Mining Application on Analyzing Users’ Interests from Twitter

In today’s world, it is problematic to provide users of social-media with posts that are analyzed from their interest efficiently. Users are unable to see the good quality and variety of posts based on their interest. The mass adoption of smartphones along with an internet connection via wi-fi or cellular network enables to analyse users’ interest from Twitter. Twitter is used by a large number of audience to share their posts on a variety of topics as tweets. Then mining users’ interests from Twitter can amplify a number of efficacies, such as advertising, trending topics that can be analyzed by interests and recommendation of users’ posts. For this purpose, this paper provides an Android application which incorporates Web Services, Jsoup, JSON, Firebase Real-time Database and MVC. The application aids to select the posts which include spectacular images and text that are shown to users as a training set. The personalized posts can later be inferred and analyzed by the users themselves using Suffix Array Data Structure and Artificial Neural Network (ANN). Under ANN, we have used Backpropagation methodology that fires neurons as posts. Kosaraju algorithm and Palette library then help in removing redundant posts while later one also retaining relevant posts with specified hashtags more efficiently and accurately.