A Clickbait Detection Method on News Sites

Abstract

The use of internet news sites increases day by day. The internet has gone beyond institutions and organizations to provide a different service and there have been attempts to provide services only through the internet and thus to earn money. It can also be a organization or an individual who opens an account on the social network and provides financial gain on these accounts. The financial gain on the internet is increasing in parallel with the number of people entering the site in general or the number of people reading the content on the site. Clickbait is a click technique in which a user manipulates the curiosity of a person in order to open more pages in a web site, usually by writing exaggerated and unreal headlines. In this study, headlines or subheadings for news were collected. In these news articles, the Clickbait headline identified by TF-IDF has summarized by looking at the content of the Clickbait news with text feature extraction based on ontology method. The summarized version has been shown to the user without having to click on the newsletter. In this study, news in 4 news sites with Turkish and English content were examined. This study is the first Turkish study about Clickbait detection. In the English tests, results have given in comparison with equivalent algorithms and explained in detail.