Structured-information extraction from patent-claim sentences

作者:

Highlights:

摘要

This paper presents a method of structured-information extraction from patent-claim sentences. A patent-claim sentence is converted to the internal form of a case structure by a semi-automatic parsing. The internal form is normalized and the specified information is efficiently extracted by referring to a specification table given every special branch. The specification table consists of several subframes which specify the items to be extracted in a standard case structure. The extracted information is stored in relational tables for relational retrieval and some precision-oriented retrieval using role-indicators. The feasibility and the effectiveness of this method are examined by a computer simulation.

论文关键词:

论文评审过程:Received 15 October 1981, Available online 18 July 2002.

论文官网地址:https://doi.org/10.1016/0306-4573(82)90046-2