Effective Deep Learning Approach for Genuineness of Twitter's Views Using Sentiment Analysis

Abstract Social Media has tremendously achieved digital advancements in the field of Information technology. The users have tools such as Twitter, micro blogging sites and blogs for sharing everyday life events, views regarding products and services, and news. Individual shares opinions, views and thoughts posts in digital form using social media sites. The most advanced and widely used social media site is Twitter, through people share their views, reviews, and opinions and broadcast them. Due to this activity, ample views and comments are being share by users on various social media platforms, e-commerce platforms, websites, and blogs. The user posts the opinion or view as positive or negative of any product sometimes leading to appraise the product or damage a product or business reputations. These views shared by the user on various platforms are not genuine or true. It may be fabricated, false, or spam. In this paper we proposed BiLSTM deep learning techniques with incorporating and combined twitter users’ features to predict the genuineness of views. The results obtained in experimentation were optimum over the traditional state of art methods. The newly build hybrid approach presented the best performance considering accuracy, F1 score, Precision and Recall .The results of the conducted experiments demonstrates ample accuracy , precision and recall.