Intelligent data transaction matching method

Abstract

The invention relates to an intelligent data transaction matching method. The method comprises the following steps: constructing a user portrait through personal information and historical behavior record information of a data transaction buyer, and generating a user tag set for describing user features; performing multi-dimensional analysis on each data product on the platform to generate a product label set for describing product features; adopting a natural language processing technology and a machine learning algorithm to calculate the matching degree between the user label and the product label; according to the label matching degree, generating a product recommendation list according to a preset weight algorithm; and dynamically adjusting the recommendation strategy according to user feedback. According to the method, more accurate personalized recommendation is realized by comprehensively analyzing the user tags and the product tags, the accuracy and the user satisfaction of a recommendation system can be remarkably improved, the probability of successful business matching is improved by accurately matching user interests and product characteristics, and compared with a traditional recommendation algorithm, the method has the advantage that the recommendation efficiency is improved. The method has obvious advantages in the aspects of product characteristic identification, recommendation strategy adjustment and the like.

Description

Technical Field

[0001] The present invention belongs to the technical field of data transaction methods, and in particular, relates to an intelligent data transaction matching method and system. Background Art

[0002] With the rapid development of the digital economy, data has become an important resource to promote social progress and economic growth. As an important way to circulate data resources, data trading realizes the information and currency exchange of data elements by classifying, pricing, circulating and buying and selling data as products. However, how to efficiently match the needs of data buyers and data sellers is one of the core issues facing the data trading market.

[0003] Data transaction matching refers to matching data suppliers (sellers) and data demanders (buyers) on data trading platforms through technical means to achieve effective data transactions. This process includes not only data discovery and matching, but also transaction execution, settlement and subsequent services.

[0004] Data transaction matching is of great significance in the data transaction market, including but not limited to improving transaction efficiency, quickly matching supply and demand parties through technical means, shortening transaction cycles, and increasing transaction success rates; optimizing resource allocation, achieving efficient allocation of data resources, and maximizing data value, etc. Therefore, promoting market development, especially promoting the healthy development of the data transaction market and attracting more participants, will help form a virtuous cycle of data transactions.

[0005] The commonly used technologies for data transaction matching include blockchain technology, artificial intelligence technology, big data processing technology, data desensitization technology, etc.

[0006] Among them, blockchain technology plays an important role in data transaction matching, which is mainly reflected in the following aspects: (1) Data confirmation and traceability: Through distributed ledgers, each data transaction record is permanently stored on the blockchain, which can be traced back to the source of the data to ensure the authenticity and integrity of the data. This helps solve the problem of data confirmation and enhances the trust between the two parties to the transaction.

[0007] (2) Smart contracts: Smart contracts can automatically execute transaction terms, reduce human intervention, and improve transaction efficiency. For example, when a buyer successfully pays, the smart contract automatically delivers the data to the buyer and transfers the money to the seller.

[0008] (3) Data privacy protection: Blockchain technology can achieve "data available but invisible", that is, data is calculated and analyzed without leaving its original location, ensuring data privacy and security. For example, through zero-knowledge proof technology, buyers can verify the authenticity of data without accessing the original data.

[0009] Artificial intelligence technology is mainly used for intelligent matching and risk assessment in data transaction matching, which is mainly reflected in the following points: (1) Intelligent matching: Predict market demand through machine learning models, optimize trading strategies, and automatically match supply and demand. For example, the platform can recommend the data products that best meet the buyer's needs based on the buyer's historical transaction records and preferences. Commonly used algorithms include collaborative filtering and deep learning.

[0010] (2) Risk assessment: AI technology can be used to assess the risks of data transactions, identify potential fraud, and improve transaction security. For example, by analyzing the historical behavior of both parties to the transaction, the credibility of the transaction can be assessed. Common technologies include anomaly detection and credit scoring.

[0011] Big data processing technology is mainly used for data cleaning, integration and analysis in data transaction matching. Data cleaning and integration mainly refers to cleaning, integrating and standardizing massive data to improve data consistency and accuracy. This helps to improve data availability and reduce misunderstandings and disputes in transactions. Data mining and analysis refers to extracting valuable information and insights from large amounts of data through machine learning and data mining technology to provide support for decision-making. For example, by analyzing market trends and predicting changes in data demand, sellers can be helped to adjust data products. Commonly used algorithms include cluster analysis, regression analysis, etc.

