News Topic Classification Using Machine Learning Techniques

摘要

News topic classification is a method of classifying news articles available in text data into some predefined classes or labels. This is one of the applications of text classification. Text classification can be applied in the fields of spam filtering, language recognition, segmenting customer feedbacks, segregating technical documents, etc. This paper discusses news topic classification on AG's News Topic Classification Dataset using machine learning algorithms such as linear support vector machine, multinomial Naive Bayesian classifier, K-Nearest Neighbor, Rocchio, bagging, and boosting. This paper discusses three steps for classification, namely pre-processing of text, then applying feature extraction techniques, and finally implementing machine learning algorithms. These algorithms are compared using evaluation metrics like Accuracy, Recall, Precision, and F1 Score.