SlideShare a Scribd company logo
1 of 1
RESEARCH POSTER PRESENTATION DESIGN © 2012
www.PosterPresentations.com
Recommendation system (RS) plays an
important role in many areas of today's digital
world. A recommender system for ecommerce
websites not only helps sellers to increase their
sells but it also helps buyers to purchase most
accurate item. In last one decade; several attacks
came into existence that can affect the accuracy
of the system. We have used a statistical based
approach to detect the attacks. We compare the
accuracy of six models and measure the accuracy
of models. Then by ensembling the top three
performing models using voting technique we
build our ensemble model.
ABSTRACT
Our implementation consists three steps. In first step, we measure the effectiveness of the attack in
terms of the prediction shift in the average rating of a particular item. In second step, we find the
attributes to distinguish the authentic and attack profiles. Two types of attributes generic
(RDMA,WDMA, Degree Similarity, Length Variance) [4] and model specific(Mean Variance, FMTD)
are used. In last step, we compare the accuracy of statistical models in distinguishing the profiles. It is
performed in two different ways. In first approach, we compare six machine learning models i.e.
decision tree, random forest, ada boost, SVM, linear regression and neural networks. We use precision
and recall to measure their accuracy. We find top three performing models and ensemble them using the
voting technique to build our ensemble model. In second approach, we give a comparative study of
four clustering models i.e. EM, farthest first, hierarchical and simple K mean.
IMPLEMENTATION & RESULTS
We use MovieLens 100K dataset. This
dataset contains 943 users who has given
100,000 ratings to 1682 items. All the
ratings are in the range between 1 to 5.
To measure the robustness of the system
we use K-fold cross validation.
CONCLUSION & FUTURE SCOPE
In this work, we examine six machine learning
classification models and measure their
performance for the detection of attack profiles
in recommender system. Based on their
performance we find out top three performing
models are neural network, SVM and random
forest. We combine them to make our ensemble
model. We also give a comparative study of four
unsupervised models. The accuracy of proposed
ensemble model is more than 90% in most of the
cases. It is expected that optimizing of model
parameters may leads to better results. This
approach can be used in other areas also like
spam filtering, intrusion detections etc. Similar
approach can be used in the detection of DOS
attacks or in the identification of intrusions in the
system.
REFERENCES
1. Davoodi, Fatemeh Ghiyafeh, and Omid
Fatemi. "Tag based recommender system for
social bookmarking sites." In Proceedings of
the 2012 International Conference on
Advances in Social Networks Analysis and
Mining (ASONAM 2012), pp. 934-940. IEEE
Computer Society, 2012.
2. Bobadilla, Jesús, Fernando Ortega, Antonio
Hernando, and Abraham Gutiérrez.
"Recommender systems survey." Knowledge-
Based Systems 46 (2013): 109-132.
3. Burke, Robin, Bamshad Mobasher, Runa
Bhaumik, and Chad Williams. "Segment-
based injection attacks against collaborative
filtering recommender systems." In Data
Mining, Fifth IEEE International Conference
on, pp. 4-pp. IEEE, 2005.
4. Chirita, Paul-Alexandru, Wolfgang Nejdl, and
Cristian Zamfir. "Preventing shilling attacks
in online.
Ashish Kumar
Master of Engineering (Software Engineering), CSED, Thapar University, Patiala
Profile Injection Attack Detection in Recommender System
INTRODUCTION
Recommender system predicts the preference
that a user would give to an item. Content based
and collaborative filtering based are two widely
used types of recommender systems. Content
based filtering approach uses the properties of an
item to recommend additional items with similar
properties [1]. Collaborative filtering works by
building a database of preferences for various
items in the system by users [2]. It works on the
principle that two users who had similar tastes in
the past will also have similar taste in future
also. As collaborative filtering based system is
open to the users input, so they have high
chances of profile injection attacks. The aim of
the attacker is to interact with the recommender
system to either push or nuke a particular item
[3]. Attack profile structure:
Types of attacks:
1. Push Attacks: Average, Random,
Bandwagon and Segment Attack.
2. Nuke Attacks: Love-Hate and Reverse-
bandwagon Attack.
Filler Size 1% 10% 20% 30% 40% 50%
Models P R P R P R P R P R P R
Decision Tree .90 .892 .921 .93 .928 .919 .939 .912 .94 .94 .961 .968
Random Forest .929 .930 .939 .92 .948 .948 .952 .956 .962 .961 .973 .979
Ada Boost .9 .908 .914 .918 .93 .924 .935 .938 .948 .934 .950 .943
SVM .938 .927 .949 .94 .959 .959 .971 .975 .979 .981 .988 .989
Linear Regression .89 .862 .89 .907 .918 .91 .925 .918 .929 .902 .93 .931
Neural Network .949 .938 .95 .943 .951 .954 .959 .958 .968 .967 .979 .98
Ensemble .939 .932 .946 .934 .953 .953 .961 .963 .968 .969 .98 .982

