Research And Implementation Of Relevance-Oriented Patent Retrieval System

Modern society has entered an era of digitization,informatization,and intelligence.Technological innovation is a huge driving force for social progress.Therefore,it is also very important for enterprises and universities to protect their intellectual property rights effectively and rationally while continuously improving their scientific and technological strength.As an intangible property,patent intellectual property covers a large amount of scientific and technological information.Whether it is the research and development of new products or the writing and application of patents,the search of patent data is inseparable.With the continuous development of technology,the patent database is becoming larger and larger.How to quickly and accurately find out the patent information that meets the needs of searchers has become a hot research direction.On the one hand,the patent search system can help patent applicants obtain information related to their own research and development technology in advance,whether it infringes other patent owners,on the other hand,it can help analyze the latest developments in the current industry,institution or regional technology development.The following technical development direction provides certain reference information.This paper proposes a patent search system based on relevance ranking in the field of patent search analysis.The entire system includes five modules: the data processing module includes two parts of data processing: search term processing based on named entity recognition and relational database direction The process of importing search engine data;the retrieval statistics module is the main part of the system,used to complete the search,relevance ranking and patent recommendation functions,and returns the patent data information and patent recommendation data related to the search;the intellectual property situation analysis report module is Based on the previous search sorting rules and aggregation statistics function,the data is sorted and analyzed to reflect the development of intellectual property rights in the current institutions and regions;other modules include favorites module,patent keyword module,laws and regulations modules,among which the favorites module is used for Temporarily save patents that searchers are interested in,providing a batch download function of patents;the patent keyword module is used to extract keywords that can summarize the main content of the patent;the laws and regulations module provides search and viewing of laws and regulations related to intellectual property rights.As the main module of the patent search system,the search statistics module is also divided into seven parts: the patent data storage part builds a set of search engines for patent search based on the captured patent data and patent search rules and is used to output detailed patent information Relational database;named entity recognition section proposes a named entity recognition method based on the Bi-LSTM-CC model to search for word entities to improve the accuracy of retrieval;the input and retrieval generation section is used to detect and obtain user input Search keywords and search methods;search and relevance ranking module is used to search patent data according to the search formula and re-rank the candidate results based on relevance score;the recommended part is to extract the technical features of retrieval based on user input and click information,based on sentence concept map Judging the similarity to screen out recommended patents;the aggregation part is to analyze the intellectual property situation report of an institution or region based on the authorization and disclosure of patents;the output module is used to output recommended search keywords and search results to users.