News Location Classification

Abstract

Every day there are a lot of things that happen around the world. There are various ways to record every event that is occurring around the world, such as news, blogs, and articles. Over the past few years, there are multiple news available on every event that has occurred. It adds to the size of information that is available for human beings to consume. People are, moving from paper-based newspapers to digital newspapers to get their daily feed of news and digitization has a role to play in this behaviour. These days every person is preoccupied with a lot of work, online and offline, as mentioned earlier the amount of information is being increased with every passing day. For this reason, people are only interested in news that match their interests. A large amount of data in the form of text is available online, hence its classification based on its hidden features can lead to the better recommendation of news to individuals. In this research work, we have used focus area and temporal features to classify news using a Convolutional Neural Network (CNN). The results of the proposed methodology in the form of precision, accuracy, recall, and F1-Score show that these features indeed can be used for recommender systems.