More Related Content

What's hot

Hybrid sentiment and network analysis of social opinion polarization icoict
Hybrid sentiment and network analysis of social opinion polarization   icoictHybrid sentiment and network analysis of social opinion polarization   icoict
Hybrid sentiment and network analysis of social opinion polarization icoictAndry Alamsyah
 
Collaborative Filtering Recommendation Algorithm based on Hadoop
Collaborative Filtering Recommendation Algorithm based on HadoopCollaborative Filtering Recommendation Algorithm based on Hadoop
Collaborative Filtering Recommendation Algorithm based on HadoopTien-Yang (Aiden) Wu
 
Effective Navigation of Query Results Based On Hierarchies
Effective Navigation of Query Results Based On HierarchiesEffective Navigation of Query Results Based On Hierarchies
Effective Navigation of Query Results Based On HierarchiesAkhil Ambekar
 
Dynamic personalized recommendation on sparse data
Dynamic personalized recommendation on sparse dataDynamic personalized recommendation on sparse data
Dynamic personalized recommendation on sparse dataJPINFOTECH JAYAPRAKASH
 
Movie recommendation system using collaborative filtering system
Movie recommendation system using collaborative filtering system Movie recommendation system using collaborative filtering system
Movie recommendation system using collaborative filtering system Mauryasuraj98
 
Collaborative filtering
Collaborative filteringCollaborative filtering
Collaborative filteringNeha Kulkarni
 
Tag recommendation in social bookmarking sites like deli
Tag recommendation in social bookmarking sites like deliTag recommendation in social bookmarking sites like deli
Tag recommendation in social bookmarking sites like deliVinay Singri
 
DYNAMIC LARGE SCALE DATA ON TWITTER USING SENTIMENT ANALYSIS AND TOPIC MODELING
DYNAMIC LARGE SCALE DATA ON TWITTER USING SENTIMENT ANALYSIS AND TOPIC MODELINGDYNAMIC LARGE SCALE DATA ON TWITTER USING SENTIMENT ANALYSIS AND TOPIC MODELING
DYNAMIC LARGE SCALE DATA ON TWITTER USING SENTIMENT ANALYSIS AND TOPIC MODELINGAndry Alamsyah
 
Finding Pattern in Dynamic Network Analysis
Finding Pattern in Dynamic Network AnalysisFinding Pattern in Dynamic Network Analysis
Finding Pattern in Dynamic Network AnalysisAndry Alamsyah
 
Collaborative Filtering 1: User-based CF
Collaborative Filtering 1: User-based CFCollaborative Filtering 1: User-based CF
Collaborative Filtering 1: User-based CFYusuke Yamamoto
 
Tag recommendation in social bookmarking sites like deli
Tag recommendation in social bookmarking sites like deliTag recommendation in social bookmarking sites like deli
Tag recommendation in social bookmarking sites like deliVinay Singri
 
Online Learning to Rank
Online Learning to RankOnline Learning to Rank
Online Learning to Rankewhuang3
 
[Final]collaborative filtering and recommender systems
[Final]collaborative filtering and recommender systems[Final]collaborative filtering and recommender systems
[Final]collaborative filtering and recommender systemsFalitokiniaina Rabearison
 
Collaborative Filtering
Collaborative FilteringCollaborative Filtering
Collaborative FilteringTayfun Sen
 
Movies Recommendation System
Movies Recommendation SystemMovies Recommendation System
Movies Recommendation SystemShubham Patil
 
INFO4990_Hossain
INFO4990_HossainINFO4990_Hossain
INFO4990_Hossainwebuploader
 
Hybrid recommender systems
Hybrid recommender systemsHybrid recommender systems
Hybrid recommender systemsrenataghisloti
 

