Recommendation datasets

The recommendation algorithm development, evaluation, and benchmarking are performed on standardized and common recommendation datasets. One way to assess the maturity of a domain in personalization research is by examining the context of standardized recommendation datasets used by researchers at premier recommendation systems and information retrieval journals, conferences, and data challenges such as, RecSys Data Challenge, SIGIR ECOM, TREC, and KDD cup. We have compiled the standardized and benchmark datasets in the recently published recommendation research and survey articles. Standard personalized recommendation datasets and their categorization across different domains is mentioned in ‘available datasets.xlsx’. Whereas ‘benchmark datasets.xlsx’ provides the details of benchmark recommendation datasets.