Intelligent transaction matching method and system for commodity service based on large model

Abstract

The application discloses a commodity service intelligent transaction matching method and system based on a large model, relates to the technical field of intelligent information processing, and comprises the following steps: collecting multi-source transaction data of demand parties and supply parties and preprocessing the multi-source transaction data to generate structured transaction data; based on the structured transaction data, using a large model to perform semantic analysis on demand information and supply information, and constructing corresponding demand portraits and supply portraits; based on the demand portraits and the supply portraits, performing candidate recall through semantic matching, category matching and association expansion modes to generate a candidate matching object set; filtering the candidate matching object set according to preset transaction constraint rules, and comprehensively scoring and sorting the remaining candidate objects based on multi-dimensional matching indexes. The application introduces a large model to perform deep semantic analysis on demand information and supply information and generate a transaction condition adjustment strategy, thereby improving the matching accuracy of transaction matching and the self-adaptive optimization capability of transaction conditions.

Description

Technical Field

[0001] This invention relates to the field of intelligent information processing technology, and in particular to a method and system for intelligent transaction matching of goods and services based on a large model. Background Technology

[0002] With the continuous development of e-commerce platforms and digital service systems, the transaction of goods and services is gradually evolving from the traditional manual matching model to an intelligent matching model. In existing technologies, common transaction matching methods mainly rely on matching mechanisms based on rule engines and recommendation algorithms, such as keyword retrieval, category filtering, tag matching, or collaborative filtering algorithms based on historical behavior, to initially match demanders and suppliers. However, existing technologies are still primarily based on structured data processing and rule-driven approaches, with limited semantic understanding capabilities, making it difficult to fully explore the deep relationships between demand and supply. Furthermore, there is still significant room for improvement in the ability to process unstructured information in complex transaction scenarios.

[0003] While existing technologies can facilitate transaction matching to some extent, they still have limitations when dealing with multi-source heterogeneous data and complex transaction conditions. Especially when there are inconsistencies in the information expressed by the demand and supply sides, strong semantic implicitness, or multi-dimensional constraints, traditional keyword or tag-based matching methods struggle to accurately understand transaction intent, thus affecting the recall quality of candidate objects and subsequent ranking effects. Furthermore, existing technologies rely heavily on fixed rules or static parameters in determining transaction conditions, lacking the ability to dynamically analyze and adjust for differences between supply and demand, making it difficult to establish optimal transaction conditions in certain scenarios. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a smart transaction matching method for goods and services based on a large model, which solves the problems of insufficient semantic understanding ability and difficulty in dynamically optimizing transaction conditions in existing transaction matching methods.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a method for intelligent transaction matching of goods and services based on a large model, comprising, Collect and preprocess multi-source transaction data from both the demand and supply sides to generate structured transaction data. Based on the structured transaction data, use a large model to perform semantic parsing of demand and supply information and construct corresponding demand and supply profiles. Based on demand and supply profiles, candidate retrieval is performed through semantic matching, category matching, and association expansion to generate a set of candidate matching objects. The candidate matching object set is filtered according to the preset transaction constraint rules, and the remaining candidate objects are comprehensively scored and ranked based on multi-dimensional matching indicators. Based on the ranking results, a large model is used to generate a trading condition adjustment strategy. Based on the trading condition adjustment strategy, trading conditions are determined, and trading orders are generated.

[0007] As a preferred embodiment of the intelligent transaction matching method for goods and services based on a large model as described in this invention, the generation of structured transaction data includes the following steps: Collect demand information, supply information and related transaction data of commodities, integrate the collected data and preprocess it to generate structured transaction data.

[0008] As a preferred embodiment of the intelligent transaction matching method for goods and services based on a large model as described in this invention, the step of using a large model to perform semantic parsing of demand and supply information and constructing corresponding demand and supply profiles includes the following steps: Based on structured transaction data, text data and attribute fields are extracted from demand and supply information to construct semantic input items; The semantic input items are fed into the pre-trained large model and semantic parsing is performed to obtain the transaction object category, attribute features and constraints, and the parsing results are transformed into corresponding structured features; Based on the structured features, demand information and supply information are aggregated to generate demand profiles and supply profiles.

[0009] As a preferred embodiment of the intelligent transaction matching method for goods and services based on a large model described in this invention, the step of generating a candidate matching object set through candidate retrieval via semantic matching, category matching, and association expansion includes the following steps: Based on demand profiles and supply profiles, semantic features, category features, and association features are extracted and uniformly encoded to construct demand-side matching features and supply-side matching features.

