A Novel Precise Personalized Learning Recommendation Model Regularized with Trust and Influence

Algorithm 1

The learning Algorithm of ITSVD.

	users set: scoring matrix: trust matrix: influential user set
Input: and the Iteration maximum
Output:
(1)	Initialize bias vector and feature matrix with random in (0,0,1);
(2)	While Loss function j not converged or the number of iteration L do
(3)	for each do
(4)	for each do
(5)	The score of item j by user u is predicted according to formula (11);
(6)	end
(7)	end
(8)	The loss functions is calculated according to formula (12)
(9)	Use the stochastic gradient descent method to update the parameters;
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)	Iteration number ++;
(18)	end
(19)	return