What's hot (20)

Hybrid sentiment and network analysis of social opinion polarization icoict
Hybrid sentiment and network analysis of social opinion polarization   icoictHybrid sentiment and network analysis of social opinion polarization   icoict
Hybrid sentiment and network analysis of social opinion polarization icoict
 
Collaborative Filtering Recommendation Algorithm based on Hadoop
Collaborative Filtering Recommendation Algorithm based on HadoopCollaborative Filtering Recommendation Algorithm based on Hadoop
Collaborative Filtering Recommendation Algorithm based on Hadoop
 
Effective Navigation of Query Results Based On Hierarchies
Effective Navigation of Query Results Based On HierarchiesEffective Navigation of Query Results Based On Hierarchies
Effective Navigation of Query Results Based On Hierarchies
 
Dynamic personalized recommendation on sparse data
Dynamic personalized recommendation on sparse dataDynamic personalized recommendation on sparse data
Dynamic personalized recommendation on sparse data
 
Movie recommendation system using collaborative filtering system
Movie recommendation system using collaborative filtering system Movie recommendation system using collaborative filtering system
Movie recommendation system using collaborative filtering system
 
Collaborative filtering
Collaborative filteringCollaborative filtering
Collaborative filtering
 
Collaborative filtering
Collaborative filteringCollaborative filtering
Collaborative filtering
 
Tag recommendation in social bookmarking sites like deli
Tag recommendation in social bookmarking sites like deliTag recommendation in social bookmarking sites like deli
Tag recommendation in social bookmarking sites like deli
 
DYNAMIC LARGE SCALE DATA ON TWITTER USING SENTIMENT ANALYSIS AND TOPIC MODELING
DYNAMIC LARGE SCALE DATA ON TWITTER USING SENTIMENT ANALYSIS AND TOPIC MODELINGDYNAMIC LARGE SCALE DATA ON TWITTER USING SENTIMENT ANALYSIS AND TOPIC MODELING
DYNAMIC LARGE SCALE DATA ON TWITTER USING SENTIMENT ANALYSIS AND TOPIC MODELING
 
Finding Pattern in Dynamic Network Analysis
Finding Pattern in Dynamic Network AnalysisFinding Pattern in Dynamic Network Analysis
Finding Pattern in Dynamic Network Analysis
 
Collaborative Filtering 1: User-based CF
Collaborative Filtering 1: User-based CFCollaborative Filtering 1: User-based CF
Collaborative Filtering 1: User-based CF
 
Tag recommendation in social bookmarking sites like deli
Tag recommendation in social bookmarking sites like deliTag recommendation in social bookmarking sites like deli
Tag recommendation in social bookmarking sites like deli
 
Online Learning to Rank
Online Learning to RankOnline Learning to Rank
Online Learning to Rank
 
Combined queries
Combined queriesCombined queries
Combined queries
 
Anp slideshow july_2001
Anp slideshow july_2001Anp slideshow july_2001
Anp slideshow july_2001
 
[Final]collaborative filtering and recommender systems
[Final]collaborative filtering and recommender systems[Final]collaborative filtering and recommender systems
[Final]collaborative filtering and recommender systems
 
Collaborative Filtering
Collaborative FilteringCollaborative Filtering
Collaborative Filtering
 
Movies Recommendation System
Movies Recommendation SystemMovies Recommendation System
Movies Recommendation System
 
INFO4990_Hossain
INFO4990_HossainINFO4990_Hossain
INFO4990_Hossain
 
Hybrid recommender systems
Hybrid recommender systemsHybrid recommender systems
Hybrid recommender systems
 

Viewers also liked

Administración Financiera
Administración FinancieraAdministración Financiera
Administración Financieragabrielaorcellet
 
Reckoken .. ^ ^ for linaa :* :*
Reckoken .. ^ ^ for linaa :* :*Reckoken .. ^ ^ for linaa :* :*
Reckoken .. ^ ^ for linaa :* :*M Reckoken
 
Coursecompletion (22)
Coursecompletion (22)Coursecompletion (22)
Coursecompletion (22)Kris Owen
 
Teorema pythagoras
Teorema pythagorasTeorema pythagoras
Teorema pythagorasdintadanti
 
Texas Tech diploma 2014
Texas Tech diploma 2014Texas Tech diploma 2014
Texas Tech diploma 2014Laurinda Abbie
 
