism) to comprehensively evaluate the logical connection, terminology consistency, and industry language compliance among the segments in the bidding XML document; when a user manually adjusts the content of an XML segment, the system can predict its subsequent possible structure and content in real time and generate a set of structured suggested segments or templates for the user to choose to replace or confirm; the recommendation mechanism supports multi-dimensional scoring of content conciseness, industry matching degree, and language standardization, guiding users to generate higher-quality and more standardized bidding documents.

[0183] The dynamic tag weight adjustment unit provides a function to dynamically adjust the weights of structured tag variables to adapt to the different bidding priorities under different industries and project types. Its mechanism includes: based on the user-specified industry type (such as software development, civil engineering, equipment procurement, etc.), the system automatically increases the priority of XML tags highly relevant to that industry; for example, in software projects, tags such as "technical architecture design," "security mechanism requirements," and "delivery cycle control" will receive higher recommendation priority; while in civil engineering, tags such as "construction site management capabilities," "qualification level," and "large equipment configuration" will have higher weights. The dynamic weight adjustment is based on: analysis of the system's built-in industry corpus, correlation analysis of tag hit rate and bid success rate in past projects, and historical bidder response data.

[0184] To achieve industry-adaptability and intelligent recommendation driven by historical data, this system introduces a formula for calculating tag semantic weights based on multi-source information fusion:

[0185]

[0186] Among them, A i H represents the semantic attention score obtained by the tag in the current industry corpus. i This indicates the frequency of its appearance in historical winning projects, T i The formula represents the average hit rate of bidder responses in history, with k1 and k2 being industry-adaptive adjustment parameters. This formula ensures a dynamic balance between semantic adaptation, industry experience, and bidder response behavior in structured tag recommendation. Through this model, the system can calculate and rank recommended tags in real time, improving the industry matching accuracy and response efficiency of the bidding structure.

[0187] The interactive evaluation preference learning unit automatically extracts the expert's scoring preferences across multiple dimensions (technical response, commercial terms, service commitments, etc.) during the expert scoring and comment input process, constructing a personalized "evaluation preference profile" for each expert. In a new round of evaluation tasks, the system can detect deviations between the expert's scores and the project's scoring criteria, and issue contradictory alerts when anomalies occur (such as low scores for high-weight indicators). Simultaneously, based on the expert's historical scoring patterns, the system can generate preliminary reference scores or suggest reasonable ranges for the expert's confirmation, improving the consistency and fairness of the evaluation. This unit can also identify behaviors that significantly deviate from general scoring logic during the evaluation process, allowing bidding administrators to review or recalibrate weights.

[0188] The real-time verification and intelligent error correction unit is used during the generation and submission of bidding documents or tender documents to perform comprehensive real-time verification and intelligent error correction based on the semantic logic and contextual consistency between structured XML fragments. Specifically, this includes: detecting whether there are contradictions or conflicting references between contract terms and technical specifications; detecting whether there are missing items or important fields in the scoring criteria; identifying common logical errors such as redundant or repetitive fragments, empty fields, and incomplete key content; the system uses a graph structure to express the logical dependencies between structured XML tags and uses graph matching algorithms to detect structural breaks and promptly identify potential errors; the system also supports user-defined sensitive words and keyword groups to achieve content risk filtering (such as inconsistent contract periods, ambiguous payment methods, etc.).

[0189] Through the integration and collaborative work of the above-mentioned units, the system described in this embodiment can realize personalized auxiliary generation, semantic-level content optimization, industry tag adaptation, bid evaluation logic modeling and intelligent document verification throughout the entire bidding process, significantly improving the standardization, efficiency and user experience of electronic bidding, and providing a highly intelligent solution for various electronic bidding processes such as government procurement, engineering construction and informatization projects.

[0190] Example 3

[0191] The electronic bidding system provided in this embodiment three is a further optimization and expansion based on the system described in embodiment one. It not only has the ability to digitize, structure, template, and intelligently manage bidding documents, but also enhances the system's ability to adjust bidding content in real time, recommend optimizations, and provide interactive feedback, aiming to achieve more efficient, intelligent, and personalized electronic bidding process management.

