Extraction of physical effects based on the semantic analysis of the patent texts

Abstract

The paper represents new methodology of semantic analysis for physical effects extracting. This methodology is based on the Tuzov ontology that formally describes the Russian language. In this paper, semantic patterns were described to extract structural physical information in the form of physical effects. A new algorithm of text analysis was described. The approach is applied to the database of physical effects and to the patent texts. The results of the proposed method compared with the results of the IOFEE system that is used for the same tasks. The method described in this article allowed increasing efficiency of the physical effect elements extracting. The semantic analyzer based on the Tuzov ontology was created to increase the accuracy and completeness of the method.