Analysis of the KNN Classifier Distance Metrics for Bulgarian Fake News Detection

Abstract

The problem with fake news and click-bait headlines is not new. Despite that this kind of news has existed for many years now, it was just a few years ago that the focus of the scientific community turned to them. Even though there is quite a lot of research on the topic of detecting fake news in general, there is a limited number of studies on fake news in Cyrillic. This paper focuses on the classification of fake news and click-bait headlines from Bulgarian Facebook Pages. The KNN machine-learning algorithm is chosen and four different distance metrics are tested and analyzed.