A study of the demographic differences of instructors in using e-Textbooks in...
A study of the demographic differences of instructors in using e-Textbooks in...A study of the demographic differences of instructors in using e-Textbooks in...
A study of the demographic differences of instructors in using e-Textbooks in...Sirui Wang
 
training cert Avalontest
training cert Avalontesttraining cert Avalontest
training cert Avalontestterry montrey
 
Grow Hormones
Grow HormonesGrow Hormones
Grow Hormonespaviadoeg
 
Can innonvation be monetized?
Can innonvation be monetized?Can innonvation be monetized?
Can innonvation be monetized?Dipanjan "DJ" Nag
 
10 Business people make their voices heard
10 Business people make their voices heard10 Business people make their voices heard
10 Business people make their voices heardGreg Eicke
 
DETERMINING THE PRICE FOR YOUR PROPERTY
DETERMINING THE PRICE FOR YOUR PROPERTYDETERMINING THE PRICE FOR YOUR PROPERTY
DETERMINING THE PRICE FOR YOUR PROPERTYBangalore Property
 
Slide on Singapore's Lost Soul
Slide on Singapore's Lost SoulSlide on Singapore's Lost Soul
Slide on Singapore's Lost SoulLLoyd Yeo
 
Effective language translation services with expert translators in u.s
Effective language translation services with expert translators in u.sEffective language translation services with expert translators in u.s
Effective language translation services with expert translators in u.sBel Translation Services
 

Viewers also liked (17)

Administración Financiera
Administración FinancieraAdministración Financiera
Administración Financiera
 
Deven Dewald
Deven DewaldDeven Dewald
Deven Dewald
 
Reckoken .. ^ ^ for linaa :* :*
Reckoken .. ^ ^ for linaa :* :*Reckoken .. ^ ^ for linaa :* :*
Reckoken .. ^ ^ for linaa :* :*
 
Coursecompletion (22)
Coursecompletion (22)Coursecompletion (22)
Coursecompletion (22)
 
Teorema pythagoras
Teorema pythagorasTeorema pythagoras
Teorema pythagoras
 
Texas Tech diploma 2014
Texas Tech diploma 2014Texas Tech diploma 2014
Texas Tech diploma 2014
 
A study of the demographic differences of instructors in using e-Textbooks in...
A study of the demographic differences of instructors in using e-Textbooks in...A study of the demographic differences of instructors in using e-Textbooks in...
A study of the demographic differences of instructors in using e-Textbooks in...
 
training cert Avalontest
training cert Avalontesttraining cert Avalontest
training cert Avalontest
 
Dia de muertos 2
Dia de muertos 2Dia de muertos 2
Dia de muertos 2
 
Grow Hormones
Grow HormonesGrow Hormones
Grow Hormones
 
Bab iii
Bab iiiBab iii
Bab iii
 
Can innonvation be monetized?
Can innonvation be monetized?Can innonvation be monetized?
Can innonvation be monetized?
 
10 Business people make their voices heard
10 Business people make their voices heard10 Business people make their voices heard
10 Business people make their voices heard
 
DETERMINING THE PRICE FOR YOUR PROPERTY
DETERMINING THE PRICE FOR YOUR PROPERTYDETERMINING THE PRICE FOR YOUR PROPERTY
DETERMINING THE PRICE FOR YOUR PROPERTY
 
Slide on Singapore's Lost Soul
Slide on Singapore's Lost SoulSlide on Singapore's Lost Soul
Slide on Singapore's Lost Soul
 
Marketsoft - Data Services
Marketsoft - Data Services Marketsoft - Data Services
Marketsoft - Data Services
 
Effective language translation services with expert translators in u.s
Effective language translation services with expert translators in u.sEffective language translation services with expert translators in u.s
Effective language translation services with expert translators in u.s
 

Similar to presentation

A REVIEW PAPER ON BFO AND PSO BASED MOVIE RECOMMENDATION SYSTEM | J4RV4I1015
A REVIEW PAPER ON BFO AND PSO BASED MOVIE RECOMMENDATION SYSTEM | J4RV4I1015A REVIEW PAPER ON BFO AND PSO BASED MOVIE RECOMMENDATION SYSTEM | J4RV4I1015
A REVIEW PAPER ON BFO AND PSO BASED MOVIE RECOMMENDATION SYSTEM | J4RV4I1015Journal For Research
 