[0012] Data desensitization technology is mainly used to protect personal privacy and business secrets in data transaction matching. The commonly used technical methods are static desensitization and dynamic desensitization. Static desensitization refers to a one-time processing of data to generate a desensitized copy of the data. For example, sensitive information is replaced with random characters to ensure the security of data during transmission and use. Dynamic desensitization refers to real-time desensitization processing when data is accessed to ensure the security of data during transmission and use. For example, when a buyer queries the data, the system automatically desensitizes the sensitive information.

[0013] In the process of data trading, matching the needs of buyers and sellers is a complex and critical link. In order to improve the accuracy and efficiency of matching, machine learning algorithms, deep learning algorithms, statistical and probability models, search and sorting algorithms and other algorithmic capabilities are widely used to achieve algorithmic implementation of data trading supply and demand matching.

[0014] Machine learning algorithms include collaborative filtering, content-based recommendations, hybrid recommendations and other capabilities.

[0015] Collaborative filtering algorithms analyze users' historical behaviors and preferences to identify user groups with similar behaviors, thereby recommending similar items or data products. They are suitable for situations with rich user behavior data, such as e-commerce platforms and social media. They can capture implicit relationships between users and recommend personalized data products.

[0016] The content-based recommendation algorithm recommends new data products that are similar to those that users liked in the past by analyzing the characteristics of data products and the historical behavior of users. It is suitable for situations where the characteristics of data products are clear, such as news articles and scientific research data. It can explain the reasons for the recommendation and the recommendation results are highly interpretable.

[0017] The hybrid recommendation algorithm combines the advantages of collaborative filtering and content-based recommendation, comprehensively considers user behavior and data product characteristics, improves the accuracy and diversity of recommendations, and is suitable for scenarios with rich user behavior data and data product characteristics. It combines the advantages of multiple recommendation algorithms and has better recommendation effects.

[0018] Deep learning algorithms include deep neural networks, convolutional neural networks, recurrent neural networks and other technologies.

[0019] Deep neural network technology learns the complex relationship between user behavior and data product characteristics through multi-layer neural networks to achieve more accurate matching. It is suitable for large-scale data sets and complex relationship modeling, such as recommendation systems, advertising, etc. It can capture complex nonlinear relationships and has good recommendation effects.

[0020] Convolutional neural networks extract local features of data products through convolutional layers. They are suitable for matching structured data such as images and texts, as well as matching multimedia data such as image data and text data, and can effectively extract local features.

[0021] Recurrent neural networks process sequence data through a cyclic structure and are suitable for matching time series data, such as stock prices, user behavior sequences, etc. They can process sequence data and capture time dependencies.

[0022] Statistical and probabilistic modeling techniques include hidden Markov models, probabilistic graphical models and other techniques.

[0023] The Hidden Markov Model models user behaviors and data product characteristics through hidden states and observed states. It is suitable for matching time series data, suitable for user behavior sequences and time series data, and can model time dependencies.

[0024] The probabilistic graph model models the probabilistic relationship between users and data products through a graph structure, and is suitable for matching complex relationship networks, such as social networks, recommendation systems, and other complex relationship networks. It can model complex relationships.

[0025] Search and sorting algorithms include techniques such as sorting learning and search tree algorithms.

[0026] Ranking learning uses machine learning algorithms to learn users' preferences for data products, sort candidate data products, and recommend top-ranked data products. It is suitable for search engines, recommendation systems, and other scenarios that require sorting. It can perform personalized sorting based on user preferences, and has good recommendation effects.

[0027] The search tree algorithm gradually narrows the search scope by constructing a search tree to find the optimal matching solution. It is suitable for searching and matching large-scale data sets and has high search efficiency.

