Text Summarization in Social Networks by using Deep Learning

Abstract

Social networks are user-based platforms where users share their thoughts and feelings visually or in writing. Social networks are also networks where users can comment on the topic (s). Emotion analysis and text summarization on shared data and big data in social networks have become popular today. In this study, Word2Vec model was created with data collected from many social networks. With the Word2vec model, a semantic context is created on the text that users share. With all the data collected, the GRU (Gated Recurrent Unit) neural network was trained and a 2-output negative-positive model was formed. The aim is to generate summary text for the user by gathering comments under a label on Twitter. At the same time, the generated abstract text is given a sense with the created model. In the study, 94% success rate was found and the text was found successfully with LSA text summarization algorithm. © 2019 IEEE.