Semantic citation for paper correlation using recurrent neural networks.

Paper Correlation is the study of creating an automated machine to assess the relationship between two or more papers. This study continues to develop in the natural language processing field. In this paper, we begin to enter into this study by conducting research to find the relationship between two citation sentences. The citation sentences contained in the paper can be used as a basis for assessing the relationship between two papers. In this research, the semantic similarity between citation sentences is included in.8 semantic classes. We use the Recurrent Neural Networks method and its variations and use SMOTE to deal with imbalanced data. The data representation that we use is word2vec to convert the sentence into a vector. We develop this research by adding data to see the algorithm's capabilities from previous research. [ABSTRACT FROM AUTHOR]