237 research outputs found
Privacy-Preserving Concordance-based Recommendations on Vertically Distributed Data
No full text10th International Conference on ICT and Knowledge Engineering (ICT&KE) -- NOV 21-23, 2012 -- Siam Univ, Bangkok, THAILANDWOS: 000320918300004Recommender systems are attractive components of e-commerce. Customers apply such systems to get help for choosing the appropriate product to purchase. To provide accurate and dependable referrals, recommender systems require sufficient user data. On the other hand, since people purchase products from different online vendors, collected user data for recommendation purposes might be distributed among several e-companies. Consequently, due to distributed data, such companies having inadequate data cannot provide truthful predictions. To overcome this challenge, data holders might want to collaborate. However, due to privacy and financial fears, they might hesitate to partnership. In this paper, we propose a concordance measure-based solution that enables data holders to produce recommendations without jeopardizing their privacy. We perform real data set-based experiments and analyze the solution in terms of privacy and extra costs. The experimental results show that e-companies can produce more accurate recommendations by employing the provided scheme.Pacific Distance Multimedia Educ Network (APDMEN)), IEEE, IEEE Thailand Sect, Comp Assoc Thailan
An entropy-based neighbor selection approach for collaborative filtering
No full textWOS: 000331160200023Collaborative filtering is an emerging technology to deal with information overload problem guiding customers by offering recommendations on products of possible interest. Forming neighborhood of a user/item is the crucial part of the recommendation process. Traditional collaborative filtering algorithms solely utilize entity similarities in order to form neighborhoods. In this paper, we introduce a novel entropy-based neighbor selection approach which focuses on measuring uncertainty of entity vectors. Such uncertainty can be interpreted as how a user perceives rating domain to distinguish her tastes or diversification of items' rating distributions. The proposed method takes similarities into account along with such uncertainty values and it solves the optimization problem of gathering the most similar entities with minimum entropy difference within a neighborhood. Described optimization problem can be considered as combinatorial optimization and it is similar to 0-1 knapsack problem. We perform benchmark data sets-based experiments in order to compare our method's accuracy with the conventional user- and item-based collaborative filtering algorithms. We also investigate integration of our method with some of previously introduced studies. Empirical outcomes substantiate that the proposed method significantly improves recommendation accuracy of traditional collaborative filtering algorithms and it is possible to combine the entropy-based method with other compatible works introducing new similarity measures or novel neighbor selection methodologie
Predicting Instructor Performance by Feature Selection and Machine Learning Methods
No full textGünümüzde hayatın her sektöründe işlenen veri miktarının artması, veri madenciliğin giderek daha popüler hale gelmesine yol açmış ve yüksek miktarda verinin artan bir karmaşıklıkta işlenmesi ihtiyacı doğmuştur. Finanstan, sağlığa, savunmadan eğitime onlarca sektörün sorunlarını çözmek adına gün geçtikçe farklı yöntemler geliştirilmekte, sosyal, ekonomik, bilimsel birçok problemin çözümü adına veri madenciliği yöntemlerine başvurulmaktadır. Eğitilen ve eğiten sayısının gün geçtikçe arttığı eğitim sektöründe ise, sistemin başarısının geliştirilebilmesi için, gerek eğitilen gerekse eğitimcilerinin performanslarının takip edilmesi ve kıymetlendirilmesi ihtiyacı, eğitimsel veri madenciliği kavramını doğurmuştur. Bu alanda yapılan çalışmalar genel olarak, öğrenci performansı konularına yoğunlaştığından, eğitmen performansı konusunda daha çok çalışmaya ihtiyaç duyulmaktadır. Eğitimsel veri madenciliği alanında öznitelik seçme ile birleştirilmiş makine öğrenmesi kullanan çalışmaların genel olarak öğrenci performansı üzerine yoğunlaştığı, ancak az sayıdaki çalışmanın eğitmen performansı üzerinde durduğu görülmüştür. Bu çalışmamızda, eğitmen performansının eğitimsel veri madenciliği yöntemleriyle nasıl tespit edilebileceği üzerinde durulmuştur. Çalışma kapsamında Gazi Üniversitesi öğrencilerinin eğitmenleri hakkında doldurdukları bir Likert Ölçekli Anket veri seti üzerinde çalışılmış, çeşitli öznitelik indirgeme algoritmaları ve farklı makine öğrenme yöntemleriyle veri seti kıymetlendirilmiş ve eğitmenlerin performansları tahmin edilmiştir. Elde edilen sonuçlara göre genetik algoritma ile öznitelik seçmenin, kullanılan veri seti için diğer yöntemlere kıyasla en iyi sonucu verdiğini göstermiş ve 33 tane öznitelik yerine 19 öznitelik kullanılabileceği ortaya çıkarılmıştır. Genetik algoritma ile birlikte makine öğrenmesi yöntemi olarak derin öğrenme kullanımı ile birlikte %97,70 bir tahmin doğruluk performansına ulaşılmış ve bu değerin tüm özniteliklerin kullanılması ile elde edilebilecek değerden yüksek olduğu görülmüştür. Bu çalışmayı diğerlerinden farklı kılan özelliği ise, indirgenmiş öznitelik sayısı ve makine öğrenmesini birleştirmesinin yanında, eğitmen performanslarının