Preprocessed experimental dataset

The three publicly available implicit feedback datasets, Last.FM, Delicious, and CiaoDVD, were processed to create the experimental dataset in this paper. Last.FM is a music sharing platform, and the dataset records user tags for artists from August 1, 2005, to May 9, 2011. The Delicious dataset consists of user-saved bookmarks and tags from the Delicious website, capturing user interactions with saved links from November 20, 2003, to November 9, 2010. CiaoDVD is a movie rating dataset that records user ratings for movies from May 31, 2000, to November 28, 2013. All datasets underwent uniform preprocessing, where user ratings or tags for items were considered as interactions between the user and the item. Additionally, user interactions with items were sorted based on the interaction time to generate user behavior sequences, and users with sequences shorter than three interactions were excluded. Each user's behavior sequence was divided into three parts: the last interacted item was used as the test set, the second-to-last item was used as the validation set, and all remaining items formed the training set.