[0028] In the process of data trading, it is necessary to use a variety of technologies to match the needs of buyers and sellers. Machine learning algorithms and deep learning algorithms can capture the complex relationship between user behavior and data product characteristics, and improve the accuracy and personalization of matching. Statistical and probabilistic models can model time dependencies and complex relationships, and are suitable for time series data and relationship networks. Search and sorting algorithms can improve search efficiency and are suitable for matching large-scale data sets. By comprehensively applying these technologies, the efficiency and success rate of data transactions can be effectively improved.

[0029] In the process of data transactions, although a variety of advanced technologies have been applied to match the needs of buyers and sellers, there are still some technical problems that need to be solved urgently, including data sparsity and cold start problems, lack of in-depth understanding of user needs, data quality and standardization issues, data security and privacy protection, technical barriers and cost issues, etc.

[0030] 1. Data sparsity and cold start problem In the data trading market, the behavioral data of buyers and sellers is often very sparse, especially for new users and new products, which lack sufficient historical data for effective matching. Data sparsity leads to a decrease in the accuracy and coverage of the recommendation system, affecting user experience and transaction success rate. New users and new data products lack historical data, making it difficult for the recommendation system to provide accurate matching suggestions for them. The cold start problem leads to a poor first experience for new users and new products, affecting market activity and user retention.

[0031] 2. Lack of in-depth understanding of user needs Buyers’ needs are often complex and changeable, and existing matching algorithms have limitations in understanding and capturing user needs. Users may use vague or imprecise search terms, resulting in unsatisfactory matching results. Insufficient understanding of needs may lead to poor buyer experience and reduce transaction success rates.

[0032] The data market changes rapidly, and buyers' needs and market conditions may change at any time. Existing matching algorithms are insufficient in terms of real-time and adaptability, and it is difficult to respond to market changes in a timely manner. Dynamic demand changes lead to delayed matching results and missed opportunities for trading.

[0033] 3. Data quality and standardization issues The data quality provided by different data suppliers varies greatly, and there is a lack of unified quality assessment standards and mechanisms. Data quality issues may cause buyers to purchase data products that do not meet their needs, increasing transaction costs and risks. The uneven data quality reduces buyers' willingness to buy and affects the healthy development of the market. The data formats and standards between different data trading platforms are not unified, which makes data exchange and sharing difficult. The lack of unified data standards and interface specifications increases the complexity of data processing and analysis. The lack of unified data standards limits the liquidity and availability of data, affecting the scale and efficiency of the market.

[0034] 4. Data security and privacy protection Data is vulnerable to hacker attacks and improper operations by insiders during transmission and use, leading to data leakage and abuse. Data leakage not only harms the interests of both parties to the transaction, but may also lead to legal proceedings and a crisis of social trust. Although existing data desensitization and encryption technologies can protect data privacy to a certain extent, there are still technical loopholes and attack risks. Especially in the scenarios of large-scale data processing and real-time data transmission, the performance and effect of privacy protection technology need to be improved. Insufficient privacy protection may lead to the leakage of personal privacy and business secrets, affecting users' willingness to use and market development.

[0035] 5. Technical barriers and cost issues The research and development and application of data processing and analysis technologies require high professional knowledge and technical capabilities, which are difficult for small and medium-sized enterprises and individual developers to complete independently. The high technical threshold limits the market participation and innovation vitality. The high technical threshold leads to a single market participant, which is not conducive to market diversification and competition. In addition, the research and development and application of data processing and analysis technologies require high costs, which many small and medium-sized enterprises and individual developers cannot afford. The high technical costs limit the expansion and popularization of the market. The high cost leads to limited market expansion, affecting the scale and activity of the market. Summary of the invention

[0036] In order to solve the above problems existing in the existing data transaction matching methods, the present invention proposes a new intelligent data transaction matching method. This method realizes more accurate personalized recommendations by comprehensively analyzing user tags and product tags, thereby significantly improving the accuracy of the recommendation system and user satisfaction.

[0037] Special note: The present invention aims to provide a data transaction matching method, whose service object is "data transaction". The user referred to in this method refers to the "data transaction buyer" and the product refers to the "data product".

[0038] The main application scenarios of the method of the present invention include: (1) Based on the intelligent matching method of the present invention, intelligent recommendations for multi-person and multi-faceted data products are realized, and matching of user tags and product tags is supported to realize data product recommendations that meet customer purchase needs, such as Figure 1 shown.

