Commodity recommendation method and system based on large model and block chain

Abstract

The embodiment of the invention discloses a commodity recommendation method and device based on a large model and a block chain. The method comprises the following steps: receiving a circle admission application initiated by a user to a specific circle owner; after entering the circle, receiving commodity information published by the seller and demand information published by the buyer; and carrying out supply and demand intelligent matching by adopting a large model and a block chain technology, and carrying out comprehensive sorting and commodity recommendation on matching results. The invention provides a solution which takes a circle owner as a core driving force, establishes a trust foundation through a block chain technology and realizes intelligent matching by using a large model, and has the advantages that the digital identity certificate serves as an unforgeable chain qualification and provides a solid trust foundation for all transaction behaviors. The owner of the circle is used as a professional identity auditor, so that the identity verification process of the members is greatly simplified, and the professional degree and threshold of the circle are improved. When a large model is recommended, the digital identity reputation of a user is used as an important weighting factor, and buyers and sellers with high reputation and high qualification are preferentially matched.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and specifically to a product recommendation method and system based on large models and blockchain. Background Technology

[0002] While traditional centralized e-commerce platforms aggregate goods and services, they suffer from fragmented communities, difficulties in building trust, and opaque recommendation algorithms. Merchants and consumers passively accept platform recommendations, lacking the ability to proactively build relationships and segment vertical markets. This is especially true in highly specialized or trust-dependent sectors, such as cross-border trade and niche collectibles, where a transaction environment led by experts or authorities, with transparent rules and trustworthy members, is lacking. Summary of the Invention

[0003] In view of the deficiencies of the prior art mentioned in the background, the purpose of this invention is to provide a product recommendation method and system based on a large model and blockchain.

[0004] To achieve the above objectives, in a first aspect, embodiments of the present invention provide a product recommendation method based on a large model and blockchain, comprising:

[0005] The system receives circle access requests from users to specific circle owners, enabling users to enter the circle; the circle owner is responsible for creating the circle and the circle governance rules, and the circle governance rules are encoded in a smart contract.

[0006] After entering the community, you can receive product information posted by seller users and demand information posted by buyer users.

[0007] The system uses large-scale models and blockchain technology to intelligently match the product and demand information, and then sorts and recommends products based on the matching results.

[0008] As one specific implementation of this application, the user entering the circle is as follows:

[0009] Receive decentralized tokens created by users through DApps;

[0010] Receive digital identity credential applications submitted by users to a specific community leader, so that the community leader can verify the user's qualifications offline or online based on the digital identity credential applications;

[0011] After successful verification, the community leader invokes the smart contract DIDContract to issue a digital identity credential to the applicant's DID address; the smart contract DIDContract is used to generate, manage, and verify digital identities.

[0012] Meanwhile, applicants are allowed to join the circle based on the circle governance rules; the circle governance rules include a member approval mechanism, content posting standards, and dispute arbitration procedures.

[0013] As a specific implementation of this application, when using a large model for intelligent supply and demand matching, the digital identity credentials and reputation records of buyer and seller users are read, and combined with the semantic information of the seller's product description for comprehensive sorting and product recommendation; the digital identity credentials and reputation records are all stored on the blockchain.

[0014] Secondly, embodiments of the present invention also provide a product recommendation system based on a large model and blockchain, including a processor, an input device, an output device, and a memory, wherein the processor, input device, output device, and memory are interconnected, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions to execute the method described in the first aspect.

[0015] Thirdly, embodiments of the present invention also provide another product recommendation system based on large-scale models and blockchain, including:

[0016] The application layer is used to provide a user interface, based on which seller users / buyer users can apply for, view and manage digital identity credentials, as well as the circle owner can issue digital identity credentials.

[0017] The middle layer is used to cooperate with the community leader through smart contracts to issue digital identity credentials, verify identity, and record on-chain behavior, and generate on-chain data based on the on-chain behavior; the on-chain data includes buyer users and their reputation scores.