[0192] Based on the functional modules described in Embodiment 1, namely: structured content segmentation unit, tag variable definition unit, data binding unit, template storage unit, tender document generation unit, bid document generation unit, and bid evaluation assistance unit, this Embodiment 3 further adds the following key functional units to achieve a higher level of intelligence and user customization services: template management unit, tender document preparation unit, bid document acquisition unit, bid evaluation processing unit, bid opening assistance unit, dynamic template combination unit, terminology expansion unit, and a client-platform collaborative deployment architecture.

[0193] The template management unit supports the standardization and structured output of bidding documents. In its implementation, the template management unit first provides a template creation interface, allowing users (such as template administrators) to issue new template creation commands based on parameters such as industry and project category. The bookmark annotation sub-unit within the template management unit automatically identifies content areas in the document that require user input and inserts uniquely identified bookmark variables (bookmark-id) at the corresponding locations, ensuring a one-to-one correspondence between subsequent information input and XML fields. The XML template generation sub-unit takes over the content after bookmark processing, transforming the document content into a parsable data template with chapter hierarchy, semantic tags, and based on preset XML structure rules. The generated XML structured templates are automatically archived by the material library interface sub-unit and categorized into the system material library, enabling template access, preview, permission settings, and version management for rapid reuse or dynamic combination.

[0194] The tender document preparation unit is an interactive tool for tenderers or project managers. Users fill in key information item by item in a graphical interface, including but not limited to project name, tender number, evaluation criteria, technical parameter requirements, and quotation details. After receiving the user's data, the tender document preparation unit automatically invokes template and variable mapping logic to structure the content and write it into the corresponding XML data fields. The system also integrates a quotation calculation logic subunit, which automatically calculates the tax-inclusive and tax-exclusive prices of each cost item after the user fills in the quotation-related elements (such as quantity, tax rate, and unit price excluding tax), generates a total quotation, and fills it into the XML and the generated Word tender document, ensuring calculation accuracy and form consistency. This unit enables rapid generation of tender documents, structured output, visual preview, and dual output of document and data.

[0195] The tender document acquisition unit is primarily aimed at bidders. Based on the tender XML structure and tender content specifications, the tender document acquisition unit generates a tender interface with navigation guidance, guiding users to fill in the response content item by item.

[0196] Users can also access information from the historical structured XML data reference tag set through this unit. The system automatically loads similar content already submitted in historical projects through path parsing, enabling rapid referencing and information reuse, and reducing the workload of bidding. After the bid is completed, all data will be automatically saved to the corresponding bidder's exclusive XML file, forming a structured bid document and improving the manageability and format standardization of the bid data.

[0197] The bid evaluation processing unit supports expert reviewers in intelligently reviewing, comparing, and scoring multiple bidders' documents during the bid evaluation phase. Within this unit, the review indicator binding subunit first reads the set of review indicator tags from the bidding XML and identifies the binding relationship between scoring items and bid sections. The system then uses the bid information extraction subunit to read response data from multiple bidders' XML documents and centrally extracts the bound field content. Finally, the visualization comparison subunit automatically generates a horizontal comparison data table, covering multiple dimensions such as technical parameters, financial status, and project experience.

[0198] After the evaluation experts score the bids in the system interface, the report generation sub-unit can automatically generate a complete draft of the bid evaluation report, and supports remote signing and result archiving, thereby improving the objectivity, efficiency and traceability of the review work.

[0199] During the bid opening phase, the bid opening assistance unit receives system instructions and calls the corresponding logic to automatically verify, read, and display the bid documents: the system will perform legality verification on the XML structure, including mandatory field checks and signature verification; and display and compare key fields such as quotation content, technical solutions, and implementation plans; realizing online assistance, data protection, and full-process information archiving in the bid opening process, replacing the traditional manual unpacking and manual comparison.

[0200] In this embodiment, the electronic bidding system supports collaborative operation between the client and the platform:

[0201] The client-side component is developed using Windows Forms technology, based on a three-tier architecture: Presentation Layer (UI): responsible for user interaction; Business Logic Layer (BLL): responsible for data processing and business rule execution; Data Access Layer (DAL): responsible for structured XML reading and writing, database access, etc. The platform-side component is based on the Spring Cloud Alibaba microservice architecture, using Vue + Element UI to build the front-end. The back-end services support service discovery, configuration center, unified authentication, workflow engine, task scheduling, log management, and other functions, possessing good scalability, security, and maintainability.

