• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권9호
• /
• pp.4684-4705
• /
• 2019

Collaborative filtering recommender systems are vulnerable to shilling attacks in which malicious users may inject biased profiles to promote or demote a particular item being recommended. To tackle this problem, many robust collaborative recommendation methods have been presented. Unfortunately, the robustness of most methods is improved at the expense of prediction accuracy. In this paper, we construct a robust Bayesian probabilistic matrix factorization model for collaborative filtering recommender systems by incorporating the detection of user anomaly rating behaviors. We first detect the anomaly rating behaviors of users by the modified K-means algorithm and target item identification method to generate an indicator matrix of attack users. Then we incorporate the indicator matrix of attack users to construct a robust Bayesian probabilistic matrix factorization model and based on which a robust collaborative recommendation algorithm is devised. The experimental results on the MovieLens and Netflix datasets show that our model can significantly improve the robustness and recommendation accuracy compared with three baseline methods.

• Journal of Computing Science and Engineering
• /
• 제7권1호
• /
• pp.21-29
• /
• 2013

The introduction of recommender systems to existing online services is now practically inevitable, with the increasing number of items and users on online services. Popular recommender systems have successfully implemented satisfactory systems, which are usually based on collaborative filtering. However, collaborative filtering-based recommenders suffer from well-known problems, such as popularity bias, and the cold-start problem. In this paper, we propose an innovative collaborative-filtering based recommender system, which uses the concepts of Experts and Novices to create fine-grained recommendations that focus on being novel, while being kept relevant. Experts and Novices are defined using pre-made clusters of similar items, and the distribution of users' ratings among these clusters. Thus, in order to generate recommendations, the experts are found dynamically depending on the seed items of the novice. The proposed recommender system was built using the MovieLens 1 M dataset, and evaluated with novelty metrics. Results show that the proposed system outperforms matrix factorization methods according to discovery-based novelty metrics, and can be a solution to popularity bias and the cold-start problem, while still retaining collaborative filtering.

• 한국지능정보시스템학회:학술대회논문집
• /
• 한국지능정보시스템학회 2004년도 추계학술대회
• /
• pp.203-210
• /
• 2004

Due to the explosion of e-commerce, recommender systems are rapidly becoming a core tool to accelerate cross-selling and strengthen customer loyalty. There are two prevalent approaches for building recommender systems - content-based recommending and collaborative filtering. Collaborative filtering recommender systems have been very successful in both information filtering domains and e-commerce domains, and many researchers have presented variations of collaborative filtering to increase its performance. However, the current research on recommendation has paid little attention to the use of time related data in the recommendation process. Up to now there has not been any study on collaborative filtering to reflect changes in user interest. This paper proposes dynamic fuzzy clustering algorithm and apply it to collaborative filtering algorithm for dynamic recommendations. The proposed methodology detects changes in customer behavior using the customer data at different periods of time and improves the performance of recommendations using information on changes. The results of the evaluation experiment show the proposed model's improvement in making recommendations.

• Journal of the Korean Data and Information Science Society
• /
• 제21권1호
• /
• pp.185-194
• /
• 2010

In recommender systems which are used widely at e-commerce, collaborative filtering needs the information of user-ratings and neighbor user-ratings. These are an important value for recommendation in recommender systems. We investigate the in-formation of rating in NBCFA (neighbor Based Collaborative Filtering Algorithm), we suggest new algorithm that improve prediction accuracy of recommender system. After we analyze relations between two variable and Error Value (EV), we suggest new algorithm and apply it to fitted line. This fitted line uses Least Squares Method (LSM) in Exploratory Data Analysis (EDA). To compute the prediction value of new algorithm, the fitted line is applied to experimental data with fitted function. In order to confirm prediction accuracy of new algorithm, we applied new algorithm to increased sparsity data and total data. As a result of study, the prediction accuracy of recommender system in the new algorithm was more improved than current algorithm.

• Management Science and Financial Engineering
• /
• 제11권3호
• /
• pp.137-153
• /
• 2005

Collaborative filtering-based recommendation systems make personalized recommendations based on users' ratings on products. Recommender systems must collect sufficient rating information from users to provide relevant recommendations because less user rating information results in poorer performance of recommender systems. To learn about new users, recommendation systems must first present users with an initial item list. In this study, we designed and analyzed seven selection strategies including the popularity, favorite, clustering, genre, and entropy methods. We investigated how these strategies performed using MovieLens, a public dataset. While the favorite and popularity methods tended to produce the highest average score and greatest average number of ratings, respectively, a hybrid of both favorite and popularity methods or a hybrid of demographic, favorite, and popularity methods also performed within acceptable ranges for both rating scores and numbers of ratings.