Profile Injection Attack Detection in Recommender System
Profile Injection Attack Detection in Recommender SystemProfile Injection Attack Detection in Recommender System
Profile Injection Attack Detection in Recommender SystemASHISH PANNU
 
A Hybrid Approach For Phishing Website Detection Using Machine Learning.
A Hybrid Approach For Phishing Website Detection Using Machine Learning.A Hybrid Approach For Phishing Website Detection Using Machine Learning.
A Hybrid Approach For Phishing Website Detection Using Machine Learning.vivatechijri
 
Cyb 5675 class project final
Cyb 5675   class project finalCyb 5675   class project final
Cyb 5675 class project finalCraig Cannon
 
Evaluating and Enhancing Efficiency of Recommendation System using Big Data A...
Evaluating and Enhancing Efficiency of Recommendation System using Big Data A...Evaluating and Enhancing Efficiency of Recommendation System using Big Data A...
Evaluating and Enhancing Efficiency of Recommendation System using Big Data A...IRJET Journal
 
Collaborative Filtering Recommendation System
Collaborative Filtering Recommendation SystemCollaborative Filtering Recommendation System
Collaborative Filtering Recommendation SystemMilind Gokhale
 
Developing a Secured Recommender System in Social Semantic Network
Developing a Secured Recommender System in Social Semantic NetworkDeveloping a Secured Recommender System in Social Semantic Network
Developing a Secured Recommender System in Social Semantic NetworkTamer Rezk
 
Movie Recommender System Using Artificial Intelligence
Movie Recommender System Using Artificial Intelligence Movie Recommender System Using Artificial Intelligence
Movie Recommender System Using Artificial Intelligence Shrutika Oswal
 
Profile injection attack detection in recommender system
Profile injection attack detection in recommender systemProfile injection attack detection in recommender system
Profile injection attack detection in recommender systemASHISH PANNU
 
A Brief Survey on Recommendation System for a Gradient Classifier based Inade...
A Brief Survey on Recommendation System for a Gradient Classifier based Inade...A Brief Survey on Recommendation System for a Gradient Classifier based Inade...
A Brief Survey on Recommendation System for a Gradient Classifier based Inade...Christo Ananth
 
Evaluating Collaborative Filtering Recommender Systems
Evaluating Collaborative Filtering Recommender SystemsEvaluating Collaborative Filtering Recommender Systems
Evaluating Collaborative Filtering Recommender SystemsMegaVjohnson
 
Analysis on Recommended System for Web Information Retrieval Using HMM
Analysis on Recommended System for Web Information Retrieval Using HMMAnalysis on Recommended System for Web Information Retrieval Using HMM
Analysis on Recommended System for Web Information Retrieval Using HMMIJERA Editor
 
Security evaluation of pattern classifiers under attack
Security evaluation of pattern classifiers under attackSecurity evaluation of pattern classifiers under attack
Security evaluation of pattern classifiers under attackShakas Technologies
 
A Novel Latent Factor Model For Recommender System
A Novel Latent Factor Model For Recommender SystemA Novel Latent Factor Model For Recommender System
A Novel Latent Factor Model For Recommender SystemAndrew Parish
 
Multidirectional Product Support System for Decision Making In Textile Indust...
Multidirectional Product Support System for Decision Making In Textile Indust...Multidirectional Product Support System for Decision Making In Textile Indust...
Multidirectional Product Support System for Decision Making In Textile Indust...IOSR Journals
 
Building a usage profile for anomaly detection on computer networks
Building a usage profile for anomaly detection on computer networksBuilding a usage profile for anomaly detection on computer networks
Building a usage profile for anomaly detection on computer networksNathanael Asaam
 
Presentation_Malware Analysis.pptx
Presentation_Malware Analysis.pptxPresentation_Malware Analysis.pptx
Presentation_Malware Analysis.pptxnishanth kurush
 
VTU final year project report Main
VTU final year project report MainVTU final year project report Main
VTU final year project report Mainathiathi3
 

Similar to presentation (20)

Major
MajorMajor
Major
 
A REVIEW PAPER ON BFO AND PSO BASED MOVIE RECOMMENDATION SYSTEM | J4RV4I1015
A REVIEW PAPER ON BFO AND PSO BASED MOVIE RECOMMENDATION SYSTEM | J4RV4I1015A REVIEW PAPER ON BFO AND PSO BASED MOVIE RECOMMENDATION SYSTEM | J4RV4I1015
A REVIEW PAPER ON BFO AND PSO BASED MOVIE RECOMMENDATION SYSTEM | J4RV4I1015
 