[0202] To accommodate the diverse bidding needs of projects across various industries and types, this embodiment of an electronic bidding system based on XML structured data also incorporates a dynamic template combination mechanism: a rule-engine-based combination logic subunit can automatically select and combine suitable templates according to industry category and project characteristics; a user-defined control subunit provides a graphical selection interface, supporting user-defined chapter structures, scoring indicators, and other content; and a template version control subunit is used to archive, compare, and synchronize various template versions with standards, ensuring the legality and adaptability of the templates used.

[0203] To support the identification and application of professional terms in different fields, this embodiment of an electronic bidding system based on XML structured data also provides a terminology extension unit, including: a tag terminology library management subunit for building an industry terminology library and associating terms with field meanings, units, and contexts; and a plug-in extension subunit for supporting the loading of industry extension packages to enable the system to quickly identify and process specific industry terms.

[0204] It should be noted that Embodiment 3 is not only an enhancement based on Embodiment 1, but it is also a complete independent embodiment with newly added functional units.

[0205] In the description of this invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0206] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0207] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can mean that the first and second features are in direct contact, or that they are in indirect contact through an intermediate medium. Furthermore, "above," "over," or "on top" the second feature can mean that the first feature is directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," or "beneath" the second feature can mean that the first feature is directly below or diagonally below the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0208] In the description of this specification, the terms "one embodiment," "some embodiments," "embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0209] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make modifications, alterations, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. An electronic tendering system based on XML structured data, characterized in that, The system comprises: a structured content division unit configured to divide the bidding content information input by the bidding administrator into a plurality of content segments and convert each content segment into a corresponding XML segment; a tag variable definition unit configured to define a structured tag variable in the XML segment, the structured tag variable being an XML placeholder tag with a semantic identifier and a unique bookmark-id; a data binding unit configured to establish a mapping relationship between the structured tag variable and associated data in a system database or project information input by a user; a template storage unit configured to combine the XML segments selected by the bidding administrator into a complete XML bidding template and store the complete XML bidding template according to industry categories; a bidding document generation unit configured to receive project information input by a user, call a complete XML bidding template matching the project information input by the user according to the project information input by the user, match the project information input by the user with the structured tag variable, and automatically fill the project information input by the user and the associated data in the system database into the corresponding structured tag variable in the matched complete XML bidding template according to the mapping relationship established by the data binding unit, to obtain a complete bidding XML file; a bid document generation unit configured to classify a plurality of project data filled in the complete bidding XML file by a bidder according to the project data required to be filled in the bidding XML file, to obtain bid XML file data; an evaluation assistance unit configured to extract project data in each bid XML file data bound with the structured tag variable based on a path of the project data in the bid XML file data bound with the structured tag variable, and generate a horizontal comparison analysis table containing the project data in each bid XML file data bound with the structured tag variable; wherein the structured content division unit divides the bidding content information input by the bidding administrator into a plurality of content segments and converts each content segment into a corresponding XML segment, and specifically comprises: the structured content division unit is configured to preliminarily segment the bidding content information input by the bidding administrator according to a sentence level, to obtain a plurality of sentence-level content segments; the plurality of sentence-level content segments are combined based on a preset target function, to obtain a plurality of structured and reasonable bidding content segments, each bidding content segment serving as a basic unit for generating an XML segment; wherein: the target function is used to select an optimal combination mode of the plurality of sentence-level content segments, to obtain a content division result with a maximum total correlation score and a minimum redundancy; the target function is: ； represents the optimal set of bidding content segments obtained by optimization, including a set of finally selected bidding content segments; ; is the i-th selected bid content segment of the target function in the optimization process; represents the i-th bidding content segment of the candidate; score the semantic relevance of the i-th candidate bid content segment; score the redundancy for the i-th candidate bid content segment; is a semantic weight; is the redundancy penalty term weight; n represents the total number of selected bidding content segments.

