Neighborhood Attentional Memory Networks for Recommendation Systems

Table 1

Notations and descriptions.

Notations	Descriptions
M, E	Memory matrix for users and items
C, Y	External memory matrix for users and items
	Embedding vectors for user , user , item I, and item j
N(i)	User neighborhood set of all users who have accessed item i
	Item neighborhood set of all items that user has accessed
h, H	The number of hops and the last hop
	h-hop user preference vector given item i and item attribute vector given user
	The similarity of the user with user given item i and the similarity of the item i with item j given user in h-hop
	Attention weights vectors of user neighborhood given item i and attention weights vectors of item neighborhood given user in h-hop
	Attention weight of user given item i and attention weight of item j given user in h-hop
	h-hop user neighborhood representation given item i and item neighborhood representation given user
	h-hop user neighborhood preference representation given item i and item neighborhood attribute representation given user
	The parameters to be learned in neural networks