sıralanması işlemini de somut olarak yapmasıdır.Today, increasing amount of data in all sector of life, make data mining more popular, and high amount of data in increasing complexity demanded to acquit. Different methods developed day by day, for solving problems at many sectors like finance, health, defense, and education, applied to data mining for many social, economic, and scientific issues. In the education area, where both number of instructors and students always increase, for enhancing system performance, it is needed to observe and evaluate the performance of students and instructors and such situation causes to reveal a new concept Educational Data Mining. Research in this area generally focuses on student performance. Thus, there is a need for research in instructor performance. Research using machine learning combined with attribute selection in the field of educational data mining have focused on student performance in general, but few studies have focused on instructor performance. In this paper, it was discussed how the performance of the instructor can be determined by educational data mining methods. A Likert type questionnaire dataset on opinions of the Gazi University’s student regarding their instructor’s teaching performance is used in this research and different feature reduction, and machine learning algorithms are used for evaluating the data set and performances of instructors. According to the obtained results, it has been revealed that the feature selection with genetic algorithm gives the best result for the used data set compared to the other methods and 19 attributes can be used instead of 33 attributes. Utilizing genetic algorithm and deep learning as a machine learning method has achieved a predictive accuracy performance of 97.70 %, which is higher than the value that can be achieved by using all the attributes. This study differs from the others in that it combines the reduced number of attributes and machine learning, as well as the ordering of instructor performances in concrete terms
Privacy-Preserving Trust-Based Recommendations on Vertically Distributed Data
No full text5th Annual IEEE International Conference on Semantic Computing (ICSC) -- SEP 18-22, 2011 -- Stanford Univ, Palo Alto, CAWOS: 000410187400066Providing recommendations on trusts between entities is receiving increasing attention lately. Customers may prefer different online vendors for shopping. Thus, their preferences about various products might be distributed among multiple parties. To provide more accurate and reliable referrals, such companies might decide to collaborate. Due to privacy, legal, and financial reasons, however, they do not want to work jointly. In this paper, we propose a method for providing trust-based predictions on vertically distributed data while preserving data owners' confidentiality. We analyze our scheme in terms of privacy and performance. We also perform experiments for accuracy analysis. Our analyses show that our scheme is secure and able to provide accurate and reliable predictions efficiently.IEEE, IEEE Comp Soc, Microsoft, Wells Fargo, Franz IncTUBITAK [108E211]This work is supported by Grant 108E211 from TUBITAK
Providing naïve Bayesian classifier-based private recommendations on partitioned data
No full text11th European Conference on Principles and Practice of Knowledge Discovery in Databases, PKDD 2007 -- 17 September 2007 through 21 September 2007 -- Warsaw -- 71124Data collected for collaborative filtering (CF) purposes might be split between various parties. Integrating such data is helpful for both e-companies and customers due to mutual advantageous. However, due to privacy reasons, data owners do not want to disclose their data. We hypothesize that if privacy measures are provided, data holders might decide to integrate their data to perform richer CF services. In this paper, we investigate how to achieve naïve Bayesian classifier (NBC)-based CF tasks on partitioned data with privacy. We perform experiments on real data, analyze our outcomes, and provide some suggestion
Robustness Analysis of Naïve Bayesian Classifier-Based Collaborative Filtering
No full textIn this study, binary forms of previously defined basic shilling attack models are proposed and the robustness of naïve Bayesian classifierbased collaborative filtering algorithm is examined. Real data-based experiments are conducted and each attack type's performance is explicated. Since existing measures, which are used to assess the success of shilling attacks, do not work on binary data, a new evaluation metric is proposed. Empirical outcomes show that it is possible to manipulate binary rating-based recommender systems' predictions by inserting malicious user profiles. Hence, it is shown that naïve Bayesian classifier-based collaborative filtering scheme is not robust against shilling attack
P2P collaborative filtering with privacy