2. The electronic tendering system based on XML structured data according to claim 1, wherein, The system further comprises: an evaluation screening unit configured to call an embedded scoring model to calculate the score of each bid XML file data based on the project data in the bid XML file data bound with the structured tag variable, and select the bid file data with the highest score as a winning candidate.

3. The electronic bidding system based on XML structured data according to claim 2, wherein The bidding content information includes basic information of the project to be bid, bidding scope, technical specifications, qualification requirements, scoring standards, bidding document format requirements, and draft content of contract terms; The project information input by the user includes a specific project name, a project budget, an industry category, an implementation location, a planned construction period, a project leader, and bidding unit information. 4.The electronic bidding system based on XML structured data according to claim 3, characterized in that, wherein semantic relevance score of the i-th candidate bid content segment is calculated by the first approach; the first mode is to use a pre-trained language model to vectorize the i th candidate bidding content segment and its adjacent bidding content segments; the semantic similarity between the i th candidate bidding content segment and its adjacent bidding content segments is calculated based on cosine similarity, and the sum of the semantic similarity between the i th candidate bidding content segment and its adjacent bidding content segments is taken as a first similarity value, and the reciprocal of the first similarity value is taken as the semantic relevance score of the i th candidate bidding content segment; wherein the redundancy score of the i-th candidate bidding content segment is calculated using the second approach; the second mode is: a sentence vector set is constructed for each candidate bidding content segment; the average semantic similarity between the i th candidate bidding content segment and all other candidate bidding content segments is calculated, and the average similarity is taken as the redundancy score of the i th candidate bidding content segment. 5.The electronic bidding system based on XML structured data according to claim 3, characterized in that, a bidding document generation unit receives project information input by a user, and calls a complete XML bidding template matching the project information input by the user according to the project information input by the user; the bidding document generation unit uses formula (1) to obtain the fitness score of the complete XML bidding template for the project information input by the user; ； wherein, a fitness score of the complete XML bid template to the project information entered by the user; m is the total number of specified fields involved in the complete XML bidding template; Industry weight factor for the ith specified field in the complete XML bid template; a description of the i-th item of information input by the user; a structure field definition corresponding to the i-th item of the requirement description in the item information input by the user in the complete XML bidding template; For With Vector cosine similarity of the complete XML bidding template with the highest fitness score for the project information input by the user is selected as the matching complete XML bidding template. 6.The electronic bidding system based on XML structured data according to claim 5, characterized in that, a bidding document generation unit automatically fills the project information input by the user and associated data in the system database into corresponding structured label variables in the matching complete XML bidding template according to the mapping relationship established by the data binding unit to obtain a complete bidding XML file, specifically including: the bidding document generation unit automatically fills the project information input by the user and associated data in the system database into corresponding structured label variables in the matching complete XML bidding template in order of filling priority scores from high to low; wherein the filling priority score is calculated by formula (2); the formula (2) is: ； Fields corresponding to the user-entered item information and associated data in the system database Frequency of occurrence in matched full XML bid templates; are weight parameters; is an adjustable threshold value; the fields corresponding to the user-entered item information and the associated data in the system database a fill priority score. 7.The electronic bidding system based on XML structured data according to claim 5, characterized in that, the various project data filled in by the bidder include: data on the composition of technical personnel of the bidding company, the number of successful cases of the bidding company, the total bid price of the bidding company, the warranty period of the bidding company, the qualification grade of the bidding company, and the registered capital of the bidding company.

8. The electronic bidding system based on XML structured data according to claim 7, characterized in that, the scoring model is: ； wherein ; ； ； wherein a is the number of doctors among the technical personnel of the bidding company; b is the number of masters among the technical personnel of the bidding company; c is the number of bachelors among the technical personnel of the bidding company; ； Number of successful cases for the bidding company; ； ； Total bid price for the bidding company; a lowest bid of the bid interval; a highest bid of the bid interval ； Guarantee period for the bidding company; ； ； a bid company qualification level, wherein the bid company qualification level is A or B or C; ； Register the capital of the bidding company.

9. An electronic tendering method based on XML structured data, characterized in that, the method is executed by the electronic bidding system based on XML structured data according to any one of claims 1-8.

Images