• 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
• /
• 제19권3호
• /
• pp.142-147
• /
• 2021

In this paper, we propose an intelligent research fund management system that applies artificial intelligence technology to an integrated research fund management system. By defining research fund management rules as work rules, a detection model learned using deep learning is designed, through which the disapproval status is presented for each research fund usage history. The disapproval detection system of the RCMS implemented in this study predicts whether the newly registered usage details are recognized or disapproved using an artificial intelligence model designed based on the use of an 8.87 million research fund registered in the RCMS. In addition, the item-detail recommendation system described herein presents the usage details according to the usage history item newly registered by the artificial intelligence model through a correlation between the research cost usage details and the item itself. The accuracy of the recommendation was shown to be 97.21%.

• 한국정보통신학회논문지
• /
• 제18권2호
• /
• pp.379-386
• /
• 2014

모바일 기기의 사용이 급증하면서 앱 스토어를 이용하는 사용자들 또한 증가하고 있다. 그러나 앱 스토어들은 대부분 단순한 랭킹 방식의 추천을 사용하므로 추천의 정확성이 떨어진다. 본 논문에서는 사용자에게 더 적합한 아이템을 추천하기 위해 사용자 정보 가중치와 아이템의 최근 선호 정도를 반영한 기법을 제안한다. 제안하는 기법은 데이터 셋을 카테고리별로 구분한 후 협업필터링 기법에 사용자 정보 가중치를 적용하여 예측값을 추출한다. 카테고리별로 아이템에 대한 최근 선호 정도를 반영하기 위해 특정 기간을 지정한 아이템 평가값 평균을 구한다. 최종적으로 두 결과 값을 결합하여 아이템을 추천한다. 실험을 통해 제안한 기법이 기존의 아이템 기반, 사용자 기반 기법보다 추천의 정확성과 적합성이 향상되는 것을 확인하였다.

• 한국IT서비스학회지
• /
• 제13권3호
• /
• pp.309-321
• /
• 2014

Recently, due to the advent of ubiquitous computing and the spread of intelligent portable device such as smart phone, iPad and PDA has been amplified, a variety of services and the amount of information has also increased fastly. It is becoming a part of our common life style that the demands for enjoying the wireless internet are increasing anytime or anyplace without any restriction of time and place. And also, the demands for e-commerce and many different items on e-commerce and interesting of associated items are increasing. Existing collaborative filtering (CF), explicit method, can not only reflect exact attributes of item, but also still has the problem of sparsity and scalability, though it has been practically used to improve these defects. In this paper, using a implicit method without onerous question and answer to the users, not used user's profile for rating to reduce customers' searching effort to find out the items with high purchasability, it is necessary for us to analyse the segmentation of customer and item based on customer data and purchase history data, which is able to reflect the attributes of the item in order to improve the accuracy of recommendation. We propose the method of recommendation system using association rule and weighted preference so as to consider many different items on e-commerce and to refect the profit/weight/importance of attributed of a item. To verify improved performance of proposing system, we make experiments with dataset collected in a cosmetic internet shopping mall.

• 한국IT서비스학회지
• /
• 제6권2호
• /
• pp.113-123
• /
• 2007

In ubiquitous computing environment such as ubiquitous marketplace (u-market), there is a need of providing context-based personalization service while considering the nomadic user preference and corresponding requirements. To do so, the recommendation systems should deal with the tremendous amount of context data. Hence, the purpose of this paper is to propose a novel recommendation method which provides the products-group list of the customers in u-market based on the shopping intention and preferences. We have developed FREPIRS(FREquent Purchased Item-sets Recommendation Service), which makes recommendation listof product-group, not individual product. Collaborative filtering and apriori algorithm are adopted in FREPIRS to build product-group.

• Journal of Information Technology Applications and Management
• /
• 제24권4호
• /
• pp.1-12
• /
• 2017

Item-based collaborative filtering (IBCF) is an important technology that is widely used in recommender system of online shopping malls. It uses historical information to compute item-item similarity and make predictions. However, in offline shopping each customer's purchasing pattern can be occurred continuously and repeatedly due to time and space constraints contrast to online shopping. Those facts can make IBCF to have limitations from being applied to offline shopping malls directly. In order to improve the quality of recommendations made by IBCF in offline shopping mall, we propose an ensemble approach that considers both item-item similarity of IBCF and each customer's purchasing patterns which are modeled by item networks. Our experimental results show that this approach produces recommendation results superior to those of existing works such as pure IBCF or bestseller approaches.

