An ensemble detection method for shilling attacks based on features of automatic extraction

Hybrid CNN and RNN-based shilling attack framework in social recommender networks

“…As a whole, these detection scheme based on the detection of attack profiles is categorized into classification, clustering, graph mining and statistical methods. Unfortunately, these existing shilling attacks detections approaches inherited some inadequacies that do not consider the difference between the attacker and normal users or confines to the attackers ratings pattern, attack size sensitivity and restricted attack types [13]. When the difference between the normal and the attacker is not identified by the shilling attack detection approach, then mis-classification rate and false rate will surely get increased.…”