• Journal of information and communication convergence engineering
• /
• 제22권1호
• /
• pp.7-13
• /
• 2024

User preferences and ratings may be anticipated by recommendation systems, which are widely used in social networking, online shopping, healthcare, and even energy efficiency. Constructing trustworthy recommender systems for various applications, requires the analysis and mining of vast quantities of user data, including demographics. This study focuses on holding elections with vague voter and candidate preferences. Collaborative user ratings are used by filtering algorithms to provide suggestions. To avoid information overload, consumers are directed towards items that they are more likely to prefer based on the profile data used by recommender systems. Better interactions between governments, residents, and businesses may result from studies on recommender systems that facilitate the use of e-government services. To broaden people's access to the democratic process, the concept of "e-democracy" applies new media technologies. This study provides a framework for an electronic voting advisory system that uses machine learning.

• International journal of advanced smart convergence
• /
• 제9권3호
• /
• pp.42-48
• /
• 2020

These days people are overwhelmed by information on the Internet thus searching for useful information becomes burdensome, often failing to acquire some in a reasonable time. Recommender systems are indispensable to fulfill such user needs through many practical commercial sites. This study proposes a novel similarity measure for user-based collaborative filtering which is a most popular technique for recommender systems. Compared to existing similarity measures, the main advantages of the suggested measure are that it takes all the ratings given by users into account for computing similarity, thus relieving the inherent data sparsity problem and that it reflects the uncertainty or vagueness of user ratings through fuzzy logic. Performance of the proposed measure is examined by conducting extensive experiments. It is found that it demonstrates superiority over previous relevant measures in terms of major quality metrics.

• 지능정보연구
• /
• 제27권2호
• /
• pp.1-15
• /
• 2021

추천시스템은 사용자의 기호를 파악하여 물품 구매 결정을 도와주는 역할을 할 뿐만 아니라, 비즈니스 전략의 관점에서도 중요한 역할을 하기에 많은 기업과 기관에서 관심을 갖고 있다. 최근에는 다양한 추천시스템 연구 중에서도 NLP와 딥러닝 등을 결합한 하이브리드 추천시스템 연구가 증가하고 있다. NLP를 이용한 감성분석은 사용자 리뷰 데이터가 증가함에 따라 2000년대 중반부터 활용되기 시작하였지만, 기계학습 기반 텍스트 분류를 통해서는 텍스트의 특성을 완전히 고려하기 어렵기 때문에 리뷰의 정보를 식별하기 어려운 단점을 갖고 있다. 본 연구에서는 기계학습의 단점을 보완하기 위하여 BERT 기반 감성분석을 활용한 추천시스템을 제안하고자 한다. 비교 모형은 Naive-CF(collaborative filtering), SVD(singular value decomposition)-CF, MF(matrix factorization)-CF, BPR-MF(Bayesian personalized ranking matrix factorization)-CF, LSTM, CNN-LSTM, GRU(Gated Recurrent Units)를 기반으로 하는 추천 모형이며, 실제 데이터에 대한 분석 결과, BERT를 기반으로 하는 추천시스템의 성과가 가장 우수한 것으로 나타났다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제10권7호
• /
• pp.3192-3212
• /
• 2016

With the advent and popularity of e-commerce, an increasing number of consumers prefer to order tourism products online. A recommender system can help these users contend with information overload; however, such a system is affected by the cold start problem. Online tourism destination searching is a more difficult task than others on account of its more restrictive factors. In this paper, we therefore propose a tourism destination recommender system that employs opinion-mining technology to refine user preferences and item opinion reputations. These elements are then fused into a hybrid collaborative filtering method by combining user- and item-based collaborative filtering approaches. Meanwhile, we embed an artificial interactive module in our recommender system to alleviate the cold start problem. Compared with several well-known cold start recommendation approaches, our method provides improved recommendation accuracy and quality. A series of experimental evaluations using a publicly available dataset demonstrate that the proposed recommender system outperforms existing recommender systems in addressing the cold start problem.

• Journal of Information Processing Systems
• /
• 제9권3호
• /
• pp.435-460
• /
• 2013

Recommender systems are popular applications that help users to identify items that they could be interested in. A recent research area on recommender systems focuses on detecting several kinds of inconsistencies associated with the user preferences. However, the majority of previous works in this direction just process anomalies that are intentionally introduced by users. In contrast, this paper is centered on finding the way to remove non-malicious anomalies, specifically in collaborative filtering systems. A review of the state-of-the-art in this field shows that no previous work has been carried out for recommendation systems and general data mining scenarios, to exactly perform this preprocessing task. More specifically, in this paper we propose a method that is based on the extraction of knowledge from the dataset in the form of rating regularities (similar to frequent patterns), and their use in order to remove anomalous preferences provided by users. Experiments show that the application of the procedure as a preprocessing step improves the performance of a data-mining task associated with the recommendation and also effectively detects the anomalous preferences.