[0010] Calculate the semantic similarity between demand-side matching features and supply-side matching features, and filter supply objects that meet the conditions according to the preset similarity threshold to generate a semantic matching candidate set.

[0011] Based on the category characteristics of the demand profile, the supply profile is filtered by category, and supply objects with consistent or similar categories are selected to generate a category matching candidate set.

[0012] Based on the association features of the demand profile, the supply objects are expanded and recalled through preset association rules to obtain supply objects that are related to the demand objects and generate an extended candidate set of associations.

[0013] The semantic matching candidate set, category matching candidate set, and association extension candidate set are merged and deduplicated to generate a candidate matching object set.

[0014] As a preferred embodiment of the intelligent transaction matching method for goods and services based on a large model as described in this invention, the step of filtering the candidate matching object set according to preset transaction constraint rules and comprehensively scoring and ranking the remaining candidate objects based on multi-dimensional matching indicators includes the following steps: Based on the candidate matching object set, the candidate objects are filtered according to the preset transaction constraint rules to remove the supply objects that do not meet the budget range, delivery time limit, geographical restrictions and qualification requirements, and obtain the initial candidate set; Extract semantic matching features, price matching features, timeliness matching features, and fulfillment capability features of each candidate object in the initial screening candidate set, and construct a multi-dimensional matching index system; The candidates are comprehensively scored based on a multi-dimensional matching index system, and then sorted according to the scoring results to generate a sorted list of candidate matching objects.

[0015] As a preferred embodiment of the intelligent transaction matching method for goods and services based on a large model according to the present invention, the step of generating a transaction condition adjustment strategy using a large model based on the ranking results includes the following steps: Based on the ranking results, at least one candidate object is selected, and the corresponding demand information, supply information, matching score results and transaction constraint information of the candidate object are extracted to construct the strategy generation input items; The strategy-generated input items are fed into a pre-trained large model to analyze the price differences, quantity differences, delivery time differences, and service condition differences between the demand side and the supply side, and obtain the corresponding condition difference results.

[0016] Based on the conditional differences, a pre-trained large model is used to generate trading condition adjustment suggestions for candidate objects. The generated trading condition adjustment suggestions are validated according to rules, and adjustment suggestions that do not meet the preset trading constraint rules are eliminated, and trading condition adjustment strategies are generated.

[0017] As a preferred embodiment of the intelligent transaction matching method for goods and services based on a large model as described in this invention, the step of determining transaction conditions based on a transaction condition adjustment strategy and generating a transaction order includes the following steps: Based on the transaction condition adjustment strategy, the corresponding price parameters, delivery time parameters, quantity parameters and service condition parameters are extracted, the original transaction conditions are updated, and the target transaction conditions are generated. Perform consistency verification on the target transaction conditions, and generate the corresponding transaction order based on the target transaction conditions that pass the verification.

[0018] Secondly, this invention provides a large-scale intelligent transaction matching system for goods and services, comprising: The multi-source data acquisition module collects multi-source transaction data from both the demand and supply sides and preprocesses it to generate structured transaction data. The profile building module uses a large model to perform semantic parsing of demand and supply information, and builds corresponding demand profiles and supply profiles. The candidate recall module performs candidate object recall based on demand profile and supply profile, filters out candidate objects that meet the conditions from the supply objects, and generates a set of candidate matching objects. The scoring and ranking module filters the candidate matching object set, removes objects that do not meet the transaction conditions, and performs comprehensive scoring and ranking of the remaining candidate objects based on multi-dimensional matching indicators. The adjustment strategy generation module, based on the ranking results, calls the large model to analyze the differences in transaction conditions between the demand side and the supply side, and generates corresponding transaction condition adjustment strategies. The transaction execution module determines the final transaction conditions based on the transaction condition adjustment strategy and generates transaction orders according to the determined transaction conditions.

[0019] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the intelligent transaction matching method for goods and services based on a large model as described in the first aspect of the present invention.

[0020] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the intelligent transaction matching method for goods and services based on a large model as described in the first aspect of the present invention.