Profile Injection Attack Detection in Recommender System
Profile Injection Attack Detection in Recommender SystemProfile Injection Attack Detection in Recommender System
Profile Injection Attack Detection in Recommender System
 
A Hybrid Approach For Phishing Website Detection Using Machine Learning.
A Hybrid Approach For Phishing Website Detection Using Machine Learning.A Hybrid Approach For Phishing Website Detection Using Machine Learning.
A Hybrid Approach For Phishing Website Detection Using Machine Learning.
 
Cyb 5675 class project final
Cyb 5675   class project finalCyb 5675   class project final
Cyb 5675 class project final
 
Evaluating and Enhancing Efficiency of Recommendation System using Big Data A...
Evaluating and Enhancing Efficiency of Recommendation System using Big Data A...Evaluating and Enhancing Efficiency of Recommendation System using Big Data A...
Evaluating and Enhancing Efficiency of Recommendation System using Big Data A...
 
Spam email filtering
Spam email filteringSpam email filtering
Spam email filtering
 
Collaborative Filtering Recommendation System
Collaborative Filtering Recommendation SystemCollaborative Filtering Recommendation System
Collaborative Filtering Recommendation System
 
Developing a Secured Recommender System in Social Semantic Network
Developing a Secured Recommender System in Social Semantic NetworkDeveloping a Secured Recommender System in Social Semantic Network
Developing a Secured Recommender System in Social Semantic Network
 
Movie Recommender System Using Artificial Intelligence
Movie Recommender System Using Artificial Intelligence Movie Recommender System Using Artificial Intelligence
Movie Recommender System Using Artificial Intelligence
 
Profile injection attack detection in recommender system
Profile injection attack detection in recommender systemProfile injection attack detection in recommender system
Profile injection attack detection in recommender system
 
A Brief Survey on Recommendation System for a Gradient Classifier based Inade...
A Brief Survey on Recommendation System for a Gradient Classifier based Inade...A Brief Survey on Recommendation System for a Gradient Classifier based Inade...
A Brief Survey on Recommendation System for a Gradient Classifier based Inade...
 
Evaluating Collaborative Filtering Recommender Systems
Evaluating Collaborative Filtering Recommender SystemsEvaluating Collaborative Filtering Recommender Systems
Evaluating Collaborative Filtering Recommender Systems
 
Analysis on Recommended System for Web Information Retrieval Using HMM
Analysis on Recommended System for Web Information Retrieval Using HMMAnalysis on Recommended System for Web Information Retrieval Using HMM
Analysis on Recommended System for Web Information Retrieval Using HMM
 
Security evaluation of pattern classifiers under attack
Security evaluation of pattern classifiers under attackSecurity evaluation of pattern classifiers under attack
Security evaluation of pattern classifiers under attack
 
A Novel Latent Factor Model For Recommender System
A Novel Latent Factor Model For Recommender SystemA Novel Latent Factor Model For Recommender System
A Novel Latent Factor Model For Recommender System
 
Multidirectional Product Support System for Decision Making In Textile Indust...
Multidirectional Product Support System for Decision Making In Textile Indust...Multidirectional Product Support System for Decision Making In Textile Indust...
Multidirectional Product Support System for Decision Making In Textile Indust...
 
Building a usage profile for anomaly detection on computer networks
Building a usage profile for anomaly detection on computer networksBuilding a usage profile for anomaly detection on computer networks
Building a usage profile for anomaly detection on computer networks
 
Presentation_Malware Analysis.pptx
Presentation_Malware Analysis.pptxPresentation_Malware Analysis.pptx
Presentation_Malware Analysis.pptx
 
