Classification of News Using Artificial Intelligence Methods

Many organizations use analysis and forecasting systems that rely on human labor for data collection and subsequent processing. The use of artificial intelligence algorithms improves the efficiency of these systems by reducing the risk of suboptimal decisions and erroneous judgments. For such systems to be useful, their costs must not exceed the generated revenue. This article presents experiments on news classification using intelligent methods as part of a larger system for automated forecasting of critical event. A model and experimental framework for a software system utilizing artificial intelligence methods are proposed. A prototype has been developed to experiment with the Multinomial Naive Bayes and Support Vector Machines algorithms, combined with the vectorizers CountVectorizer, HashingVectorizer, and TfidfVectorizer. The tests were conducted using a parallel computer consisting of four Raspberry Pi 4 nodes. The best results were achieved with the Multinomial Naive Bayes method and the CountVectorizer.