Abstract:
One of the most well-liked applications of natural language processing is text classification. Bengali is becoming more and more popular in this subject, much as many other languages, and the most well-known effort here is the categorization of various unlabeled news items. categories, such as national, international, IT and so on. Bengali news portals are becoming more and more prevalent today. The ease of access to web has made browsing news online a common activity.
The news site features a variety of news categories. This article presents a technique for categorizing news headlines from websites or news portals. An algorithm for machine learning makes predictions. Many of the gathered data were tested then trained. As which was before activities like tokenization, number removal, exclamation mark withdrawal, sign removal, and stop-word elimination are completed. Additionally, a list of stop phrases is manually prepared. Effective stop words improve performance. Stop words elimination is the most important factor in feature choice. Instead of analyzing news items from various online publications, this study focuses on categorizing Bengali News Headlines. There are eight different types of news. This work is being considered, and the news headlines are being utilized to categorize it. The model is used to model the input data. The overall model was attained the best performance by the GRU method. The height of the accuracy consisted of in case 84%.
