<![CDATA[Access to information continues to experience significant developments. With the rapid advancement of the internet, the amount of news content available on digital platforms is also increasing rapidly. Internet users can quickly and easily access news and information from various sources. However, this also brings new challenges for internet users, especially digital news readers. With the vast amount of available news, readers often receive news recommendations that are irrelevant to their interests. This is due to the different preferences of each user. Additionally, each user may have more than one preference, leading to the appearance of random and unwanted news recommendations. Therefore, this research aims to enhance the personalization of news recommendations by utilizing POS-Tagger technology to analyze news content. Additionally, the content-based filtering method is used to match news with user preferences based on previously consumed content. The news matching is done after calculating vectors using TF-IDF, followed by matching using cosine similarity calculation. The recommender system demonstrates a good ability to provide recommendations that are relevant to user preferences. The performance evaluation showed satisfactory results. F1-score showed an average result of 90% from the three users, and high cosine similarity value with an average from the three users of 8% of the overall recommendation results indicating a high relevance between the recommendations and the news that users have read.]]>
