Multiview Ensemble Method for Detecting Shilling Attacks in Collaborative Recommender Systems

Table 1

Common attack models.

Attack model
	Items	Rating	Items	Rating	Rating
Random attack	Not used		Randomly chosen	System mean
Average attack	Not used		Randomly chosen	Item mean
AoP attack	Not used		Top x% of most popular items	Item mean
User random shifting	Not used		Randomly chosen	System mean+random
User average shifting	Not used		Randomly chosen	Item mean+random
Target random shifting	Not used		Randomly chosen	System mean
Target average shifting	Not used		Randomly chosen	Item mean
Power item attack	Power item	Ratings with normal distribution around item mean and standard deviation
Power user attack	Copy of the power user’s items and ratings
Bandwagon attack	Popular items		Randomly chosen	System mean
Love/Hate attack	Not used		Randomly chosen
Reverse Bandwagon attack	Unpopular items		Randomly chosen	System mean