• 한국IT서비스학회지
• /
• 제22권1호
• /
• pp.43-56
• /
• 2023

With the development of the e-commerce market, various types of products continue to be released. However, customers face an information overload problem in purchasing decision-making. Therefore, personalized recommendations have become an essential service in providing personalized products to customers. Recently, many studies on GCN-based recommender systems have been actively conducted. Such a methodology can address the limitation in disabling to effectively reflect the interaction between customer and product in the embedding process. However, previous studies mainly use implicit feedback data to conduct experiments. Although implicit feedback data improves the data scarcity problem, it cannot represent customers' preferences for specific products. Therefore, this study proposed a novel model combining explicit and implicit feedback to address such a limitation. This study treats the average ratings of customers and products as the features of customers and products and converts them into a high-dimensional feature vector. Then, this study combines ID embedding vectors and feature vectors in the embedding layer to learn the customer-product interaction effectively. To evaluate recommendation performance, this study used the MovieLens dataset to conduct various experiments. Experimental results showed the proposed model outperforms the state-of-the-art. Therefore, the proposed model in this study can provide an enhanced recommendation service for customers to address the information overload problem.

• 영어어문교육
• /
• 제8권1호
• /
• pp.39-53
• /
• 2002

Recently much research has investigated the role of attention in L2 learning, comparing the effects of explicit learning with those of implicit learning. With this background the research aims at examining the effects explicit focus on form has on L2 learning based on the acquisition of the English article system. The participants were 70 Korean college students who enrolled in English Composition classes. The experimental group received explicit focus on form including grammatical explanation, input enhancement, output practice, and negative evidence (corrective feedback) for two weeks, while the control group was exposed to sufficient input and negative evidence. Completion tasks were administered at the beginning and the end of the semester. In addition, errors in the use of English articles were analysed on their compositions both before and after the different treatments. The analyses of the results show that the explicit focus on form group improved significantly more than the control group, particularly for the definite article 'the', and some changes occurred in the distribution of article errors. These findings suggest that explicit teaching plays a more contributory role than implicit teaching in acquiring L2 knowledge in classroom-based L2 learning.

• 디지털콘텐츠학회 논문지
• /
• 제18권8호
• /
• pp.1561-1566
• /
• 2017

기계 학습과 인공 지능 기술의 발전으로 다양한 응용분야들이 가능해지고 있고, 이중에 추천 시스템은 이미 여러 업체들에서 영화 추천이나 상품 추천 등의 서비스에 적용하여 효과를 보고 있다. 이러한 서비스 중인 추천 시스템들의 대부분은 아이템의 내용을 분석하여 추천하거나 아니면 평점과 같은 직접적인 피드백에 기반하여 시스템을 학습하고 추천하고 있다. 하지만 많은 온라인 쇼핑몰 중에는 아이템의 내용을 분석하는 것이 어렵고, 직접적인 피드백 정보가 없거나 혹은 거의 없어 추천 시스템 구축이 어려운 경우가 많다. 이러한 경우에도 사용자의 상품 조회에 관한 로그 기록들은 어렵지 않게 확보할 수 있고, 로그 기록들만 가지고도 추천 서비스를 제공할 수 있다면 서비스의 질을 향상할 수 있을 것으로 기대된다. 본 논문에서는 사용자의 로그 기록으로부터 암묵적인 피드백인 상품 조회 정보를 추출하고, 암묵적인 피드백에 기반한 추천 시스템을 구현하고, 제안된 시스템은 온라인 반려동물 용품점에 적용하여 확인한다. 즉, 사용자들의 상품조회를 위한 클릭정보만을 활용하여 반려동물 용품 추천 시스템을 구축하여 서비스로 확인한다.

• 스마트미디어저널
• /
• 제12권9호
• /
• pp.89-94
• /
• 2023

추천 시스템은 사용자가 아이템에 남긴 익스플리싯 또는 임플리싯 피드백을 바탕으로 각 사용자가 선호할 법한 아이템들을 추천하는 기술이다. 최근, 추천 시스템에 사용되는 딥 러닝 기반 모델의 사이즈가 커짐에 따라, 높은 추천 정확도를 유지하며 추론 시간은 줄이기 위한 목적의 연구가 활발히 진행되고 있다. 대표적으로 지식증류기법을 이용한 추천 시스템에 관한 연구가 있으며, 지식증류기법이란 큰 사이즈의 모델(즉, 교사)로부터 추출된 지식을 통해 작은 사이즈의 모델(즉, 학생)을 학습시킨 뒤, 학습이 끝난 작은 사이즈의 모델을 추천 모델로서 이용하는 방법이다. 추천 시스템을 위한 지식증류기법들에 관한 기존의 연구들은 주로 임플리싯 피드백 환경만을 대상으로 수행되어 왔었으며, 본 논문에서 우리는 이들을 익스플리싯 피드백 환경에 적용할 경우의 성능 및 정확도를 관찰하고자 한다. 실험을 위해 우리는 총 5개의 최신 지식증류기법들과 3개의 실세계 데이터셋을 사용하였다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2008년도 하계종합학술대회
• /
• pp.641-642
• /
• 2008

In this paper, we propose a new approach for the implicit rating algorithm of finding user's intense and preference to the contents on the web. Although the explicit method dig out the user preference of specific contents based on the user's intervention, we propose the implicit method obtaining the user preference according to the user's behavioral patterns on the web implicitly and automatically without the user's intervention. The implementation results show that the proposed approach is highly valuable for supporting recommender systems in conjunction with the users lifestyle.