• 한국멀티미디어학회논문지
• /
• 제25권6호
• /
• pp.819-832
• /
• 2022

Cross-domain recommender system transfers knowledge across different domains to improve the recommendation performance in a target domain that has a relatively sparse model. However, they suffer from the "negative transfer" in which transferred knowledge operates as noise. This paper proposes a novel Multiple Fusion-based Deep Cross-Domain Recommendation named MFDCR. We exploit Doc2Vec, one of the famous word embedding techniques, to fuse data user-wise and transfer knowledge across multi-domains. It alleviates the "negative transfer" problem. Additionally, we introduce a simple multi-layer perception to learn the user-item interactions and predict the possibility of preferring items by users. Extensive experiments with three domain datasets from one of the most famous services Amazon demonstrate that MFDCR outperforms recent single and cross-domain recommendation algorithms. Furthermore, experimental results show that MFDCR can address the problem of "negative transfer" and improve recommendation performance for multiple domains simultaneously. In addition, we show that our approach is efficient in extending toward more domains.

• 한국콘텐츠학회논문지
• /
• 제7권10호
• /
• pp.19-26
• /
• 2007

협력적 필터링은 아이템에 대한 선호도를 기반으로 이웃 선정 방법을 사용하므로 내용을 반영하지 못할뿐만 아니라 희박성 및 확장성 문제를 가지고 있다. 이러한 문제를 개선하기 위하여 아이템 기반 협력적 필터링이 실용화되었으나 아이템의 속성을 반영하지는 못한다. 본 논문에서는 기존의 개인화 추천 시스템의 문제점을 해결하기 위하여 FP-Tree를 이용한 연관 군집 방법을 제안하였다. 제안된 방법으로는 FP-Tree를 이용하여 후보집합의 발생없이 빈발항목을 구성하고 연관규칙을 생성한다. 생성된 연관 규칙의 신뢰도에 따라서 $\alpha-cut$을 사용하여 효율적인 연관 군집을 한다. 성능평가를 위해 MovieLens 데이터 집합에서 Gibbs Sampling, EM, K-means와 비교 평가하였다.

• 한국전자거래학회:학술대회논문집
• /
• 한국전자거래학회 2005년도 종합학술대회
• /
• pp.147-153
• /
• 2005

최근 인터넷 쇼핑몰에서 상품을 구매하는 고객들에게 편의성과 효율성을 제공하기 위하여 구매자들의 선호도나 가격에 맞는 상품을 추천해 주는 연구들이 활발하게 진행되고 있지만추천된 상품들의 구매시점에 관한 연구는 찾아보기 어렵다. 이에 본 논문에서는 인터넷 쇼핑몰의 적극적인 마케팅 일환으로 판매가격의 흐름을 시계열 패턴으로 분석하여 상품의 구매시점 정보를 제공하는 방안을 제안한다. 이를 위하여 과거의 판매 기록 데이터베이스에 있는 판매가격의 기준이 되는 패턴과 유사한 변화를 보이는 패턴을 정규화된 유사도로써 검색하고, 검색된 가격 패턴을 기준으로 미래의 가격 패턴의 변화를 분석하여, 미래 가격 패턴의 변화 폭에 따라 상품에 대한 구매시점을 제공한다.

• 한국멀티미디어학회논문지
• /
• 제24권12호
• /
• pp.1673-1681
• /
• 2021

Item cold start is a well studied problem in the research field of recommender systems. Still, many existing collaborative filters cannot recommend items accurately when only a few user-item interaction data are available for newly introduced items (Cold items). We propose a interaction feature prediction method to mitigate item cold start problem. The proposed method predicts the interaction features that collaborative filters can calculate for the cold items. For prediction, in addition to content features of the cold-items used by state-of-the-art methods, our method exploits the interaction features of k-nearest content neighbors of the cold-items. An attention network is adopted to extract appropriate information from the interaction features of the neighbors by examining the contents feature similarity between the cold-item and its neighbors. Our evaluation on a real dataset CiteULike shows that the proposed method outperforms state-of-the-art methods 0.027 in Recall@20 metric and 0.023 in NDCG@20 metric.