No full textWOS: 000275603100008With the evolution of the Internet and e-commerce, collaborative filtering (CF) and privacy-preserving collaborative filtering (PPCF) have become popular The goal in CF is to generate predictions with decent accuracy, efficiently. The main issue in PPCF, however, is achieving such a goal while preserving users' privacy Many implementations of CF and PPCF techniques proposed so far are centralized In centralized systems, data is collected and Stored by a central server for CF purposes Centralized storage poses several hazards to Users because the central server controls users' data In this work, we investigate how to produce naive Bayesian classifier (NBC)-based recommendations while preserving users' privacy unthout using a central server In a community of people, user might create a peer-to-peer (P2P) network Through P2P network, users can communicate with each other and exchange data to produce predictions We share the workload of prediction process and offer referrals efficiently using P2P network We propose privacy-preserving schemes and analyze them in terms of accuracy, privacy, and efficiency Our real data-based results show that our schemes offer accurate NBC-based predictions with privacy eliminating central serve
Privacy-preserving SOM-based recommendations on horizontally distributed data
No full textWOS: 000305719900011To produce predictions with decent accuracy, collaborative filtering algorithms need sufficient data. Due to the nature of online shopping and increasing amount of online vendors, different customers' preferences about the same products can be distributed among various companies, even competing vendors. Therefore, those companies holding inadequate number of users' data might decide to combine their data in such a way to present accurate predictions with acceptable online performance. However, they do not want to divulge their data, because such data are considered confidential and valuable. Furthermore, it is not legal disclosing users' preferences: nevertheless, if privacy is protected, they can collaborate to produce correct predictions. We propose a privacy-preserving scheme to provide recommendations on horizontally partitioned data among multiple parties. In order to improve online performance, the parties cluster their distributed data off-line without greatly jeopardizing their secrecy. They then estimate predictions using k-nearest neighbor approach while preserving their privacy. We demonstrate that the proposed method preserves data owners' privacy and is able to suggest predictions resourcefully. By performing several experiments using real data sets, we analyze our scheme in terms of accuracy. Our empirical outcomes show that it is still possible to estimate truthful predictions competently while maintaining data owners' confidentiality based on horizontally distributed dataTUBITAK [108E221]This work was supported by the Grant 108E221 from TUBITAK
Privacy-Preserving Naive Bayesian Classifier-Based Recommendations on Distributed Data
No full textWOS: 000349390400002Data collected for recommendation purposes might be distributed among various e-commerce sites, which can collaboratively provide more accurate predictions. However, because of privacy concerns, they might not want to work together. If privacy measures are provided, they may decide to become involved in prediction generation processes. We propose privacy-preserving schemes eliminating e-commerce sites' privacy concerns for providing predictions on distributed data. We investigate how to achieve naive Bayesian classifier-based recommendations when data are distributed horizontally or vertically among multiple parties, even competing ones, without greatly violating their confidentiality. We analyze our schemes in terms of privacy and additional costs and show that they do not deeply violate online vendors' secrecy and they cause insignificant overhead costs. We also perform experiments on real data, evaluate our outcomes, and provide suggestions. Our empirical results show that our schemes produce more accurate predictions.TUBITAK [108E221]This work is supported by grant 108E221 from TUBITAK
Privacy-Preserving Random Projection-Based Recommendations Based on Distributed Data
No full textWOS: 000317850400002Providing recommendations based on distributed data has received an increasing amount of attention because it offers several advantages. Online vendors who face problems caused by a limited amount of available data want to offer predictions based on distributed data collaboratively because they can surmount problems such as cold start, limited coverage, and unsatisfactory accuracy through partnerships. It is relatively easy to produce referrals based on distributed data when privacy is not a concern. However, concerns regarding the protection of private data, financial fears due to revealing valuable assets, and legal regulations imposed by various organizations prevent companies from forming collaborations. In this study, we propose to use random projection to protect online vendors' privacy while still providing accurate predictions from distributed data without sacrificing online performance. We utilize random projection to eliminate the aforementioned issues so vendors can work in partnerships. We suggest privacy-preserving schemes to offer recommendations based on vertically or horizontally partitioned data among multiple companies. The recommended methods are analyzed in terms of confidentiality. We also analyze the superfluous loads caused by privacy concerns. Finally, we perform real data-based trials to evaluate the accuracy of the proposed schemes. The results of our analyses show that our methods preserve privacy, cause insignificant overheads, and offer accurate predictions.TUBITAK [108E221]This work is supported by grant 108E221 from TUBITAK
- …
