News Category Classification using Natural Language Processing Transformer

In the information age of era, online news has rapidly increased. While grouping the news, each news has its own categorization. News categorization is the process that humans are interested one from online news articles. But the readers have the difficulty if the news is not in a categorized order. For news category classification, a new model is proposed. Large unorganized text data is explored by the procedure of Text mining. Hot topic of nowadays news has numerous amount of text data got from newspapers, media and other online blogs. To convert unstructured data to structured data, and do classification is quite challenging in text mining. Users now find it challenging to access news that is relevant to them due to the proliferation of news sources, making categorization of news essential for easy access. In real time, it will help the users to access news without wasting time. As news is continuously appearing it is much more difficult that news could be never-seen-before and could fall into new category. By using the NLP transformer, Bi-directional Encoding Representational Transformers (BERT), categorization can be done. The BERT assisted computers in deciphering the meaning of language in text using the surrounding text to establish the content. The proposed model gives a high accuracy of 91.0%.