[0018] The service layer is used for:

[0019] Obtain on-chain data based on a large language model;

[0020] Receive product information posted by sellers and demand information posted by buyers;

[0021] Based on the on-chain data, intelligent supply and demand matching is performed on the product information and demand information, and the matching results are comprehensively sorted and product recommendations are made.

[0022] The circle owner is responsible for creating the circle and its governance rules, which are encoded in a smart contract. The circle owner issues digital identity credentials specifically as follows:

[0023] Receive decentralized tokens created by users through DApps;

[0024] Receive digital identity credential applications submitted by users to a specific community leader, so that the community leader can verify the user's qualifications offline or online based on the digital identity credential applications;

[0025] After successful verification, the community leader invokes the smart contract DIDContract to issue a digital identity credential to the applicant's DID address; the smart contract DIDContract is used to generate, manage, and verify digital identities; the digital identity credential is stored on the blockchain.

[0026] As one specific implementation of this application, the service layer includes:

[0027] The intent recognition module is used to perform deep intent understanding of buyer needs and obtain buyer intent;

[0028] The multidimensional semantic matching module is used to perform multidimensional semantic matching based on the buyer's intent and the seller's product information to obtain the matching result.

[0029] The reputation weight fusion module is used to obtain buyer weight and seller credibility, and combine the matching results to perform reputation weight matching and risk adjustment;

[0030] The intelligent fusion module is used to calculate the final score of intelligent matching based on the result of the reputation weight fusion module, and to sort and recommend products based on the final score.

[0031] Implementing embodiments of the present invention provides a solution that uses community leaders as the core driving force, establishes a trust foundation through blockchain technology, and utilizes a large model to achieve intelligent matching. This solution has the following advantages:

[0032] A strengthened trust system: Digital identity credentials, as unforgeable on-chain credentials, provide a solid trust foundation for all transactions.

[0033] Efficient identity verification: As a professional identity verifier, the circle owner greatly simplifies the identity verification process for members, thereby increasing the professionalism and entry barrier of the circle.

[0034] User data autonomy: Users control their own digital identity and can selectively disclose some identity information to different circles or trading parties to protect their personal privacy.

[0035] Precise matching and reputation weighting: When making recommendations, the large model uses the user's digital identity reputation as an important weighting factor, prioritizing the matching of buyers and sellers with high reputation and strong qualifications. Attached Figure Description

[0036] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below.

[0037] Figure 1 This is a flowchart of a product recommendation method based on a large model and blockchain provided in an embodiment of the present invention;

[0038] Figure 2 This is a structural diagram of a product recommendation device based on a large model and blockchain provided in an embodiment of the present invention;

[0039] Figure 3 yes Figure 3 Another structural diagram of the device shown;

[0040] Figure 4 This is a structural diagram of the main modules of the recommendation algorithm;

[0041] Figure 5 This is the algorithm flowchart for the intent recognition module;

[0042] Figure 6 This is the algorithm flowchart for the multidimensional semantic matching module;

[0043] Figure 7 This is the algorithm flowchart for the reputation weight fusion module;

[0044] Figure 8 This is the algorithm flowchart for the intelligent fusion module. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0047] The embodiments of the present invention specifically address the following problems:

[0048] Lack of trust: By leveraging the authority of the circle leader and the transparency of the blockchain, trust among members can be quickly established and strengthened within a specific circle.

[0049] Market segmentation is challenging: empowering community leaders to build vertical market segments, accurately aggregating merchants and buyers with similar needs, interests, or professional backgrounds.

[0050] Information asymmetry: By leveraging the circle leader and a large language model, we ensure high-quality and transparent information within the circle, breaking down traditional information barriers.

[0051] Decentralized ecosystem: Decentralizing operational power to community leaders to form a decentralized business network consisting of multiple micro-ecosystems.

[0052] To solve the above problems, the main technical means adopted by the present invention include the following four aspects:

[0053] 1. Circle creation and governance based on smart contracts:

[0054] Authorized users (circle owners) are allowed to create "circles" with specific themes or entry thresholds on the blockchain by calling "factory contracts".

