To run this, download the WIKI10+ dataset from http://nlp.uned.es/social-tagging/wiki10+/ and extract the files.
First run the tag extraction file (follow the comments, the file needs to be run multiple times), then the dataset maker, then the partition file and finally the feature extractor.