Semantic analysis of article cites using deep learning approaches

Semantic publishing; the use of web and semantic web technologies to improve the meaning of a published journal article, facilitate its automatic discovery, connect to semantically related articles, provide processable access to data within the article, and facilitate the integrity of data between articles. Semantic web technologies and information display enhance information restructuring, sharing in a structured and machine-readable way. It shares the goal of creating semantic web technologies and intelligent artificial objects that imitate human intelligence via computer, such as deep learning, reasoning, validation and prediction. The aim of this thesis is to develop a deep learning based search method in citation network that provides the advantages of both semantic web technologies and deep learning applications by using deep learning applications together with citation ontology developed using semantic web technologies. The citation ontology developed within the scope of the study provides information about citations to journal articles. Citation ontology was created using the Protégé ontology development editor. The protégé ontology development editor's graphical interface allows visual identification of citation issues, so that the desired area can be modeled. In the semantic analysis of the citation network proposed in the thesis study, the system developed by using deep learning methods found more similar journal articles in the matching process than the exact matches, and the effects of the proposed method to increase the journal article matching performance were determined.