[0039] (2) Using the intelligent matching method of the present invention to recommend data products of potential purchase needs of users based on the products currently browsed by the users, such as Figure 2 shown.

[0040] (3) Based on the product description filled in by the user, the product label suitable for the current product is automatically matched when the product is put on the shelf, such as Figure 3 shown.

[0041] Specifically, the present invention provides an intelligent data transaction matching method, such as Figure 5 As shown, the method comprises the following steps: S1. User tag acquisition: Build a user profile through the personal information and historical behavior record information of the data transaction buyer (referred to as the user in this invention), and generate a user tag set that describes the user's characteristics; S2. Product label acquisition: Perform multi-dimensional analysis on each data product (referred to as product in this invention) on the platform to generate a set of product labels that describe product characteristics; S3. Tag matching calculation: Natural language processing technology and machine learning algorithms are used to calculate the matching degree between user tags and product tags; based on the user's search path, people can find products or products can find products; S4. Recommendation result generation: Generate a product recommendation list based on the tag matching degree and the preset weight algorithm; S5. Dynamically adjust recommendation strategy: Dynamically adjust recommendation strategy based on user feedback.

[0042] Furthermore, the historical behavior record information of the data transaction buyer in step S1 of the intelligent data transaction matchmaking method of the present invention includes but is not limited to: user browsing data, search history, click record, and comment information; The multi-dimensional analysis of each data product on the platform described in step S2 includes but is not limited to: analyzing the industry attributes, usage scenarios, product categories, product descriptions, user historical evaluations, and product related information of each data product on the platform.

[0043] Furthermore, the machine learning algorithm in step S3 of the intelligent data transaction matching method of the present invention includes but is not limited to: collaborative filtering algorithm, content-based recommendation algorithm, hybrid recommendation algorithm; The preset weight algorithm in step S4 includes but is not limited to: collaborative filtering algorithm, deep learning algorithm; wherein, the deep learning algorithm includes: deep neural network, convolutional neural network, recurrent neural network.

[0044] Furthermore, the process of obtaining and using user tags and product tags in the intelligent data transaction matching method of the present invention includes the following steps: 1) The data integration module connects to the databases of different business systems and extracts the corresponding business data to the label profiling system; 2) Use the NLP model to perform word segmentation on the extracted business data; 3) Extract keywords from the segmented data according to predefined labeling rules and automatically label system users or products; 4) Conduct tag management and portrait analysis; 5) Provide unified interface access for different business systems through the tag portrait API.

[0045] Furthermore, the databases of different business systems in step 1) of the above-mentioned intelligent data transaction matching method include the following types of data sources: data products, solutions, enterprise entities, transaction orders, search logs, and trial records; The tag management described in step 4) includes but is not limited to: defining tag catalog, tag definition, tag rules, and update strategy; The portrait analysis described in step 4) includes but is not limited to: crowd definition, crowd selection, crowd portrait, and crowd comparison.

[0046] Furthermore, the intelligent data transaction matching method of the present invention also includes a data preprocessing step, which includes data cleaning and data standardization, wherein: Data cleaning: remove noise and erroneous data, fill missing values, and handle outliers; Data standardization: Convert data from different sources into a unified format and standard.

[0047] Furthermore, the intelligent data transaction matching method of the present invention also includes: Real-time data processing by using stream processing technologies such as Apache Kafka, Apache Flink, etc. Dynamically update model parameters periodically by using online learning or incremental learning techniques; Diversity recommendations are made by introducing diversity indicators (such as entropy, coverage, etc.); Use fairness indicators (such as group fairness, individual fairness, etc.) to regularly evaluate the fairness of algorithms; Encrypt sensitive data to ensure data security during transmission and storage; Desensitize user data to protect user privacy; Limit access to sensitive data by enforcing access control mechanisms.

[0048] Furthermore, the intelligent data transaction matching method of the present invention is applied to the following scenarios: (1) Recommend data products that users may need to purchase based on the products they are currently browsing; (2) Based on the product description filled in by the user, the system automatically matches the appropriate product label for the current product when the product is put on the shelf.