[0021] The beneficial effects of this invention are as follows: By introducing a large model to perform semantic analysis of demand and supply information and construct demand and supply profiles, a deep semantic understanding of transaction intentions and constraints is achieved, improving the accuracy and consistency of supply and demand matching; through a candidate recall method combining semantic matching, category matching, and association expansion, multi-path and multi-dimensional candidate object acquisition is achieved, expanding the matching coverage and enhancing the ability to discover potential matching relationships; by combining preset transaction constraint rules and multi-dimensional matching indicators for filtering and comprehensive scoring and ranking, priority determination of candidate objects is achieved, improving the rationality and executability of the matching results; by using a large model to generate transaction condition adjustment strategies based on the ranking results, dynamic analysis of the differences between supply and demand and adaptive optimization of transaction conditions are achieved, improving the intelligence level, decision-making flexibility, and transaction conversion capability of transaction matching. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart of a smart transaction matching method for goods and services based on a large model.

[0024] Figure 2 This is a schematic diagram of a smart transaction matching system for goods and services based on a large model.

[0025] Figure 3 Build flowcharts for demand profiles and supply profiles.

[0026] Figure 4 Flowchart for candidate recall and ranking. Detailed Implementation

[0027] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0028] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0029] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0030] Reference Figures 1-4 This is one embodiment of the present invention, which provides a smart transaction matching method for goods and services based on a large model, including the following steps: Collect and preprocess multi-source transaction data from both the demand and supply sides to generate structured transaction data. Based on this structured transaction data, use a large model to perform semantic parsing of demand and supply information and construct corresponding demand and supply profiles.

[0031] Specifically, the process involves collecting demand and supply information for goods, as well as relevant transaction data, integrating the collected data, preprocessing it, and generating structured transaction data.

[0032] By collecting and integrating commodity demand information, supply information, and related transaction data, and preprocessing them, unified aggregation and standardized processing of multi-source heterogeneous data are achieved. This transforms the original, scattered, and inconsistent data into structured transaction data with clear structure and unified expression, thereby improving data availability and processing efficiency.

[0033] Furthermore, based on structured transaction data, text data and attribute fields are extracted from demand and supply information to construct semantic input items; The semantic input items are fed into the pre-trained large model and semantic parsing is performed to obtain the transaction object category, attribute features and constraints, and the parsing results are transformed into corresponding structured features; Based on the structured features, demand information and supply information are aggregated to generate demand profiles and supply profiles.

[0034] By extracting textual data and attribute fields from demand and supply information based on structured transaction data and constructing semantic input items, a unified expression and semantic modeling preparation for multi-source information is achieved, providing a standardized input foundation for deep semantic analysis. By inputting these semantic input items into a pre-trained large model for semantic parsing, the categories, attribute features, and constraints of transaction objects are obtained and transformed into structured features, enabling a deep semantic understanding of transaction intentions and key elements, allowing implicit information to be explicitly expressed. Furthermore, by aggregating the structured features to generate demand and supply profiles, a systematic integration and unified characterization of the characteristics of both supply and demand sides is achieved, improving the consistency and discriminative ability of subsequent matching processes.

[0035] Based on demand and supply profiles, candidate retrieval is performed through semantic matching, category matching, and association expansion to generate a set of candidate matching objects.

[0036] Specifically, based on demand profiles and supply profiles, semantic features, category features, and association features are extracted and uniformly encoded to construct demand-side matching features and supply-side matching features.

[0037] Calculate the semantic similarity between demand-side matching features and supply-side matching features, and filter supply objects that meet the conditions according to the preset similarity threshold to generate a semantic matching candidate set.

[0038] Based on the category characteristics of the demand profile, the supply profile is filtered by category, and supply objects with consistent or similar categories are selected to generate a category matching candidate set.

[0039] Based on the association features of the demand profile, the supply objects are expanded and recalled through preset association rules to obtain supply objects that are related to the demand objects and generate an extended candidate set of associations.

[0040] The semantic matching candidate set, category matching candidate set, and association extension candidate set are merged and deduplicated to generate a candidate matching object set.

[0041] By extracting semantic, category, and association features from demand and supply profiles and performing unified encoding, a unified expression and comparability of multi-dimensional features from both supply and demand sides is achieved, providing a standardized feature foundation for matching calculations. By calculating the semantic similarity of matching features between the demand and supply sides and generating a semantic matching candidate set, accurate identification of semantic-level matching relationships between supply and demand is achieved, effectively capturing potential matching objects. Category features are used for category filtering and generating a category matching candidate set, realizing structured constraint filtering based on a business classification system and improving the rationality of matching results. Association features and association rules are used for extended recall, generating an extended association candidate set, enabling the mining and supplementation of implicit associations, and better discovering potential but not explicitly matched supply objects. By merging and deduplicating multi-source candidate sets to form a candidate matching object set, unified integration and optimization of multi-path recall results are achieved, improving candidate coverage and result quality.

