A Comparison of Deep Learning Techniques Used to Identify Social Media Reviews

Using social media is one of the easy means that can be used for untruthful people to spread rumors, fake news, and falsified information. Thus, it becomes necessary for researchers to invent intelligent tools that can deploy Machine Learning and Deep Learning techniques for this reason. this paper presents a comparative study to test GraphSAGE and Node2vec embedding methods, such methods are used to transform the input data into a meaningful format used for classification through the computation of vector representations of a node based on random walks in the graph. Such methods can be considered as feature extraction methods. In particular, using GraphSAGE outperformed the Node2vec in terms of the well-known classification evaluation measures.