Afaan Oromo Language Fake News Detection in Social Media Using Convolutional Neural Network and Long Short Term Memory

Abstract

Social media platforms are constantly faced with rapid news which are disseminated based on relative importance, tone, and intended audience. Fake news is made up of stories written to deceive the readers of these social media platforms. Stance Detection is one of the factors influencing fake news detection. It is not practical for humans to fact-check every piece of information produced by the media. Hence, the goal of this research is to use natural language processing techniques to automate stance detection to determine the quality of the news source. It considers what other organizations write about the same headline. A body of the text is claimed to agree, disagree, discuss, or be unrelated to a headline. The system was developed by preprocessing the text according to the language grammar and, mixing convolutional neural network and recurrent neural network (RNN). Therefore, it performs with an accuracy of 67.1% and, an error result of 32.9%.

Keywords

• stance
• news dataset
• afaan oromo
• detection