[0042] The candidate matching object set is filtered according to the preset transaction constraint rules, and the remaining candidate objects are comprehensively scored and ranked based on multi-dimensional matching indicators.

[0043] Specifically, based on the candidate matching object set, the candidate objects are filtered according to the preset transaction constraint rules to eliminate the supply objects that do not meet the budget range, delivery time limit, geographical restrictions and qualification requirements, and obtain the initial candidate set; Extract semantic matching features, price matching features, timeliness matching features, and fulfillment capability features of each candidate object in the initial screening candidate set, and construct a multi-dimensional matching index system; The feature indicators in the multidimensional matching index system are normalized, and a comprehensive score is calculated for each candidate object, as shown in the following formula: ; in, Index for candidate matching objects, For the first The overall score of each candidate matchmaker For the first Semantic matching feature values ​​of candidate matching objects For the first Price matching feature values ​​of candidate matching objects For the first The timeliness matching feature values ​​of each candidate matching object For the first The performance capability characteristic value of each candidate matching object. , , , These are the weights of semantic matching features, price matching features, timeliness matching features, and fulfillment capability features, which can be dynamically adjusted according to different transaction scenarios.

[0044] Sort the candidates according to the scoring results and generate a sorted list of matchmaking targets.

[0045] By filtering the candidate matching object set according to preset transaction constraint rules, the system effectively eliminates supply objects that do not meet the requirements of budget range, delivery time limit, geographical restrictions, and qualification requirements, ensuring that the candidate objects in the subsequent processing have basic transaction feasibility. By extracting the semantic matching features, price matching features, timeliness matching features, and performance capability features of each candidate object in the initial screening candidate set and constructing a multi-dimensional matching index system, the system achieves a systematic and quantitative expression of the multi-dimensional attributes of the candidate objects. By performing comprehensive scoring calculation and ranking based on the multi-dimensional matching index system, a ranked candidate matching object list is generated, which realizes reasonable differentiation and optimal selection of candidate object priorities, improving the accuracy of the matching results and the scientific nature of decision-making.

[0046] Based on the ranking results, a large model is used to generate a trading condition adjustment strategy. Based on the trading condition adjustment strategy, trading conditions are determined, and trading orders are generated.

[0047] Specifically, based on the ranking results, at least one candidate object is selected, and the corresponding demand information, supply information, matching score results and transaction constraint information of the candidate object are extracted to construct the strategy generation input items; The strategy-generated input items are fed into a pre-trained large model to analyze the price differences, quantity differences, delivery time differences, and service condition differences between the demand side and the supply side, and obtain the corresponding condition difference results.

[0048] Based on the conditional differences, a pre-trained large model is used to generate trading condition adjustment suggestions for candidate objects. The generated trading condition adjustment suggestions are validated according to rules, and adjustment suggestions that do not meet the preset trading constraint rules are eliminated, and trading condition adjustment strategies are generated.

[0049] By selecting candidate objects based on the ranking results and extracting corresponding demand information, supply information, matching score results, and transaction constraint information to construct strategy generation input items, the integration and unified input expression of key transaction elements are achieved. The strategy generation input items are then fed into a pre-trained large-scale model to analyze the differences between supply and demand sides in terms of price, quantity, delivery time, and service conditions, achieving precise identification and structured expression of supply-demand mismatch factors. The large-scale model generates transaction condition adjustment suggestions based on the condition difference results, enabling dynamic optimization and flexible adjustment of transaction conditions, making the generated solutions more targeted and adaptable. By performing rule verification on the transaction condition adjustment suggestions and eliminating content that does not conform to the constraint rules, the final transaction condition adjustment strategy is generated, achieving controllability and executability of the strategy output, improving the transaction condition optimization capability, and enhancing the effectiveness of matching decisions.

[0050] Furthermore, based on the transaction condition adjustment strategy, the corresponding price parameters, delivery time parameters, quantity parameters, and service condition parameters are extracted to update the original transaction conditions and generate the target transaction conditions. Perform consistency verification on the target transaction conditions, and generate the corresponding transaction order based on the target transaction conditions that pass the verification.

