Enhancing Indonesian News Recommendations through Metadata Integration with Neural Attentive Multi-View Learning

The news recommendation system has the potential to help users discover articles that match their interests, which is crucial to alleviate user information overload. To generate effective news recommendations, one key capability is to accurately capture the contextual meaning of the text in news articles, as this is essential for obtaining useful representations for both news content and users. In this study, we examine the effectiveness of neural news recommendation with the Neural News Recommendation with Attentive Multi-View Learning (NAML) method to perform the news recommendation task in the Indonesian language. We investigate techniques for generating suitable vector representations of named entities in news content. We also propose to incorporate news metadata such as tags and entities in the news to improve the effectiveness of the NAML method in the Indonesian news recommendation system. Our results show that the NAML method leads to significant improvement in the effectiveness of news recommendations in the Indonesian language. Further addition of news metadata has been shown to improve the performance of the NAML method up to by 5.86% in terms of the NDCG@5 metric.