| With the rapid development of text semantic understanding and text analysis technology in natural language processing,it has been used by many enterprises and institutions,and gradually combined with artificial intelligence technology.However,in the fields of patents or specialized text semantic understanding of specific disciplines,the number of applications is still scarce.In the production environment,limited by data scale and complexity of models and algorithms,patent classifing and comparing system using artificial intelligence also lags behind the development of the times because of technical resons.In China,the problems encountered by patent examiners in retrieval,verification and content analysis in the process of patent examination have eliminated the obstacles to the combination of big data technology and artificial intelligence technology.Combined with some innovations in data mining algorithm and big data architecture,this paper has solved the problems of literature analysis,content evaluation,missing parts evaluation.Many problems such as performance optimization greatly improve the efficiency of patent examination and make the results of patent retrieval more accurate.The paper first explores the implementation of the system from the direction of algorithm and model,then analyzes patent text comparison and automatic patent classification technologies such as word vector technology,and puts forward new algorithms and models such as vector index of text,hypothesis probability,molecular sieve big data filtering model and secondary verification model of automatic text classification combined with big data scenario,So that the automatic patent classification and patent comparison technology based on neural network can be applied in the big data scenario and replace manual.After the separation of these technologies and methods,they can also be used for reference by other professional fields or other large-scale and super big data scenarios. |
PDF Full Text Request