• Journal of Intelligence and Information Systems
• /
• v.15 no.1
• /
• pp.31-50
• /
• 2009

As the collected information which is static or dynamic is infinite in ubiquitous computing environments, information overload and invasion of privacy have been pressing issues in the recommendation service. In this study, we propose a recommendation service procedure through P2P, The P2P helps customer to obtain effective and secure product information because of communication among customers who have the similar preference about the products without connection to server. To evaluate the performance of the proposed recommendation service, we utilized real transaction and product data of the Korean mobile company which service character images. We developed a prototype recommender system and demonstrated that the proposed recommendation service makes an effect on recommending product in the ubiquitous environments. We expect that the information overload and invasion of privacy will be solved by the proposed recommendation procedure in ubiquitous environment.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.07a
• /
• pp.587-590
• /
• 2020

최근, 디지털 콘텐츠 산업이 폭발적으로 성장됨에 따라 고객 유치를 위한 개인화 추천 기술들이 많은 주목을 받고 있다. 개인화 추천 방식들을 큰 갈래로 나누어 본다면 협업 필터링 기술과 내용 기반 기술로 나눌 수 있다. 협업 필터링의 경우 개인화 추천에는 적합하지만 사용자 평가 데이터의 양이 방대해야 하며 초기에 평가자가 없는 콘텐츠에 대해 추천할 수 없는 초기 평가자 문제가 존재한다. 따라서 매일 방대한 양의 콘텐츠가 편입되는 분야에서 사용하기에 큰 결점이 될 수 있다. 본 논문에서는 영화들의 정보가 담긴 데이터 셋과 사용자 평가 데이터, 그리고 사용자의 선호 기준을 의미하는 메타 가중치를 활용한 내용 기반의 맞춤형 영화 추천 시스템을 제안한다. 논문에서는 먼저, 영화를 고를 때 일반적으로 중요시 보는 속성들을 활용하여 영화의 특징 벡터를 구성하고, 이를 사용자 평가와 결합하여 개인의 선호에 대한 특징 벡터를 구성하는 방법을 제안하며, 구성된 데이터와 코사인 유사도, 메타 가중치를 활용하여 사용자 선호와 유사한 영화들을 도출하는 방법을 제안한다. 또한, 평가데이터를 활용하여 구현된 추천시스템의 검증 프로세스를 구성하고, 검증 프로세스를 활용한 손실 함수를 설계하여 적합한 메타 가중치를 학습하는 방법을 제시한다. 본 논문에서 제안하는 시스템은 다수의 속성을 조합하여 활용하므로 추천 결과가 과도하게 특수화 되지 않을 수 있으며, 메타 가중치라는 요소를 통해 더욱 개인화 된 추천을 제공할 수 있다.

• Journal of Intelligence and Information Systems
• /
• v.17 no.3
• /
• pp.147-168
• /
• 2011

As personalized recommendation of products and services is rapidly growing in importance, a number of studies provided fundamental knowledge and techniques for developing recommendation systems. Among them, the CF technique has been most widely used and has proven to be useful in many practices. However, current collaborative filtering (CF) technique has still considerable rooms for improving the effectiveness of recommendation systems: 1) a similarity function most systems use to find so-called like-minded people is not well defined in that similarity is computed from a single perspective of similarity concept; and 2) temporal information that contains the changing preference of customers needs to be taken into account when making recommendations. We hypothesize that integration of multiple aspects of similarity and utilization of temporal information will improve the accuracy of recommendations. The objective of this paper is to test the hypothesis through a series of experiments using MovieLens data. The experimental results show that the proposed recommendation system highly outperforms the conventional CF-based systems, confirming our hypothesis.

• Database Research
• /
• v.34 no.3
• /
• pp.99-123
• /
• 2018

Due to the ever expanding nature of the Web, separating more valuable information from the noisy data is getting more important. Although recommendation systems are widely used for addressing the information overloading issue, their performance does not seem meaningfully improved in currently suggested approaches. Hence, to investigate the issues, this study discusses different characteristics of popular, existing recommendation approaches, and proposes a new profiling technique that uses collaborative tags and test whether it successfully compensates the limitations of the existing approaches. In addition, the study also empirically evaluates rating/tagging patterns of users in various recommendation approaches, which include the proposed approach, to learn whether those patterns can be used as effective cues for improving the recommendations accuracy. Through the sensitivity analyses, this study also suggests the potential associated with a single recommendation system that applies multiple approaches for different users or items depending upon the types and contexts of recommendations.

