Abstract:
In our present era, where the internet is pervasive, everyone relies on many online news sources for information. News quickly disseminated among millions of people within a relatively short time due to the increase in the use of social media platforms like Google, Facebook, Twitter, etc. The propagation of false information has wide-ranging effects, such as the development of prejudiced beliefs that might influence election results in favor of particular candidates. Additionally, spammers make money via clickbait ads by employing enticing news headlines. Support Vector Machines (SVM), Decision Trees (DT), and Naive Bayes are three well-known machine learning techniques used in this study to create a Bengali false news detection system. The dataset consists of labeled Bangla news articles gathered from online sources. Preprocessing techniques are applied to extract relevant features, and the algorithms are trained on the labeled dataset. Experimental evaluations using various performance metrics demonstrate the effectiveness of SVM, DT and Naive Bayes algorithms in Bangla fake news detection. A comparative analysis identifies their strengths and limitations.
