Age Based News Classification using LSTM and BERT

Abstract

News has been playing a major role in our day-to-day life in educating, informing, alerting, and even entertaining people. In our research, we aim on classifying news articles as adult news and kid news with a view to providing age-appropriate news. News classification as adult news or kid news is done by news article analysis. For fulfilling our target, we developed our news dataset and constructed two different models: one using Long short-term memory(LSTM) and the other using Bidirectional Encoder Representations from Transformers(BERT). The LSTM model achieved an accuracy of 94.84% and the BERT model managed to reach an accuracy of 95.68%, which is a bit higher than the LSTM model for news classification.