Intelligent retrieval method for international business laws and regulations

Abstract

The invention provides an international business regulation intelligent retrieval method, which relates to the technical field of international business regulation retrieval, and specifically comprises the following steps: collecting and processing data, constructing a multi-language international regulation retrieval model, inputting a retrieval problem, and extracting key information. According to the method, the multi-language international regulation retrieval model is constructed, so that quick retrieval of cross-language and cross-regulation systems can be realized, the cumbersome process that a user looks up regulations one by one is avoided, and the user experience is improved. And the query intention of the user is understood by utilizing a natural language processing technology, the accuracy and comprehensiveness of the retrieval result are improved by combining a multi-language query expansion technology, and the user is helped to better understand and apply laws and regulations by displaying the retrieval result in a manner which is easy to understand by the user and providing auxiliary information such as explanation and cases of laws and regulations.

Description

Technical Field

[0001] The present invention relates to the technical field of international business law retrieval, and particularly to an intelligent retrieval method for international business laws. Background Art

[0002] With the acceleration of the globalization process, international business activities are becoming increasingly frequent, and enterprises need to comply with the laws and regulations of different countries and regions. However, the number of international business laws is huge, updated frequently, and involves different languages and legal systems. Traditional keyword retrieval methods are difficult to meet the needs of enterprises to obtain legal information efficiently and accurately.

[0003] For a traditional law retrieval method with the existing application number: 202211653939.9, after obtaining the text vector of the query statement input by the user, it determines the corresponding law library type of the query statement according to the text vector, so as to match laws in the corresponding law library, greatly reducing the calculation amount. Then, it obtains the first law list according to the semantic vector in the law library, further reducing the calculation amount of calculating the matching similarity in the follow-up. After that, it calculates the matching score of the first law according to the classification similarity, semantic similarity, and matching similarity, and comprehensively sorts the first law from three dimensions: law library classification, semantic vector, and law matching, so that the final sorting result is accurate, excluding unnecessary laws and meeting the user's needs, greatly improving the accuracy of law matching.

[0004] During the use of the above device, it does not have a data fusion function and can only perform small query retrievals on legal documents within a single region. When it is necessary to retrieve laws and regulations in different regions simultaneously, it is time-consuming and laborious, and it is easy to miss key information. Therefore, we propose an intelligent retrieval method for international business laws to solve the above problems. Summary of the Invention

[0005] The problem to be solved by the present invention is that when it is necessary to retrieve laws and regulations in different regions simultaneously, the existing retrieval methods require multiple retrievals, which are time-consuming and laborious, and it is easy to miss key information.

[0006] To solve the above technical problems, the present invention provides an intelligent retrieval method for international business laws. The specific steps of the intelligent retrieval method are as follows: S1: Data collection and processing; S2: Construct a multilingual international law retrieval model; S3: Input a retrieval question, extract key information, and convert it into a standardized retrieval request; S4: Intelligent retrieval algorithm design; S5: Retrieval result sorting and display; S6: Evaluation and optimization; Preferably, the data collection and processing specifically include: S101: Collect regulatory data sources, and collect international business regulatory text data from the United Nations Conference on Trade and Development, the World Trade Organization, government legal websites of various countries, and professional legal databases. S102: Data cleaning, removing duplicate, incorrect, and incomplete data from the collected data set, and standardizing the data to make the data set in a unified text format for subsequent analysis.

[0007] Preferably, the construction of the multilingual international regulation retrieval model specifically includes: S201: Feature extraction, extracting key feature words that can represent the content of regulations through algorithms such as word frequency statistics and TF-IDF.

[0008] Preferably, the input of the retrieval question, extraction of key information, and conversion into a standardized retrieval request specifically include: S301: Receive the query statement input by the user, and use natural language processing technology to identify the user's query intention and key information. S302: According to the user's query intention and key information, use the multilingual international regulation retrieval model for query expansion to generate a multilingual query expression. S303: Use the multilingual query expression to perform retrieval in the multilingual international regulation retrieval model.

[0009] Preferably, the intelligent retrieval algorithm design specifically includes: S401: Boolean retrieval, which supports the user to input keywords, combines retrieval conditions through logical operators such as AND, OR, and NOT to achieve exact matching. S402: Fuzzy retrieval, which uses algorithms such as the edit distance algorithm to retrieve words similar to the keyword. S403: Semantic retrieval, which uses word vector models in natural language processing technology to understand the semantics of the user's input and retrieve regulations related to the semantics.