[0055] The community's governance rules (such as member approval mechanisms, content publishing guidelines, and dispute arbitration processes) are encoded in smart contracts, ensuring the transparency and enforceability of the rules.

[0056] Governance authority can be held by the community leader, or community co-governance can be achieved through a voting mechanism based on reputation scores.

[0057] 2. Digital reputation system based on distributed identity (DID):

[0058] The system creates a DID for each participant as their unique identity on the blockchain.

[0059] User behavior data (such as transaction history and received reviews) is associated with their DID in the form of verifiable credentials (VC) and is under the user's control.

[0060] Deploy a "reputation management contract" that automatically and dynamically calculates and updates a user's "digital reputation score" based on verifiable on-chain behavior (such as successful fulfillment of obligations and receiving positive reviews). This reputation score is publicly verifiable and becomes the user's most important intangible asset within the system.

[0061] 3. Publish goods and services within the community based on smart contracts:

[0062] The entire business process is solidified and automatically executed through a set of collaborative smart contracts (including contracts for member management, supply and demand posting, transaction processes, reputation management, etc.).

[0063] From users joining the circle, posting supply and demand, confirming the match, completing the transaction, to finally updating the reputation score, a complete, trustworthy, and automated on-chain closed loop is formed.

[0064] 4. Use large-scale models and blockchain technology for intelligent matching:

[0065] LLM is deployed in a secure off-chain computing environment and communicates with the blockchain network through a decentralized oracle or an authorized secure API.

[0066] When a match is required, the user authorizes the oracle to access their anonymized demand data and on-chain metadata (such as reputation score).

[0067] The core task of LLM is to perform "intent matching," deeply understand the true intent of user needs, and combine the semantic information of the seller's product description with the reputation scores of both parties to perform comprehensive ranking and recommendation.

[0068] Please refer to Figure 1 This invention provides a product recommendation method based on a large model and blockchain, which may include the following steps:

[0069] S1, the creation of a user's digital identity and the application for digital identity credentials.

[0070] In practice, users create a decentralized identifier (DID) through a DApp. Further, users submit an application to a specific community leader, requesting the issuance of digital identity credentials related to their identity or qualifications.

[0071] In S2, the circle owner verifies the user's identity, applies for approval, and issues a digital identity certificate, which the user then uses to enter the circle.

[0072] The circle owner, acting as the authoritative reviewer, verifies applicants' qualifications offline or online. Upon successful verification, the circle owner invokes the DIDContract to issue a digital identity credential to the applicant's DID address. This credential serves as an unforgeable on-chain record. The circle owner can set circle governance rules within the CircleContract, such as "only merchants with ISO 9001 certified digital identities are allowed to join." These governance rules, including member approval mechanisms, content publishing guidelines, and dispute arbitration processes, are encoded in a smart contract, ensuring transparency and enforceability.

[0073] It should be noted that:

[0074] DIDContract: Used to generate, manage, and verify digital identities.

[0075] CircleContract: The circle owner defines the circle rules and acts as the issuer of digital identities.

[0076] CreditScoreContract: Records on-chain behavior based on digital identity to generate a reputation score.

[0077] S3, after entering the circle, shows product information posted by seller users and demand information posted by buyer users.

[0078] S4 uses large-scale models and blockchain technology to intelligently match the supply and demand information of the goods and services, and performs comprehensive sorting and product recommendation on the matching results.

[0079] In practice, a large model is used for intelligent matching of supply and demand. The digital identity credentials and reputation records of buyer and seller users are read, and the semantic information of the seller's product description is combined to perform comprehensive sorting and product recommendation. The digital identity credentials and reputation records are stored on the blockchain.

[0080] For example, for a large purchase, the big model will prioritize recommending companies with "Advanced Supplier Certification" digital identities and high reputation scores, thereby improving the success rate and security of the transaction.

[0081] It should be noted that the intelligent matching process of supply and demand includes user intent recognition, multi-dimensional semantic matching, reputation weight fusion and intelligent fusion, etc. The specific process will be described in detail in subsequent system implementations, and will not be repeated here.

