THE PERSPECTIVE OF DATA MINING: THE STUDY OF FAKE NEWS ON SOCIAL MEDIA

Abstract

Today's rapidly evolving network environment is affected by the continuing problem of information authenticity. Since it has led many consumers astray, fake news detection is still a concern. This study aims to add additional research into the study of fake news on social media by gaining its perspective via the use of data mining due to the paucity of studies that identify and categorize false news pieces. The study concentrated on the extraction, categorization, and analysis of word context features, such as the seven character composition attributes word frequencies, vocabulary without white space, word length mean and variance, word vulgarity, word and phrase counter, and computation of word rarity. It includes 50 different pieces of data from Buzz Feed News' Face book social media. The bagging classifier from Weka's Meta algorithm was used to get an accuracy rating of 86%. The study suggests that future research collects additional information from other social networking sites, classifies them, and analyzes them. It also suggests evaluating the model developed during the study's model development phase and contrasting the results of the two assessments.