[0010] Preferably, the sorting and display of the retrieval results specifically include: S501: Result display, displaying the retrieval results in an intuitive and clear list manner, including the regulation name, release time, release agency, abstract of relevant articles, etc., and clicking can view the detailed content. S502: Sorting function: Sort the results according to factors such as the relevance to the retrieval term and the importance of the regulations, facilitating the user to quickly obtain useful information. At the same time, a regulation comparison function is also provided to help the user compare the similarities and differences between different regulations and provide a reference for decision-making.

[0011] Preferably, the evaluation and optimization specifically include: S601: Evaluation metrics, which evaluate the retrieval performance using metrics such as accuracy and recall. Accuracy is the proportion of correct results retrieved out of the total retrieved results, and recall is the proportion of relevant results retrieved out of all relevant results. S602: Method optimization, which adjusts the feature extraction method, retrieval algorithm parameters, etc. according to the evaluation results to continuously improve the retrieval effect and user experience.

[0012] Technical effects and advantages of the present invention: The present invention constructs a multilingual international regulations retrieval model to achieve fast retrieval across languages and regulatory systems, avoiding the cumbersome process of users checking regulations item by item. Moreover, it uses natural language processing technology to understand the user's query intention and combines multilingual query expansion technology to improve the accuracy and comprehensiveness of retrieval results. By presenting the retrieval results in a way that is easy for users to understand and providing auxiliary information such as interpretations of regulatory articles and cases, it helps users better understand and apply regulations. Description of the drawings

[0013] Figure 1 It is a schematic diagram of the overall structure of the present invention. Detailed implementation manners

[0014] The present invention provides an intelligent retrieval method for international business regulations, as Figure 1 shown. The specific steps of the intelligent retrieval method are as follows: S1: Data collection and processing; S101: Collect regulation data sources, and collect international business regulation text data from the United Nations Conference on Trade and Development, the World Trade Organization, government legal websites of various countries, and professional legal databases. S102: Data cleaning, removing duplicate, incorrect, and incomplete data from the collected data set, and standardizing the data to make the data set in a unified text format for subsequent analysis.

[0015] S2: Construct a multilingual international regulations retrieval model; S201: Feature extraction, extracting key feature words that can represent the content of regulations through algorithms such as word frequency statistics and TF-IDF. S202: Index construction, establishing an index for the feature words and information such as the positions of the regulation texts, just like the catalog index in a library, to facilitate quick positioning of relevant regulation content.

[0016] S3: Input a retrieval question, extract key information, and convert it into a standardized retrieval request; S301: Receive the query statement input by the user, and use natural language processing technology to identify the user's query intention and key information. S302: According to the user's query intention and key information, use a multilingual international regulations retrieval model to perform query expansion and generate a multilingual query expression; S303: Use the multilingual query expression to perform a retrieval in the multilingual international regulations retrieval model.

[0017] S4: Intelligent retrieval algorithm design; S401: Boolean retrieval, which supports the user to input keywords and combines retrieval conditions through logical operators such as AND, OR, and NOT to achieve exact matching; S402: Fuzzy retrieval, which uses algorithms such as the edit distance algorithm to retrieve words similar to the keyword; S403: Semantic retrieval, which uses word vector models in natural language processing technology to understand the semantics of the user's input and retrieve regulations related to the semantics.

[0018] S5: Retrieval result sorting and display; S501: Result display, which displays the retrieval results in an intuitive and clear list form, including the name of the regulation, the release time, the release agency, the abstract of relevant articles, etc., and clicking can view the detailed content; S502: Sorting function: Sort the results according to factors such as the relevance to the retrieval term and the importance of the regulation, so as to facilitate the user to quickly obtain useful information. At the same time, a regulation comparison function is also provided to help the user compare the similarities and differences between different regulations and provide a reference for decision-making.

[0019] S6: Evaluation and optimization; S601: Evaluation metrics, which use metrics such as accuracy and recall to evaluate the retrieval performance. Accuracy is the proportion of correct results retrieved to the total retrieval results, and recall is the proportion of relevant results retrieved to all relevant results; S602: Method optimization, which adjusts the feature extraction method, retrieval algorithm parameters, etc. according to the evaluation results to continuously improve the retrieval effect and user experience.

