Achieving Privacy-Preserving Group Recommendation with Local Differential Privacy and Random Transmission

Table 1

Notations.

Notation	Meaning
	Number of users
	Number of movies
	Number of latent factors
	Rating generated by user for movie , which could be null
	User ’s preference ratings predicted by personalized recommendation under LDP,
	A group
	Number of group members
	Group preferences of movie
	Number of iterations
	Total number of ratings
	User ’s user vector
	Movie ’s item vector
	Regularization coefficient
	Learning rate
	Privacy budget
	Gradient matrix for at each iteration
	Change value during the transmission in IntPPA
	The probability that user’s profile is sent to group members
()	Group members need to execute IntPPA algorithm after (before) this time