PARSING STRUKTUR PARAGRAF BERBASIS NEURAL NETWORK

Abstract

Paragraph parsing has an important role in the development of artificial intelligence. Parsing is the first step to reasoning paragraphs so that they can be understood by machines. The effectiveness of the paragraph position parsing method on how to decompose text into text segments. The segmentation process without taking into account the semantic structure of a paragraph will result in a structure that is not in sync with the actual meaning. To overcome this problem, this study proposes recursive neural network (RvNN) based method. This method tries to get the best binary tree that represents paragraph structure. The proposed method is applied to complete simple paragraphs about children's stories. The test results show that the proposed method can parse paragraphs with an accuracy rate of 0.9. The proposed method is also more efficient because there is no need to build the repository of paragraph structure..