[0020] Example: Suppose the user needs to query "Regulations on import car tariffs in China", and the specific execution process is as follows: Input the query statement into a pre-constructed multilingual regulations retrieval model that contains international business regulations of countries and regions such as China, the United States, and the European Union; Receive the query statement "Regulations on import car tariffs in China" input by the user, and use natural language processing technology to identify that the user's query intention is "query tariff regulations", and the key information is "China" and "import car"; According to the user's query intention and key information, query expansion is performed using the multilingual regulatory knowledge graph to generate multilingual query expressions, such as "China tariff on imported cars", "People's Republic of China, imported cars, tariff"; Use the multilingual query expression to search in the multilingual regulatory knowledge graph to find the provisions regarding the tariff on imported cars; Arrange the retrieved regulatory articles in reverse order of the effective time, and highlight the parts related to the user's query keywords and display them to the user.

[0021] It can be understood that the present invention is described through some embodiments. Those skilled in the art know that, without departing from the spirit and scope of the present invention, various changes or equivalent replacements can be made to these features and embodiments. Additionally, under the teaching of the present invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application belong to the scope protected by the present invention.

Claims

1. An intelligent retrieval method for international business regulations, characterized in that: The specific steps of the intelligent retrieval method are as follows: S1: Data collection and processing; S2: Construct a multilingual international regulations retrieval model; S3: Input the retrieval question, extract key information, and convert it into a standardized retrieval request; S4: Design of intelligent retrieval algorithms; S5: Sorting and display of retrieval results; S6: Evaluation and optimization.

2. The intelligent retrieval method for international business regulations according to claim 1, characterized in that: The data collection and processing specifically include: S101: Collect regulation data sources, and collect international business regulation text data from the United Nations Conference on Trade and Development, the World Trade Organization, the legal websites of various governments, and professional legal databases; S102: Data cleaning, remove duplicate, incorrect, and incomplete data from the collected data set, and perform standardized processing on the data to make the data set into a unified text format for subsequent analysis.

3. The intelligent retrieval method for international business regulations according to claim 1, characterized in that: The construction of the multilingual international regulations retrieval model specifically includes: S201: Feature extraction, extract key feature words that can represent the content of regulations through algorithms such as word frequency statistics and TF-IDF; S202: Index construction, establish an index for the position and other information of the feature words and the regulation text, just like the catalog index of a library, to facilitate quick positioning of relevant regulation content.

4. The intelligent retrieval method for international business regulations according to claim 1, characterized in that: The input of the retrieval question, extraction of key information, and conversion into a standardized retrieval request specifically include: S301: Receive the query statement input by the user, and use natural language processing technology to identify the user's query intention and key information; S302: According to the user's query intention and key information, use the multilingual international regulations retrieval model for query expansion to generate a multilingual query expression; S303: Use the multilingual query expression to perform retrieval in the multilingual international regulations retrieval model.

5. The intelligent retrieval method for international business regulations according to claim 1, characterized in that: The design of the intelligent retrieval algorithms specifically includes: S401: Boolean retrieval, which supports the user to input keywords, and combines retrieval conditions through logical operators such as AND, OR, and NOT to achieve exact matching; S402: Fuzzy retrieval, which uses algorithms such as the edit distance algorithm to retrieve words similar to the keyword; S403: Semantic retrieval, which uses word vector models and other technologies in natural language processing to understand the semantics of the user's input and retrieve regulations related in semantics.

6. The intelligent retrieval method for international business regulations according to claim 1, characterized in that: The sorting and display of the retrieval results specifically include: S501: Result display, display the retrieval results in an intuitive and clear list form, including the regulation name, release time, release agency, abstract of relevant articles, etc., and click to view the detailed content; S502: Sorting function: Sort the results according to factors such as relevance to the retrieval term and the importance of the regulations, so as to facilitate the user to quickly obtain useful information. At the same time, a regulation comparison function is also provided to help the user compare the similarities and differences between different regulations and provide reference for decision-making.

7. An intelligent retrieval method for international business regulations according to claim 1, characterized in that: The evaluation and optimization specifically include: S601: Evaluation indicators, use indicators such as accuracy and recall to evaluate the retrieval performance. Accuracy is the proportion of correct results retrieved to the total retrieval results, and recall is the proportion of relevant results retrieved to all relevant results; S602: Method optimization, according to the evaluation results, adjust the feature extraction method, retrieval algorithm parameters, etc., and continuously improve the retrieval effect and user experience.

Images