[0082] In S5, after both parties are matched, they communicate about the transaction, and the entire transaction process is recorded on the blockchain.

[0083] As can be seen from the above description, this invention provides a solution that uses community owners as the core driving force, establishes a trust foundation through blockchain technology, and utilizes a large model to achieve intelligent matching. This solution has the following advantages:

[0084] A strengthened trust system: Digital identity credentials, as unforgeable on-chain credentials, provide a solid trust foundation for all transactions.

[0085] Efficient identity verification: As a professional identity verifier, the circle owner greatly simplifies the identity verification process for members, thereby increasing the professionalism and entry barrier of the circle.

[0086] User data autonomy: Users control their own digital identity and can selectively disclose some identity information to different circles or trading parties to protect their personal privacy.

[0087] Precise matching and reputation weighting: When making recommendations, the large model uses the user's digital identity reputation as an important weighting factor, prioritizing the matching of buyers and sellers with high reputation and strong qualifications.

[0088] Furthermore, by deeply integrating digital identity with the community ecosystem, this invention provides a more solid foundation of trust for future business models.

[0089] Decentralized B2B platforms: Based on a trusted digital identity and reputation system, businesses can confidently find partners globally.

[0090] Professional services market: Experts and freelancers can use their qualifications (such as lawyer's licenses, designer certificates) as digital identity credentials to gain clients' trust within their circles.

[0091] Finance and Insurance: Financial institutions can leverage on-chain digital identities and credit records to provide more precise and lower-risk trade finance and insurance services to businesses within the ecosystem.

[0092] Based on the same inventive concept, embodiments of the present invention also provide a product recommendation device based on a large model and blockchain. For example... Figure 2 As shown, the product recommendation device may include one or more processors 101, one or more input devices 102, one or more output devices 103, and a memory 104. The processors 101, input devices 102, output devices 103, and memory 104 are interconnected via a bus 105. The memory 104 stores a computer program, which includes program instructions. The processor 101 is configured to invoke the program instructions to execute the method described in the above-described method embodiment.

[0093] It should be understood that, in this embodiment of the invention, the processor 101 may be a central processing unit (CPU), but it may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0094] Input device 102 may include a keyboard, etc., and output device 103 may include a display (LCD, etc.), a speaker, etc.

[0095] The memory 104 may include read-only memory and random access memory, and provides instructions and data to the processor 101. A portion of the memory 104 may also include non-volatile random access memory. For example, the memory 104 may also store device type information.

[0096] In specific implementations, the processor 101, input device 102, and output device 103 described in the embodiments of the present invention can execute the implementation methods described in the embodiments of the product recommendation method based on large models and blockchain provided in the embodiments of the present invention, which will not be repeated here.

[0097] Please refer to this again. Figure 3 This invention also provides another product recommendation system based on large-scale models and blockchain, including:

[0098] The application layer is used to provide a user interface, based on which seller users / buyer users can apply for, view and manage digital identity credentials, as well as the circle owner can issue digital identity credentials.

[0099] The middle layer is used to cooperate with the community leader through smart contracts to issue digital identity credentials, verify identity, and record on-chain behavior, and generate on-chain data based on the on-chain behavior; the on-chain data includes buyer users and their reputation scores.

[0100] The service layer is used for:

[0101] Obtain on-chain data based on a large language model;

[0102] Receive product information posted by sellers and demand information posted by buyers;

[0103] Based on the on-chain data, intelligent supply and demand matching is performed on the product information and demand information, and the matching results are comprehensively sorted and product recommendations are made.

[0104] The circle owner is responsible for creating the circle and its governance rules, which are encoded in a smart contract. The circle owner issues digital identity credentials specifically as follows:

[0105] Receive decentralized tokens created by users through DApps;

[0106] Receive digital identity credential applications submitted by users to a specific community leader, so that the community leader can verify the user's qualifications offline or online based on the digital identity credential applications;

[0107] After successful verification, the community leader invokes the smart contract DIDContract to issue a digital identity credential to the applicant's DID address; the smart contract DIDContract is used to generate, manage, and verify digital identities; the digital identity credential is stored on the blockchain.

