A Hybrid TF-IDF and N-Grams Based Feature Extraction Approach for Accurate Detection of Fake News on Twitter Data

Abstract

As there is an exponential growth of social networks and due to large usage of social media,
there is an increasing demand for data in the web for the users which leads to current inclinations
concepts in the area of research. Sentiment, text analysis and social media analysis, especially in
user reviews and a tweet has become a popular area of research. Fake and Real data classification
from user responses plays a key role in fake detection on social media platforms. Fake news and
lack of trust in the media are growing problems with huge difficulties in our society. Evidently in
a deceptive story or fake news in social Medias leads to change its description. The main
objective of this research is to detect the fake news, which is a classic text classification problem
with a straightforward proposition. There is needed to build a model that can differentiate
between “Real” news and “Fake” news with NLP (Natural Language Processing) and ML
(Machine Learning) techniques for discovering the 'fake news', or deceptive news stories that
arises from the defective bases. Often, some preprocessing steps and feature extraction
techniques are applied to obtain features from twitter data to improve the accuracy using
supervised classification algorithm. This paper elucidates the ML classification approaches with
different feature extraction techniques to obtain a text analysis and the results obtained are
compared to identify the best possible approach.