[0051] By extracting price, delivery time, quantity, and service condition parameters based on the strategy of adjusting transaction conditions and updating the original transaction conditions, a systematic reconstruction and optimization of key transaction elements is achieved, making the transaction conditions more closely match the actual matching status of supply and demand. By performing consistency verification on the target transaction conditions, the rationality and compliance with constraints of the transaction conditions are effectively verified, preventing non-compliant conditions from entering the execution stage. Based on the verified target transaction conditions, corresponding transaction orders are generated, realizing an effective transformation from strategy output to actual transaction execution, enabling the matching results to be applied to real transaction scenarios.

[0052] This embodiment also provides a large-scale intelligent transaction matching system for goods and services, including: The multi-source data acquisition module collects multi-source transaction data from both the demand and supply sides and preprocesses it to generate structured transaction data. The profile building module uses a large model to perform semantic parsing of demand and supply information, and builds corresponding demand profiles and supply profiles. The candidate recall module performs candidate object recall based on demand profile and supply profile, filters out candidate objects that meet the conditions from the supply objects, and generates a set of candidate matching objects. The scoring and ranking module filters the candidate matching object set, removes objects that do not meet the transaction conditions, and performs comprehensive scoring and ranking of the remaining candidate objects based on multi-dimensional matching indicators. The adjustment strategy generation module, based on the ranking results, calls the large model to analyze the differences in transaction conditions between the demand side and the supply side, and generates corresponding transaction condition adjustment strategies. The transaction execution module determines the final transaction conditions based on the transaction condition adjustment strategy and generates transaction orders according to the determined transaction conditions.

[0053] This embodiment also provides a computer device applicable to the intelligent transaction matching method for goods and services based on a large model, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the intelligent transaction matching method for goods and services based on a large model as proposed in the above embodiment.

[0054] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0055] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the intelligent transaction matching method for goods and services based on a large model as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0056] In summary, this invention achieves a deep semantic understanding of transaction intentions and constraints by introducing a large model to semantically analyze demand and supply information and construct demand and supply profiles, thereby improving the accuracy and consistency of supply-demand matching. Through a candidate recall method combining semantic matching, category matching, and association expansion, it achieves multi-path, multi-dimensional candidate object acquisition, expanding the matching coverage and enhancing the ability to discover potential matching relationships. By combining preset transaction constraint rules with multi-dimensional matching indicators for filtering and comprehensive scoring and ranking, it achieves priority determination of candidate objects, improving the rationality and executability of the matching results. Finally, by using the large model to generate transaction condition adjustment strategies based on the ranking results, it achieves dynamic analysis of the differences between supply and demand and adaptive optimization of transaction conditions, improving the intelligence level, decision-making flexibility, and transaction conversion rate of transaction matching.

[0057] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for intelligent transaction matching of goods and services based on a large model, characterized in that: include, Collect and preprocess multi-source transaction data from both the demand and supply sides to generate structured transaction data. Based on the structured transaction data, use a large model to perform semantic parsing of demand and supply information and construct corresponding demand and supply profiles. Based on demand and supply profiles, candidate retrieval is performed through semantic matching, category matching, and association expansion to generate a set of candidate matching objects. The candidate matching object set is filtered according to the preset transaction constraint rules, and the remaining candidate objects are comprehensively scored and ranked based on multi-dimensional matching indicators. Based on the ranking results, a large model is used to generate a trading condition adjustment strategy. Based on the trading condition adjustment strategy, trading conditions are determined, and trading orders are generated.

2. The intelligent transaction matching method for goods and services based on a large model as described in claim 1, characterized in that: The generation of structured transaction data includes the following steps: Collect demand information, supply information and related transaction data of commodities, integrate the collected data and preprocess it to generate structured transaction data.

3. The intelligent transaction matching method for goods and services based on a large model as described in claim 2, characterized in that: The process of using a large model to perform semantic parsing of demand and supply information and constructing corresponding demand and supply profiles includes the following steps: Based on structured transaction data, text data and attribute fields are extracted from demand and supply information to construct semantic input items; The semantic input items are fed into the pre-trained large model and semantic parsing is performed to obtain the transaction object category, attribute features and constraints, and the parsing results are transformed into corresponding structured features; Based on the structured features, demand information and supply information are aggregated to generate demand profiles and supply profiles.