[0108] It should be noted that the service layer is based on, for example, Figure 4 The four recommendation algorithm modules shown are implemented as follows:

[0109] The intent recognition module is used to perform deep intent understanding of buyer needs and obtain buyer intent;

[0110] The multidimensional semantic matching module is used to perform multidimensional semantic matching based on the buyer's intent and the seller's product information to obtain the matching result.

[0111] The reputation weight fusion module is used to obtain buyer weight and seller credibility, and combine the matching results to perform reputation weight matching and risk adjustment;

[0112] The intelligent fusion module is used to calculate the final score of intelligent matching based on the result of the reputation weight fusion module, and to sort and recommend products based on the final score.

[0113] like Figure 5 As shown, the intent recognition module is specifically used to perform text preprocessing, entity recognition, intent classification and vectorization on the buyer's needs to obtain the buyer's intent.

[0114] The text preprocessing includes: (1) text cleaning: noise character removal; (2) word segmentation: Jieba / BERT word segmentation; (3) part-of-speech tagging: POS tagging; (4) stop word filtering; (5) text normalization. For example, input: “I need a high-performance laptop”; output: [“need”, “high-performance”, “laptop”].

[0115] Entity recognition is implemented based on the NER algorithm, using a BiLSTM-CRF model. Entity type recognition includes: product category, attribute features, and numerical range. For example, the output is: {Product, Laptop, Attribute, Performance, Value: High} with a confidence score of 0.95.

[0116] Intent classification is implemented using a BERT-based classifier; intent categories include purchase intent, comparison intent, and inquiry intent. For example, the output might be: Intent = BUY, Probability = 0.89, Urgency: Medium.

[0117] Vectorization is achieved using a Transformer encoder, with an output dimension of 768. The feature identifiers include semantic features and contextual features.

[0118] like Figure 6 As shown, the multidimensional semantic matching module includes cosine similarity calculation, semantic distance calculation, structural similarity calculation, and comprehensive score calculation.

[0119] like Figure 7As shown, the buyer weighting component of the reputation weighting fusion module uses a weighting function based on the buyer's reputation score and historical behavior; the seller credibility component is based on a credibility model, with evaluation dimensions including transaction success rate, timely fulfillment, and product quality rating; the matching weighting component is based on adjustment factors, including community activity, transaction frequency, and relationship intimacy, with the matching weight W_match = buyer W_buyer × seller credibility T_seller. The risk adjustment component involves risk assessment factors including abnormal behavior detection, fraud risk scoring, and the impact of market fluctuations.

[0120] like Figure 8 As shown, the intelligent fusion module is specifically used for multi-factor fusion, dynamic weight adjustment, personalized ranking, and confidence assessment. Among them, the polyphonic character fusion is implemented based on the default semantic weight, reputation weight, and context weight; the dynamic weight adjustment is based on user behavior feedback learning, reinforcement learning optimization, and test verification to adjust the default weights mentioned above; the personalized ranking is implemented based on personalized factors, which include user preference history, purchase behavior patterns, and social relationship networks.

[0121] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices or units, or may be electrical, mechanical or other forms of connection.

[0122] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of the present invention, depending on actual needs.

[0123] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0124] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0125] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A product recommendation method based on a large model and blockchain, characterized in that, include: Receive circle access requests from users to specific circle owners, so that users can enter the circle; The circle owner is used to create circles and circle governance rules, and the circle governance rules are encoded in a smart contract; After entering the community, you can receive product information posted by seller users and demand information posted by buyer users. The system uses large-scale models and blockchain technology to intelligently match the product and demand information, and then sorts and recommends products based on the matching results.

