Note: Articles are clustered by their headline. Articles with potentially similar headlines should be grouped together.
Clustering is performed using the HDBSCAN algorithm. The default settings use a minimum cluster size of 2 and minimum samples of 2. Minimum cluster size: The smallest number of articles that can form a cluster. If a group of articles is fewer than this number, it will not be recognized as a cluster. Minimum samples: The smallest number of close neighbors an article must have to be considered a core point of a cluster. Determines the cluster density requirement and affects algorithm's sensitivity to noise.
Emotion keywords
Note: Emotion keywords are extracted from comments. They may not be directly related to the articles or their content.Processing time of many comments can take up to 1.5 minutes.