• 한국전문물리치료학회지
• /
• 제13권3호
• /
• pp.24-32
• /
• 2006

Theoretical framework of motor learning is used to enhance perceptual motor skill in physical therapy intervention, which can be subdivided into two main types-explicit and implicit. The purpose of this study was to examine whether stroke patients with unilateral brain damage learn implicitly a motor skill using the arm ipsilateral to the damaged hemisphere. Speculation then followed as to the formation of therapeutic plans and instructions provided to patients with stroke. 20 patients with stroke and 20 normal participants were recruited. All the subjects practiced serial reaction time tasks for 30 minutes a day and retention tests on the following day. The tasks and tests involved pressing the corresponding buttons to 4 colored circles presented on a computer screen as quickly and accurately as possible. Patients with stroke responded more slowly than controls. However, both groups showed decreased reaction time in the experimental and retention periods. Also, there was no significant difference between both groups regarding explicit knowledge of consecutive order. Therefore, patients with stoke had the ability to learn implicitly a perceptual motor skill. Prescriptive instruction using implicit and explicit feedback may be beneficial for motor skill learning in physical therapy intervention for patients with brain damage.

• 한국멀티미디어학회논문지
• /
• 제21권11호
• /
• pp.1353-1361
• /
• 2018

With the availability of real-time educational data collection and analysis techniques, the education paradigm is shifting from educator-centric to data-driven lectures. However, most offline and online education frameworks collect students' feedback from question-answering data that can summarize their understanding but requires instructor's attention when students need additional help during lectures. This paper proposes a content restructure recommendation framework based on collected student feedback. We list the types of student feedback and implement a web-based framework that collects both implicit and explicit feedback for content restructuring. With a case study of four-week lectures with 50 students, we analyze the pattern of student feedback and quantitatively validate the effect of the proposed content restructuring measured by the level of student engagement.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권5호
• /
• pp.2381-2399
• /
• 2019

Traditional recommendation algorithms on Collaborative Filtering (CF) mainly focus on the rating prediction with explicit ratings, and cannot be applied to the top-N recommendation with implicit feedbacks. To tackle this problem, we propose a new collaborative filtering approach namely Maximize MAP with Matrix Factorization (MFMAP). In addition, in order to solve the problem of non-smoothing loss function in learning to rank (LTR) algorithm based on pairwise, we also propose a smooth MAP measure which can be easily implemented by standard optimization approaches. We perform experiments on three different datasets, and the experimental results show that the performance of MFMAP is significantly better than other recommendation approaches.

• 한국문헌정보학회지
• /
• 제35권3호
• /
• pp.173-189
• /
• 2001

This paper presents a framework for observable behavior that can be used as a basis for user modeling, and it reports the results of a pair of user studies that examine the joint utility of two specific behaviors. User models can be constructed by hand, or they can be teamed automatically based on feedback provided by the user about the relevance of documents that they have examined. By observing user behavior, it is possible to obtain implicit feedback without requiring explicit relevance judgments. Four broad categories of potentially observable behavior are identified : examine, retain, reference, and annotate, and examples of specific behaviors within a category are further subdivided based on the natural scope of information objects being manipulated . segment object, or class. Previous studies using Internet discussion groups (USENET news) have shown reading time to be a useful source of implicit feedback for predicting a user's preferences. The experiments reported in this paper extend that work to academic and professional journal articles and abstracts, and explore the relationship between printing behavior and reading time. Two user studies were conducted in which undergraduate students examined articles or abstracts from the telecommunications or pharmaceutical literature. The results showed that reading time can be used to predict the user's assessment of relevance, that the mean reading time for journal articles and technical abstracts is longer than has been reported for USENET news documents, and that printing events provide additional useful evidence about relevance beyond that which can be inferred from reading time. The paper concludes with a brief discussion of the implications of the reported results.

• 방송공학회논문지
• /
• 제18권1호
• /
• pp.88-97
• /
• 2013

개인화된 추천을 제공하기 위한 협력 필터링은 추천 시스템에서 성공적으로 활용되어 온 기법이다. 그러나 협력 필터링이 주로 연구 및 적용된 분야들은 사용자로부터의 명시적 피드백이 존재하는 독립된 아이템들을 추천하는 것에 초점을 두고 있다. VOD 서비스 플랫폼에서 개인화된 TV 프로그램을 추천하기 위해서는 해당 도메인의 특성과 제한들을 고려하는 것이 필요하다. 본 논문에서는 TV 프로그램의 시리즈 속성을 이용하여, 선호를 판단하기 힘든 비명시적 피드백인 회별 프로그램 시청기록을 명시적이고 지속적인 프로그램 선호도로 변환하는 방법을 고안하였다. 데이터 수집과 최종 추천은 회별 프로그램 단위로 이루어지면서 협력 필터링 처리 단위는 프로그램으로 변경되어 TV 프로그램 VOD 추천 환경에 가장 적당한 형태로 협력 필터링을 변형 적용하였다. 실험 결과는 고안된 추천 시스템이 단순히 협력 필터링을 적용했을 때보다 높은 정확도와 더 적은 계산량을 가지는 것을 보여준다. 도메인 특화된 이러한 변형은 추천 시스템의 알고리즘 모듈로 구성되어 기존에 알려진 다양한 협력 필터링 기법과 결합하여 사용될 수 있다.