Image Based Fake Tweet Retrieval (IBFTR)

The increasing growth of information technology has led to the development of fast information sharing through social media. Due to that, the spread of fake information is taking place at a greater speed and also has serious social implications. The legitimacy of such content is highly questioned yet there is no third party involved in checking this information. There is an increasing demand in this field to detect fake news and take appropriate measures to prevent its spreading. Social media information with images makes more impact. False exploration of social media information for the disaster rescue team at the disaster time can misguide the rescue team to provide help to a fake user instead of the real needy user. The objective of this paper is to prepare the innovative system Image Based Fake Tweet Retrieval (IBFTR) to increase the efficiency to detect the fake tweet on Twitter by extracting the features based on image captioning which relates the content of an image with the tweet text and image forgery which identifies the originality of the image with reference to the tweet content followed by feature selection to retain the relevant features for the efficient classification performance. Experimentation on selected features yields the classification with less complexity of space and execution time.