Method for user categorization in social media, computer program, and computer

摘要

In order to enable categorization on the basis of content similarity, without acquiring content for all users, a method for clustering a plurality of users of social media in which a text profile and text content is associated with each user includes: a step in which a content characteristic vector is generated for each user in a portion of all the users, on the basis of associated content; a step in which a plurality of clusters and the portion of users are mapped on the basis of the content characteristic vector; a step in which a profile characteristic vector for each cluster is generated on the basis of profiles associated with the portion of users mapped to each cluster; and a step in which users other than the portion of users are categorized into the plurality of clusters on the basis of profiles associated with said other users and the profile characteristic vector.