2. The product recommendation method as described in claim 1, characterized in that, The specific steps for a user to enter the circle are as follows: Receive decentralized tokens created by users through DApps; Receive digital identity credential applications submitted by users to a specific community leader, so that the community leader can verify the user's qualifications offline or online based on the digital identity credential applications; After successful verification, the community leader invokes the smart contract DIDContract to issue a digital identity credential to the applicant's DID address; the smart contract DIDContract is used to generate, manage, and verify digital identities. Meanwhile, applicants are allowed to join the circle based on the circle governance rules; the circle governance rules include a member approval mechanism, content posting standards, and dispute arbitration procedures.

3. The product recommendation method as described in claim 1, characterized in that, When using a large model for intelligent supply and demand matching, the digital identity credentials and reputation records of buyer and seller users are read, and combined with the semantic information of the seller's product description for comprehensive sorting and product recommendation; the digital identity credentials and reputation records are all stored on the blockchain.

4. A product recommendation system based on a large model and blockchain, comprising a processor, an input device, an output device, and a memory, wherein the processor, input device, output device, and memory are interconnected, characterized in that, The memory is used to store a computer program, the computer program including program instructions, and the processor is configured to invoke the program instructions to perform the method as described in any one of claims 1-3.

5. A product recommendation system based on large-scale models and blockchain, characterized in that, include: The application layer is used to provide a user interface, based on which seller users / buyer users can apply for, view and manage digital identity credentials, as well as the circle owner can issue digital identity credentials. The middle layer is used to cooperate with the community leader through smart contracts to issue digital identity credentials, verify identity, and record on-chain behavior, and generate on-chain data based on the on-chain behavior; the on-chain data includes buyer users and their reputation scores. The service layer is used for: Obtain on-chain data based on a large language model; Receive product information posted by sellers and demand information posted by buyers; Based on the on-chain data, intelligent supply and demand matching is performed on the product information and demand information, and the matching results are comprehensively sorted and product recommendations are made.

6. The product recommendation system as described in claim 5, characterized in that, The circle owner is responsible for creating the circle and establishing circle governance rules, which are encoded in a smart contract; the circle owner issues digital identity credentials specifically as follows: Receive decentralized tokens created by users through DApps; Receive digital identity credential applications submitted by users to a specific community leader, so that the community leader can verify the user's qualifications offline or online based on the digital identity credential applications; After successful verification, the community leader invokes the smart contract DIDContract to issue a digital identity credential to the applicant's DID address; the smart contract DIDContract is used to generate, manage, and verify digital identities; the digital identity credential is stored on the blockchain.

7. The product recommendation system as described in claim 5, characterized in that, The service layer includes: The intent recognition module is used to perform deep intent understanding of buyer needs and obtain buyer intent; The multidimensional semantic matching module is used to perform multidimensional semantic matching based on the buyer's intent and the seller's product information to obtain the matching result. The reputation weight fusion module is used to obtain buyer weight and seller credibility, and combine the matching results to perform reputation weight matching and risk adjustment; The intelligent fusion module is used to calculate the final score of intelligent matching based on the result of the reputation weight fusion module, and to sort and recommend products based on the final score.

8. The product recommendation system as described in claim 7, characterized in that, The intent recognition module is specifically used for: The buyer's needs are preprocessed with text, entity recognition, intent classification, and vectorization to obtain the buyer's intent. The text preprocessing includes noise character removal, BERT word segmentation, POS part-of-speech tagging, stop word filtering, and text standardization. The intent classification results in intent categories including purchase intent, comparison intent, and consultation intent.

9. The product recommendation system as described in claim 7, characterized in that, The multidimensional semantic matching includes cosine similarity calculation, semantic distance calculation, structural similarity calculation, and comprehensive score calculation.

10. The product recommendation system as described in claim 7, characterized in that, The intelligent fusion module is specifically used for multi-factor fusion, dynamic weight adjustment, personalized ranking, and confidence assessment. Among them, the polyphonic character fusion is based on the default semantic weight, reputation weight, and context weight; the dynamic weight adjustment is based on user behavior feedback learning, reinforcement learning optimization, and test verification to adjust the default weights mentioned above; the personalized ranking is based on personalized factors, which include user preference history, purchase behavior patterns, and social relationship networks.

Images