Classification of news about Turkey in the foreign press through text mining

In recent years, many newspapers and news providers have begun presenting their content via web pages or through social media. This shift has led to a massive increase in the volume of news content available, necessitating the analysis and management of this vast information flow. In this study, 8,385 pieces of news content related to Turkey were collected from the web pages of major foreign news content providers, including Fox, The Guardian, BBC, and CNN. While traditional techniques classify news texts into categories based on their content, this study achieved an average accuracy rate of 89.36% by classifying the contents according to eight predefined areas of interest. Moreover, analyses were conducted based on the publication dates of all foreign news content, revealing relationships between the dates of publication and the classified areas of interest. Additionally, sentiment analysis was conducted on the collected foreign news content using the BERT algorithm, which identified the sentiment categories of the contents and examined the perception of Turkey in the foreign press.