The spread of fake news in the present day has been a common occurrence especially in online social networks. Unverified news on social media have huge negative impact on public trust, political processes, and societal wellbeing. This research entails developing a predictive intelligence model for detecting fake news by using machine learning techniques. The study seeks to train it using articles from reliable news sources through web mining methods and then analyze news articles via NLP and ML algorithms to categorize it as either true or fake. The research methodology entailed detailed data preprocessing, text normalization, stemming, lemmatization, and sentiment analysis, to obtain linguistic and emotional markers mostly linked with fake news. The study uses a labeled dataset obtained from Kaggle repository which is used for training and evaluation. The research employs Bidirectional Encoder Representations from Transformers particularly for binary classification of Fake and Real News articles. Data Preprocessing steps like text normalization, stemming, lemmatization and sentiment analysis to extract linguistic and emotional markers are undertaken. Model performance is rigorously evaluated through precision, recall, and F1-score metrics, with particular attention to minimizing false positives/negatives in classification
Kamwira et al. (Fri,) studied this question.