4. The intelligent transaction matching method for goods and services based on a large model as described in claim 3, characterized in that: The process of retrieving candidates and generating a set of candidate matching objects through semantic matching, category matching, and association expansion includes the following steps: Based on demand profiles and supply profiles, semantic features, category features and association features are extracted and uniformly encoded to construct demand-side matching features and supply-side matching features. Calculate the semantic similarity between demand-side matching features and supply-side matching features, and filter supply objects that meet the conditions according to the preset similarity threshold to generate a semantic matching candidate set; Based on the category characteristics of the demand profile, the supply profile is filtered by category, and supply objects with consistent or similar categories are selected to generate a category matching candidate set. Based on the association features of the demand profile, the supply objects are expanded and recalled through preset association rules to obtain supply objects that are related to the demand objects and generate an association expansion candidate set. The semantic matching candidate set, category matching candidate set, and association extension candidate set are merged and deduplicated to generate a candidate matching object set.

5. The intelligent transaction matching method for goods and services based on a large model as described in claim 4, characterized in that: The process of filtering the candidate matching object set according to preset transaction constraint rules and comprehensively scoring and ranking the remaining candidate objects based on multi-dimensional matching indicators includes the following steps: Based on the candidate matching object set, the candidate objects are filtered according to the preset transaction constraint rules to remove the supply objects that do not meet the budget range, delivery time limit, geographical restrictions and qualification requirements, and obtain the initial candidate set; Extract semantic matching features, price matching features, timeliness matching features, and fulfillment capability features of each candidate object in the initial screening candidate set, and construct a multi-dimensional matching index system; The candidates are comprehensively scored based on a multi-dimensional matching index system, and then sorted according to the scoring results to generate a sorted list of candidate matching objects.

6. The intelligent transaction matching method for goods and services based on a large model as described in claim 5, characterized in that: The process of generating a trading condition adjustment strategy based on the ranking results using a large model includes the following steps: Based on the ranking results, at least one candidate object is selected, and the corresponding demand information, supply information, matching score results and transaction constraint information of the candidate object are extracted to construct the strategy generation input items; The strategy-generated input items are fed into a pre-trained large model to analyze the price differences, quantity differences, delivery time differences, and service condition differences between the demand side and the supply side, and obtain the corresponding condition difference results. Based on the conditional differences, a pre-trained large model is used to generate trading condition adjustment suggestions for candidate objects. The generated trading condition adjustment suggestions are validated according to rules, and adjustment suggestions that do not meet the preset trading constraint rules are eliminated, and trading condition adjustment strategies are generated.

7. The intelligent transaction matching method for goods and services based on a large model as described in claim 6, characterized in that: The method of determining trading conditions and generating trading orders based on the trading condition adjustment strategy includes the following steps: Based on the transaction condition adjustment strategy, the corresponding price parameters, delivery time parameters, quantity parameters and service condition parameters are extracted, the original transaction conditions are updated, and the target transaction conditions are generated. Perform consistency verification on the target transaction conditions, and generate the corresponding transaction order based on the target transaction conditions that pass the verification.

8. A large-scale model-based intelligent transaction matching system for goods and services, based on the large-scale model-based intelligent transaction matching method for goods and services as described in any one of claims 1 to 7, characterized in that: include, The multi-source data acquisition module collects multi-source transaction data from both the demand and supply sides and preprocesses it to generate structured transaction data. The profile building module uses a large model to perform semantic parsing of demand and supply information, and builds corresponding demand profiles and supply profiles. The candidate recall module performs candidate object recall based on demand profile and supply profile, filters out candidate objects that meet the conditions from the supply objects, and generates a set of candidate matching objects. The scoring and ranking module filters the candidate matching object set, removes objects that do not meet the transaction conditions, and performs comprehensive scoring and ranking of the remaining candidate objects based on multi-dimensional matching indicators. The adjustment strategy generation module, based on the ranking results, calls the large model to analyze the differences in transaction conditions between the demand side and the supply side, and generates corresponding transaction condition adjustment strategies. The transaction execution module determines the final transaction conditions based on the transaction condition adjustment strategy and generates transaction orders according to the determined transaction conditions.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the intelligent transaction matching method for goods and services based on a large model as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the intelligent transaction matching method for goods and services based on a large model as described in any one of claims 1 to 7.

Images