추천시스템은 정보 과부하의 문제를 해결하기 위해 폭넓게 사용되어지고 있다. 지난 수십년 동안 다양한 추천시스템이 정보량이 그것을 처리할 수 있는 능력보다 더 빠르게 증가하게 됨에 따라 개발되어져 왔다. 이 같은 상황에서 본 연구의 목적은 기 개발된 추천시스템을 분석하여 시스템적 관점을 제공하고 이를 구현하는데 따르는 기본적인 이슈들을 밝히는 것이다. 이를 통하여 추천시스템의 개선을 위한 유용한 정보를 제안하며, 시스템 개발자들에게는 그러한 시스템을 개선하기 위한 아이디어를 제공하고자 한다. 특히 본 연구는 추천시스템의 이론적 관점에 집중하는데, 이를 위해 과거 추천시스템의 도메인과 목표, 주요 방법 및 평가 방법에 대해서 다루고자 하며, 이 결과는 통계치나 도표 등의 형태로 보이려고 한다.
추천시스템은 정보 과부하의 문제를 해결하기 위해 폭넓게 사용되어지고 있다. 지난 수십년 동안 다양한 추천시스템이 정보량이 그것을 처리할 수 있는 능력보다 더 빠르게 증가하게 됨에 따라 개발되어져 왔다. 이 같은 상황에서 본 연구의 목적은 기 개발된 추천시스템을 분석하여 시스템적 관점을 제공하고 이를 구현하는데 따르는 기본적인 이슈들을 밝히는 것이다. 이를 통하여 추천시스템의 개선을 위한 유용한 정보를 제안하며, 시스템 개발자들에게는 그러한 시스템을 개선하기 위한 아이디어를 제공하고자 한다. 특히 본 연구는 추천시스템의 이론적 관점에 집중하는데, 이를 위해 과거 추천시스템의 도메인과 목표, 주요 방법 및 평가 방법에 대해서 다루고자 하며, 이 결과는 통계치나 도표 등의 형태로 보이려고 한다.
Recommendation systems are widely used to help deal with the problem of information overload. Over the past decades, a variety of recommendation systems have been developed as the amount of information in the world increases far more quickly than our ability to process it. This paper aims to analyze...
Recommendation systems are widely used to help deal with the problem of information overload. Over the past decades, a variety of recommendation systems have been developed as the amount of information in the world increases far more quickly than our ability to process it. This paper aims to analyze existing developed recommendation systems, provide systemic review, and present some basic issues on improvement action. Through this, we also suggest useful implications for better recommendation systems and give some ideas to recommendation system developers to improve their system. Especially, this study focuses on researches on recommendation system. In our research, we analyze the studies along with four different keys dimensions : their domain, objective, underlying model, and evaluation method of recommendation systems and portray the results as statistics or statistical graphics or table form.
Recommendation systems are widely used to help deal with the problem of information overload. Over the past decades, a variety of recommendation systems have been developed as the amount of information in the world increases far more quickly than our ability to process it. This paper aims to analyze existing developed recommendation systems, provide systemic review, and present some basic issues on improvement action. Through this, we also suggest useful implications for better recommendation systems and give some ideas to recommendation system developers to improve their system. Especially, this study focuses on researches on recommendation system. In our research, we analyze the studies along with four different keys dimensions : their domain, objective, underlying model, and evaluation method of recommendation systems and portray the results as statistics or statistical graphics or table form.
* AI 자동 식별 결과로 적합하지 않은 문장이 있을 수 있으니, 이용에 유의하시기 바랍니다.
가설 설정
2：Novel approach：The goal is to propose an innovative method for recommendation systems regardless of domain.
제안 방법
In this paper, we have performed an analysis for recommendation systems along four different key dimensions：their domain, their objective, their underlying method, and their method of evaluation of recommendation systems. And then on the basis of the result, we provide systemic review and present some basic issues on improvement action.
And they also try to improve certain recommendation system itself through suggesting alternatives to overcome its shortcoming and redesign data source used for recommendation such as user requirement and historical preference of user. Second, researches in adopting methods to new domain dimension propose new recommendation systems applied existing recommendation methods to new area such as specific user group, specific item or content, marketing campaign, mobile environment, peer to peer Architecture and team staffing. Lastly, creative approaches are introduced to develop recommendation system in novel approach dimension.
대상 데이터
First we used the National Digital Science Library (NDSL) system, which is the biggest journal indexing database for a scientific research in Korea. The database provides 62,704 journals and 194,534 proceedings. It also provides 22,571,483 full-text articles itself - from 22,813 journals and 8,180 proceedings, respectively- and provides 41,079,468 links within the citation/abstract records to the electronic version of the publication.
이론/모형
Afterwards, we added some relevant papers which were obtained by checking reference lists. Finally, for the study quality, we confirmed our list with a double-blind method. As a result, we selected 85 papers for our meta-analysis.
This research applies a technique classification for studying effectiveness (Hundhausen, 1996) of analyzing the evaluation method of the proposed recommendation system. All of the papers considered in our study make use of one or more the techniques presented in [Figure 4].
성능/효과
Fifth, today’s recommendation systems are overall lack robustly developed personalized services.
Fourth, all of current recommendation systems only recommend what people would like to. However, when people make a decision, sometimes notrecommend list is very important.
Hence, the method to verify recommendation system itself is needed and it should be distinguished a simple accuracy metric. Third, existing developed recommendation system are mostly based solely on historical rating data and recommend the most similar one to what they chose in the past. However, there are different kinds of cases in real world.
AI-Helper
※ AI-Helper는 부적절한 답변을 할 수 있습니다.