[0049] On the other hand, the present invention also provides an intelligent data transaction matching system, which implements the steps of the above-mentioned intelligent data transaction matching method when running.

[0050] In addition, the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the above-mentioned intelligent data transaction matching method are implemented.

[0051] The intelligent data transaction matching method of the present invention has the following characteristics: (1) Enhanced product feature identification capability: In the process of constructing product tags, the present invention not only considers the traditional attributes of data products (such as industry, scenario, etc.), but also includes user comments, product related information, product description, extracted key business information, etc., so as to more accurately understand the characteristics and advantages of each data product. This multi-dimensional product description helps to improve the relevance of recommendations.

[0052] (2) Implementing an efficient tag matching algorithm: The present invention uses advanced natural language processing technology and machine learning models to calculate the matching degree between user tags and product tags. This method can quickly find the best match on a large-scale data set and continuously optimize its performance as the data grows.

[0053] (3) Flexible adjustment of recommendation strategies: Based on the results of tag matching, the present invention can dynamically adjust the recommendation strategy and respond quickly to users’ real-time feedback and behavior changes. This immediacy makes the recommendation system more intelligent and able to adapt to changes in user needs in different scenarios.

[0054] (4) Better explainability: Since the recommendation process is based on clear label matching, when a user asks why a product is recommended to them, the system can provide a clear explanation, such as "You often pay attention to marketing data products, and this product is a highly rated option in this category", thereby increasing the user's trust in the recommendation results.

[0055] (5) Higher recommendation accuracy: The present invention can improve recommendation accuracy while maintaining high user satisfaction. By accurately matching user interests with data product characteristics, irrelevant recommendations are reduced and the probability of successful business matchmaking is increased.

[0056] In summary, the present invention realizes a more personalized recommendation service through tag matching. Compared with traditional recommendation algorithms, it has obvious advantages in product feature identification, recommendation strategy adjustment, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the description of the present invention. Obviously, the following drawings are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without paying creative work.

[0058] Figure 1 This is an example diagram of the first application scenario of the method of the present invention.

[0059] Figure 2 This is an example diagram of a second application scenario of the method of the present invention.

[0060] Figure 3 This is an example diagram of the third application scenario of the method of the present invention.

[0061] Figure 4 The figure is a schematic diagram of the process of obtaining and using user tags and product tags in the method of the present invention.

[0062] Figure 5 It is a schematic diagram of the implementation process of the method of the present invention. DETAILED DESCRIPTION

[0063] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with specific embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments, and the present invention can also be implemented or applied through other different specific implementation methods. The details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention.

[0064] At the same time, it should be understood that the protection scope of the present invention is not limited to the specific embodiments described below; it should also be understood that the terms used in the embodiments of the present invention are for describing specific embodiments rather than for limiting the protection scope of the present invention.

[0065] Embodiment: A smart data transaction matching method The implementation process of this method is as follows: Step 1. Obtain user tags: Build a user profile through the personal information, historical behavior records and other information of the data transaction buyer to form a series of tag sets that describe the user.

[0066] Step 2. Product label acquisition: Perform multi-dimensional analysis on each data product on the platform, including but not limited to the attributes, categories, historical evaluations, etc. of the data products, to form a product label set.

[0067] Step 3. Tag matching calculation: Use natural language processing technology and machine learning models to calculate the matching degree between user tags and product tags. According to the user's search path, the solution of people looking for products or products looking for products is realized.

[0068] Step 4. Generate recommendation results: Generate a recommendation list based on the label matching degree and a certain weight algorithm (such as collaborative filtering, deep learning model, etc.).

[0069] Step 5. Dynamic adjustment mechanism: adjust the recommendation strategy in real time based on user feedback to improve the recommendation effect.

[0070] In the specific implementation, we first need to build a tag database, which includes user tags and product tags. Then, we use appropriate algorithms to process these tags to calculate the correlation between them. Finally, based on the calculation results, we present the user with product recommendations that best match his or her interests.