• Journal of the Korean Society for information Management
• /
• v.23 no.3 s.61
• /
• pp.91-125
• /
• 2006

Recent advancements in information technology and the Internet have caused an explosive increase in the information available and the means to distribute it. However, such information overflow has made the efficient and accurate search of information a difficulty for most users. To solve this problem, an information retrieval and filtering system was developed as an important tool for users. Libraries and information centers have been in the forefront to provide customized services to satisfy the user's information needs under the changing information environment of today. The aim of this study is to propose an efficient information service for libraries and information centers to provide a personalized recommendation system to the user. The proposed method overcomes the weaknesses of existing systems, by providing a personalized hybrid recommendation method for multimedia contents that works in a large-scaled data and user environment. The system based on the proposed hybrid method uses an effective framework to combine Association Rule with Collaborative Filtering Method.

• Information Systems Review
• /
• v.19 no.2
• /
• pp.71-94
• /
• 2017

Recommender systems reduce information overload and enhance choice quality. This technology is used in many services and industry. Previous studies did not consider recommendation quantity and the repetitive recommendations of an item. This study is the first to examine recommender systems by considering recommendation quantity and repetitive recommendations. Only a limited number of items are displayed in offline stores because of their physical limitations. Determining the type and number of items that will be displayed is an important consideration. In this study, I suggest the use of a user-based recommender system that can recommend the most appropriate items for each store. This model is evaluated by MAE, Precision, Recall, and F1 measure, and shows higher performance than the baseline model. I also suggest a new performance evaluation measure that includes Quantity Precision, Quantity Recall, and Quantity F1 measure. This measure considers the penalty for short or excess recommendation quantity. Novelty is defined as the proportion of items in a recommendation list that consumers may not experience. I evaluate the new revenue creation effect of the suggested model using this novelty measure. Previous research focused on recommendations for customer online, but I expand the recommender system to cover stores offline.

• Smart Media Journal
• /
• v.5 no.2
• /
• pp.84-89
• /
• 2016

For a place-recommendation model based on user's behavior and multi-attribute attitude in this thesis. We focus groups that show similar patterns of visiting restaurants and then compare one and the other. We make use of The Fishbein Equation, Pearson's Correlation Coefficient to calculate multi-attribute attitude scores. Furthermore, We also make use of Preference Prediction Algorithm and Distance based method named "Euclidean Distance" to provide accurate results. We can demonstrate how excellent this system is through several experiments carried out with actual data.

• Journal of Intelligence and Information Systems
• /
• v.14 no.2
• /
• pp.157-177
• /
• 2008

Even in a single day, an enormous amount of content including digital videos, posts, photographs, and wikis are generated on the web. It's getting more difficult to recommend to a user what he/she prefers among these contents because of the difficulty of automatically grasping of content's meanings. CF (Collaborative Filtering) is one of useful methods to recommend proper content to a user under these situations because the filtering process is only based on historical information about whether or not a target user has preferred an item before. Collaborative Tagging is the process that allows many users to annotate content with descriptive tags. Recommendation using tags can partially improve, such as the limitations of CF, the sparsity and cold-start problem. In this research, a CF method with user-created tags is proposed. Collaborative tagging is employed to grasp and filter users' preferences for items. Empirical demonstrations using real dataset from del.icio.us show that our algorithm obtains improved performance, compared with existing works.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2019.01a
• /
• pp.437-438
• /
• 2019

스마트폰을 중심으로 한 모바일 기기의 보급과 온라인 소셜 네트워크 서비스의 이용자들이 증가하면서 사용자들은 많은 콘텐츠를 소비하고 공유한다. 이는 콘텐츠 사용자들의 개별적 기호에 맞지 않거나 만족도가 떨어지는 콘텐츠를 소비하게 한다. 이와 같은 문제를 해결하기 위해 소셜 네트워크 사용자에게 적합한 콘텐츠를 추천하기 위한 기법에 대한 연구가 활발하게 진행되고 있다. 본 논문에서는 온라인 상에 존재하는 다양한 정보 중에서 공연과 관련한 콘텐츠들을 중심으로 사용자 성향별로 추천을 해줄 수 있는 협업필터링 방법에 대하여 제안한다.

• Journal of Intelligence and Information Systems
• /
• v.12 no.4
• /
• pp.1-14
• /
• 2006

This study proposed a customer preference estimation model for production recommendation and a method to enhance the performance of product recommendation using the estimated customer preference information. That is, we suggested customer preference estimation model to estimate exactly customer's product preference with his behavior. This model shows the relationship of customer's behaviors with his preferences. The proposed estimation model is optimized by learning the relative weights of customer's behavior variables to have an effect on his preference and enables to estimate exactly his preference. To validate our proposed models, we collected virtual book store data and then made a comparative analysis of our proposed models and a benchmark model in terms of performance results of collaborative filtering for product recommendation. The benchmark model means a prior preference weighting model. The results of our empirical analysis showed that our proposed model performed better results than the benchmark model.