VTU final year project report Main
VTU final year project report MainVTU final year project report Main
VTU final year project report Main
 

presentation

  • 1. RESEARCH POSTER PRESENTATION DESIGN © 2012 www.PosterPresentations.com Recommendation system (RS) plays an important role in many areas of today's digital world. A recommender system for ecommerce websites not only helps sellers to increase their sells but it also helps buyers to purchase most accurate item. In last one decade; several attacks came into existence that can affect the accuracy of the system. We have used a statistical based approach to detect the attacks. We compare the accuracy of six models and measure the accuracy of models. Then by ensembling the top three performing models using voting technique we build our ensemble model. ABSTRACT Our implementation consists three steps. In first step, we measure the effectiveness of the attack in terms of the prediction shift in the average rating of a particular item. In second step, we find the attributes to distinguish the authentic and attack profiles. Two types of attributes generic (RDMA,WDMA, Degree Similarity, Length Variance) [4] and model specific(Mean Variance, FMTD) are used. In last step, we compare the accuracy of statistical models in distinguishing the profiles. It is performed in two different ways. In first approach, we compare six machine learning models i.e. decision tree, random forest, ada boost, SVM, linear regression and neural networks. We use precision and recall to measure their accuracy. We find top three performing models and ensemble them using the voting technique to build our ensemble model. In second approach, we give a comparative study of four clustering models i.e. EM, farthest first, hierarchical and simple K mean. IMPLEMENTATION & RESULTS We use MovieLens 100K dataset. This dataset contains 943 users who has given 100,000 ratings to 1682 items. All the ratings are in the range between 1 to 5. To measure the robustness of the system we use K-fold cross validation. CONCLUSION & FUTURE SCOPE In this work, we examine six machine learning classification models and measure their performance for the detection of attack profiles in recommender system. Based on their performance we find out top three performing models are neural network, SVM and random forest. We combine them to make our ensemble model. We also give a comparative study of four unsupervised models. The accuracy of proposed ensemble model is more than 90% in most of the cases. It is expected that optimizing of model parameters may leads to better results. This approach can be used in other areas also like spam filtering, intrusion detections etc. Similar approach can be used in the detection of DOS attacks or in the identification of intrusions in the system. REFERENCES 1. Davoodi, Fatemeh Ghiyafeh, and Omid Fatemi. "Tag based recommender system for social bookmarking sites." In Proceedings of the 2012 International Conference on Advances in Social Networks Analysis and Mining (ASONAM 2012), pp. 934-940. IEEE Computer Society, 2012. 2. Bobadilla, Jesús, Fernando Ortega, Antonio Hernando, and Abraham Gutiérrez. "Recommender systems survey." Knowledge- Based Systems 46 (2013): 109-132. 3. Burke, Robin, Bamshad Mobasher, Runa Bhaumik, and Chad Williams. "Segment- based injection attacks against collaborative filtering recommender systems." In Data Mining, Fifth IEEE International Conference on, pp. 4-pp. IEEE, 2005. 4. Chirita, Paul-Alexandru, Wolfgang Nejdl, and Cristian Zamfir. "Preventing shilling attacks in online. Ashish Kumar Master of Engineering (Software Engineering), CSED, Thapar University, Patiala Profile Injection Attack Detection in Recommender System INTRODUCTION Recommender system predicts the preference that a user would give to an item. Content based and collaborative filtering based are two widely used types of recommender systems. Content based filtering approach uses the properties of an item to recommend additional items with similar properties [1]. Collaborative filtering works by building a database of preferences for various items in the system by users [2]. It works on the principle that two users who had similar tastes in the past will also have similar taste in future also. As collaborative filtering based system is open to the users input, so they have high chances of profile injection attacks. The aim of the attacker is to interact with the recommender system to either push or nuke a particular item [3]. Attack profile structure: Types of attacks: 1. Push Attacks: Average, Random, Bandwagon and Segment Attack. 2. Nuke Attacks: Love-Hate and Reverse- bandwagon Attack. Filler Size 1% 10% 20% 30% 40% 50% Models P R P R P R P R P R P R Decision Tree .90 .892 .921 .93 .928 .919 .939 .912 .94 .94 .961 .968 Random Forest .929 .930 .939 .92 .948 .948 .952 .956 .962 .961 .973 .979 Ada Boost .9 .908 .914 .918 .93 .924 .935 .938 .948 .934 .950 .943 SVM .938 .927 .949 .94 .959 .959 .971 .975 .979 .981 .988 .989 Linear Regression .89 .862 .89 .907 .918 .91 .925 .918 .929 .902 .93 .931 Neural Network .949 .938 .95 .943 .951 .954 .959 .958 .968 .967 .979 .98 Ensemble .939 .932 .946 .934 .953 .953 .961 .963 .968 .969 .98 .982