[0071] The process of obtaining and using user tags and product tags in this method is as follows: Figure 4 As shown, the following steps are included: 1. Data integration: connect to the databases of different business systems (available data sources include but are not limited to data products, solutions, corporate entities, transaction orders, search logs, trial records, etc.), and extract the corresponding business data into the label portrait system.

[0072] 2. Natural language processing large model (NLP model) processing: Use the NLP model to perform word segmentation processing to extract keywords from the data.

[0073] 3. Label calculation / crowd calculation / crowd profiling: Extract keywords from segmented data based on predefined labeling rules and automatically label system users or products.

[0074] 4. Tag management and portrait analysis: Tag management can facilitate system administrators to define tags and tag rules (including but not limited to defining tag directories, tag definitions, tag rules, update strategies, etc.); portrait analysis can view the user's tag distribution from multiple dimensions (including but not limited to crowd definition, crowd selection, crowd portrait, crowd comparison, etc.).

[0075] 5. Tag portrait API: API is the external interface of the tag portrait system, responsible for providing unified interface access to the business system.

[0076] The user operation process of this method is as follows: 1. After the user logs in, the system loads their personal information and historical behavior data from the database.

[0077] 2. The analysis engine extracts user tags based on the above data.

[0078] 3. Extract corresponding labels for each data product in the target product pool.

[0079] 4. Apply matching algorithm to calculate the similarity between user tags and product tags.

[0080] 5. Sort by similarity and select the top N data products as recommendation results to display to users.

[0081] In order to enable the intelligent data transaction matching method to be implemented smoothly, play its vital role in the data transaction process, efficiently and accurately match the needs of buyers and sellers, and improve data transaction efficiency and user satisfaction. The following strategies and technologies are used in the intelligent data transaction matching method of the present invention: 1. Data preprocessing The present invention requires efficient and accurate cleaning and standardization of data to ensure the quality of input data. The measures adopted include data cleaning and data standardization; Data cleaning: remove noise and erroneous data, fill in missing values, and handle outliers.

[0082] Data standardization: Convert data from different sources into a unified format and standard.

[0083] 2. Understanding of user needs The present invention requires a comprehensive and accurate understanding of user needs to improve the degree of personalization of matching. The measures adopted include the use of multimodal data fusion, natural language processing and other technologies; Multimodal data fusion: Combine multiple data sources such as user behavior data, search history, click records, etc.

[0084] Natural Language Processing: Use NLP technology to parse users’ natural language queries and improve the accuracy of demand understanding.

[0085] 3. Model selection and training The implementation of the present invention requires the selection of appropriate machine learning and deep learning algorithms to improve the accuracy and efficiency of matching. The measures adopted include the selection of machine learning algorithms and deep learning algorithms; Machine learning algorithms: Based on the differences in actual data transaction scenarios, select a variety of machine learning algorithms such as collaborative filtering, content-based recommendation, and hybrid recommendation, and combine them with label matching technology.

[0086] Deep learning algorithm: According to the differences in actual data trading scenarios, we select a variety of deep learning algorithms such as deep neural network, convolutional neural network, recurrent neural network, etc. to automatically extract features and make recommendations based on feature matching results.

[0087] 4. Real-time and dynamic adaptability The implementation of the present invention needs to be able to perform real-time data processing and dynamic model updating to improve response speed and adaptability; Real-time data processing: Use stream processing technologies such as Apache Kafka, Apache Flink, etc.

[0088] Dynamic model update: Use online learning, incremental learning and other technologies to regularly update model parameters.

[0089] 5. Diversity and Equity The present invention needs to ensure the diversity and fairness of the recommendation results, and avoid excessive concentration and bias in the results. The measures adopted include diversity recommendation and fairness assessment; Diversity recommendation: Introduce diversity indicators such as entropy and coverage.

[0090] Fairness assessment: Regularly evaluate the fairness of the algorithm, using fairness indicators such as group fairness, individual fairness, etc.

[0091] 6. Security and Privacy Protection The implementation of this invention needs to protect the security and privacy of user data and prevent data leakage and abuse. The measures adopted include data encryption, data desensitization, and access control; Data encryption: Encrypt sensitive data to ensure data security during transmission and storage.

