A Novel Supervised Approach to Detection of Shilling Attack in Collaborative Filtering Based Recommendation System

Collaborative filtering is widely used recommendation algorithm to generate variety of recommendation for target users. With increasing popularity of collaborative filtering recommendation, number of users started to insert fake shilling profiles into the system. Due to shilling attack or profile in...

Full description

Saved in:
Bibliographic Details
Published in:International journal of computer science and information security 2016-04, Vol.14 (4), p.208-208
Main Authors: Patel, Krupa, Thakkar, Amit, Shah, Chandni, Makvana, Kamlesh
Format: Article
Language:English
Subjects:
Algorithms
Classification
Collaboration
Computer information security
Computer simulation
Decision trees
Filtering
Filtration
Inserts
Mathematical models
Neural networks
Ratings & rankings
Recommender systems
Teaching methods
User profiles
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Collaborative filtering is widely used recommendation algorithm to generate variety of recommendation for target users. With increasing popularity of collaborative filtering recommendation, number of users started to insert fake shilling profiles into the system. Due to shilling attack or profile injection attack, accuracy of collaborative filtering recommendation will reduce. This paper attempts to proposed method to detection of shilling attack in collaborative filtering recommendation system using supervised approach. Our proposed method use statistical parameters RDMA, DigSim and LengthVar to identify shilling attack profiles from genuine profile. This parameters are use to train the model for detection of attacker profiles. Then our proposed method will identify genuine profile those are classified as attacker profiles.
ISSN:1947-5500