Table 3. Properties of MovieLens-1M (ML), Douban Book (DB), and Yelp datasets.

Dataset	# of users	# of items	# of ratings	Sparsity (%)	Rating average	# of Ratings per item	# of Ratings per user
ML	6,040	3,952	1,000,209	95.8	3.58	253.1	165.6
DB	13,024	22,347	792,062	99.7	4.05	35.4	60.8
Yelp	16,239	14,284	198,397	99.9	3.77	13.9	12.2