그룹 추천은 개인이 아닌 그룹의 특성 및 성향을 분석하여 구성원

그룹 추천은 개인이 아닌 그룹의 특성 및 성향을 분석하여 구성원들에게 적합한 정보를 제공하는 추천 방식이다. 하지만 기존의 그룹 추천 방식은 높은 선호도나 선호 횟수에만 기반하여 고른 분포로 높은 선호도를 갖는 관심사를 고려하지 않는 문제점이 있어, 이를 개선하고자 관심사에 대한 평균 선호도와 선호도 편차를 고려한 그룹 추천 방법을 제안한다. 제안 방법은 평균 선호도가 높고 선호도 편차가 적은 관심사들을 추천 결과로 제공해줌으로써 기존 그룹 모델링 전략에 비해 그룹 내 구성원 전체의 성향이 반영되며, 많은 수의 사용자가 형성하는 그룹에서도 높은 성능을 갖는다.

결과 (영어) 1:

Characteristics of the group rather than the individual and the Group recommended to analyze the tendencies of members featured to provide the appropriate information. But the way I prefer the existing number of high affinity or referral group only picked on the basis of having high affinity interests distribution does not consider issues, seeking to improve the average rating and affinity for concern considering the variance Group offers a way to recommend. The proposed method is the average rating, high affinity offers as a result of the deviation is less concerns recommendation allows an existing group members within a group-wide inclination compared to the modeling strategy is reflected in the large number of users to deliver high performance, even at the group forming.

결과 (영어) 2:

Group recommendation is a recommendation how to analyze the characteristics and nature of the group rather than an individual provide the appropriate information to its members. However, existing groups like manner, we propose a group like method based on the average affinity and preference deviation for there is a problem that does not take into account the interest has a high affinity to the even distribution based only on a high preference or favorite number, to improve this concern . The proposed method has a high average rating is provided by giving preference bias is reflected in the tendency of the whole group within an existing group members compared to modeling strategies as a result of recommendation little interest, it has a high performance in a large number of users to form groups.

결과 (영어) 3:

the individual is not recommended to the group of characteristics and tendency of members, for information, to provide the recommended way. however, the existing group of high preference for the good in me, and based on the distribution of the highest preference to get concern to do not consider the problem, improve the concern of 호도와 preference on average in the considered group recommendation method proposal. the proposed method has high preference preference deviation is less concerned about the recommendation to provide the results to the existing group compared with the modelling strategy to reflect the tendency of the members of the group, the group of the 형성하 in high performance.