The prediction of user's sharing behavior by deep learning method via analysis the content of tweet

Subject Areas : Futurology

hosniyeh safiarian ¹ , Mohamad jafar Tarokh ² , Mohammad Ali Afshar Kazemi ³

1 - head
2 - Khajeh Nasir
3 - Associate Professor, Faculty Member of Islamic Azad University, Central Tehran Branch

Received: 2020-12-23 Accepted : 2024-05-20 Published : 2025-08-23

Keywords: social media , user , deep learning , share , content,

Abstract :

Recently, Social media is an important way for the future of research that is why users follow events and trends on it. When users study their interesting tweets, there is likelihood of them sharing the tweets. Research shows that only 1% users can create 50% tweets and 25% users share the tweets. According to it, sharing tweet is an important tool for diffusion of information and news. The prediction of user’s sharing behavior is one of the most important challenges. The many of methods pay attention to statistical features also these methods do not focus on the content of tweets. Although there are many different content of issues , many of methods are not accurate as well as can not predict fast. This paper defines the content of feature and measure them to predict the value of tweet for users .This method is a binary classification based on preposed features to predict user’s sharing behavior individually. The classification problem is solved by deep learning . The proposed method is evaluated by three datasets .The accuracy of features is 85% as well as it is effective.