[0092] Data desensitization: Desensitize user data to protect user privacy.

[0093] Access Control: Implement strict access control mechanisms to limit access to sensitive data.

[0094] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Any technician familiar with the profession can make some changes or modifications to the technical contents disclosed above without departing from the scope of the technical solution of the present invention to obtain equivalent embodiments of equivalent changes. However, any simple modification, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention should be included in the scope of protection of the claims of the present invention.

Claims

1. An intelligent data transaction matching method, characterized in that: The method comprises: S1. User tag acquisition: Build a user profile through the personal information and historical behavior record information of the data transaction buyer, and generate a user tag set that describes the user's characteristics; S2. Product label acquisition: Perform multi-dimensional analysis on each data product on the platform to generate a set of product labels that describe product characteristics; S3. Tag matching calculation: Natural language processing technology and machine learning algorithms are used to calculate the matching degree between user tags and product tags; S4. Recommendation result generation: Generate a product recommendation list based on the tag matching degree and the preset weight algorithm; S5. Dynamically adjust recommendation strategy: Dynamically adjust recommendation strategy based on user feedback.

2. The intelligent data transaction matching method according to claim 1 is characterized in that: The historical behavior record information of the data transaction buyer in step S1 includes: user browsing data, search history, click record, and comment information; The multi-dimensional analysis of each data product on the platform described in step S2 includes: analyzing the industry attributes, usage scenarios, product categories, product descriptions, user historical evaluations, and product related information of each data product on the platform.

3. The intelligent data transaction matching method according to claim 1 is characterized in that: The machine learning algorithm in step S3 includes: collaborative filtering algorithm, content-based recommendation algorithm, and hybrid recommendation algorithm; The preset weight algorithms described in step S4 include: collaborative filtering algorithm, deep learning algorithm; wherein, the deep learning algorithm includes: deep neural network, convolutional neural network, recurrent neural network.

4. The intelligent data transaction matching method according to claim 1, characterized in that: The process of obtaining and using user tags and product tags in the intelligent data transaction matching method includes the following steps: 1) The data integration module connects to the databases of different business systems and extracts the corresponding business data to the label profiling system; 2) Use the NLP model to perform word segmentation on the extracted business data; 3) Extract keywords from the segmented data according to predefined labeling rules and automatically label system users or products; 4) Conduct tag management and portrait analysis; 5) Provide unified interface access for different business systems through the tag portrait API.

5. The intelligent data transaction matching method according to claim 4 is characterized in that: The databases of different business systems described in step 1) include the following types of data sources: data products, solutions, enterprise entities, transaction orders, search logs, and trial records; The tag management described in step 4) includes: defining tag catalog, tag definition, tag rules, and update strategy; The portrait analysis described in step 4) includes: crowd definition, crowd selection, crowd portrait, and crowd comparison.

6. The intelligent data transaction matching method according to claim 1, characterized in that: The intelligent data transaction matching method also includes a data preprocessing step, which includes data cleaning and data standardization, wherein: Data cleaning: remove noise and erroneous data, fill missing values, and handle outliers; Data standardization: Convert data from different sources into a unified format and standard.

7. The intelligent data transaction matching method according to claim 1, characterized in that: The intelligent data transaction matching method also includes: Real-time data processing by using stream processing techniques; Regularly and dynamically update model parameters by using online learning or incremental learning techniques; Diversity recommendations are made by introducing diversity indicators; Regularly assess the fairness of algorithms using fairness metrics; Encrypt sensitive data to ensure data security during transmission and storage; Desensitize user data to protect user privacy; Limit access to sensitive data by enforcing access control mechanisms.

8. The intelligent data transaction matching method according to claim 1, characterized in that: The intelligent data transaction matching method is applied to the following scenarios: (1) Recommend data products that users may need to purchase based on the products they are currently browsing; (2) Based on the product description filled in by the user, the system automatically matches the appropriate product label for the current product when the product is put on the shelf.

9. An intelligent data transaction matching system, characterized in that: When the system is running, the steps of the intelligent data transaction matching method described in any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the steps of the intelligent data transaction matching method according to any one of claims 1 to 8.

Images