• Journal of Internet Technology
• /
• 제19권6호
• /
• pp.1691-1698
• /
• 2018

In this paper, we are to address the problem of item recommendations to users in shopping malls selling several different kinds of items, e.g., daily necessities such as cosmetics, detergent, and food ingredients. Most of current recommendation algorithms are developed for sites selling only one kind of items, e.g., music or movies. To devise efficient recommendation algorithms suitable for repetitively purchasing items, we give a method to implicitly assign ratings for these items by making use of repetitive purchase counts, and then use these ratings for the purpose of recommendation prediction with the help of user-based collaborative filtering and item-based collaborative filtering algorithms. We also propose associate item-based recommendation algorithm. Items are called associate items if they are frequently bought by users at the same time. If a user is to buy some item, it is reasonable to recommend some of its associate items. We implement user-based (item-based) collaborative filtering algorithm and associate item-based algorithm, and compare these three algorithms in view of the recommendation hit ratio, prediction performance, and recommendation coverage, along with computation time.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2003년도 컴퓨터소사이어티 추계학술대회논문집
• /
• pp.175-178
• /
• 2003

Two defects have been pointed out in existing user-based collaborative filtering such as sparsity and scalability, and the research has been also made progress, which tries to improve these defects using item-based collaborative filtering. Actually there were many results, but the problem of sparsity still remains because of being based on an explicit data. In addition, the issue has been pointed out. which attributes of item arenot reflected in the recommendation. This paper suggests a recommendation method using nave Bayesian algorithm in hybrid user and item-based collaborative filtering to improve above-mentioned defects of existing item-based collaborative filtering. This method generates a similarity table for each user and item, then it improves the accuracy of prediction and recommendation item using naive Bayesianalgorithm. It was compared and evaluated with existing item-based collaborative filtering technique to estimate the accuracy.

• 대한설비관리학회지
• /
• 제23권4호
• /
• pp.57-64
• /
• 2018

User-based collaborative filtering is a method of recommending an item to a user based on the preference of the neighbor users who have similar purchasing history to the target user. User-based collaborative filtering is based on the fact that users are strongly influenced by the opinions of other users with similar interests. Item-based collaborative filtering is a method of recommending an item by comparing the similarity of the user's previously preferred items. In this study, we create a recommendation model using user-based collaborative filtering and item-based collaborative filtering with consumer's consumption data. Collaborative filtering is performed by using RFM (recency, frequency, and monetary) technique with purchasing data to recommend items with high purchase potential. We compared the performance of the recommendation system with the purchase amount and the performance when applying the RFM method. The performance of recommendation system using RFM technique is better.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권12호
• /
• pp.5826-5841
• /
• 2019

In our previous research, we proposed a tag emotion-based item recommendation scheme. The ternary associations among users, items, and tags are described as a three-order tensor in order to capture the emotions in tags. The candidates for recommendation are created based on the latent semantics derived by a high-order singular value decomposition technique (HOSVD). However, the tensor is very sparse because the number of tagged items is smaller than the amount of all items. The previous research do not consider the previous behaviors of users and items. To mitigate the problems, in this paper, the item-based collaborative filtering scheme is used to build an extended data. We also apply the probabilistic ranking algorithm considering the user and item profiles to improve the recommendation performance. The proposed method is evaluated based on Movielens dataset, and the results show that our approach improves the performance compared to other methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권1호
• /
• pp.1-19
• /
• 2021

The graph-based tag recommendation algorithm FolkRank can effectively utilize the relationships between three entities, namely users, items and tags, and achieve better tag recommendation performance. However, FolkRank does not consider the internal relationships of user-user, item-item and tag-tag. This leads to the failure of FolkRank to effectively map the tagging behavior which contains user neighbors and item neighbors to a tripartite graph. For item-item relationships, we can dig out items that are very similar to the target item, even though the target item may not have a strong connection to these similar items in the user-item-tag graph of FolkRank. Hence this paper proposes an improved FolkRank algorithm named FolkRank++, which fully considers the user-user and item-item internal relationships in tag recommendation by adding the correlation information between users or items. Based on the traditional FolkRank algorithm, an initial weight is also given to target user and target item's neighbors to supply the user-user and item-item relationships. The above work is mainly completed from two aspects: (1) Finding items similar to target item according to the attribute information, and obtaining similar users of the target user according to the history behavior of the user tagging items. (2) Calculating the weighted degree of items and users to evaluate their importance, then assigning initial weights to similar items and users. Experimental results show that this method has better recommendation performance.

• 인터넷정보학회논문지
• /
• 제19권3호
• /
• pp.7-14
• /
• 2018

As SNS(Social Network Service) becomes a part of our life, new information can be derived through various information provided by SNS. Through the public timeline analysis of SNS, we can extract the latest tour trends for the public and the intimacy through the social relationship analysis in the SNS. The extracted intimacy can also be used to make the personalized recommendation by adding the weights to friends with high intimacy. We apply SNS elements such as analyzed latest trends and intimacy to item-based collaborative filtering techniques to achieve better accuracy and satisfaction than existing travel recommendation services in a new way. In this paper, we propose a social travel recommendation system using item - based collaborative filtering.

• 한국빅데이터학회지
• /
• 제2권1호
• /
• pp.41-48
• /
• 2017

최근 추천 시스템의 인기와 함께 추천 시스템의 알고리즘의 성능에 대한 평가가 중요해 졌다. 본 연구는 영화 데이터에서 다양한 알고리즘 중 어떤 알고리즘의 효과적인지 판단하기 위하여 모델링과 RMSE를 통한 모델 검증을 하였다. 본 연구의 데이터는 무비렌즈의 평가 데이터 10만건을 활용하여 피어슨 상관계수를 활용한 사용자 기반 협업 필터링, 코사인 상관계수를 활용한 아이템 기반 협업 필터링 그리고 특이 값분해를 활용한 아이템 기반 협업 필터링 모델을 만들었다. 세가지 추천 모델로 평점을 예측한 결과 사용자 기반 협업 필터링보다 아이템 기반 협업 필터링의 정확도가 월등히 높은 것을 확인했고, 행렬 분해를 사용했을 때 더 정확한 추천을 할 수 있었다.

• 지능정보연구
• /
• 제20권2호
• /
• pp.73-92
• /
• 2014

기존 협업 필터링 기법은 사용자들의 아이템에 대한 선호도를 기반으로 유사 아이템 집합 또는 유사 사용자 집합을 구성하고, 이를 이용해 예측된 사용자의 특정 아이템에 대한 선호도를 기반으로 추천을 수행한다. 이로 인해, 사용자 선호도 정보가 부족하게 되면, 유사 아이템 사용자 집합의 신뢰도가 낮아지고, 추천 서비스의 신뢰도 또한 따라서 낮아진다. 또한, 서비스의 규모가 커질수록, 유사 아이템, 사용자 집합의 생성에 걸리는 시간은 기하급수적으로 증가하고 추천서비스의 응답시간 또한 그에 따라 증가하게 된다. 위와 같은 문제점을 해결하기 위해 본 논문에서는 적응형 군집화 기법을 제안하고 이를 적용한 협업 필터링 기법을 제안하고 있다. 이 기법은 크게 네 가지 방법으로 이루어진다. 첫째, 사용자와 아이템의 특성 벡터를 기반으로 사용자와 아이템 각각을 군집화 하여, 기존 협업 필터링 기법에서 유사 아이템, 사용자 집합을 생성하는데 소요되는 시간을 절약하며, 사용자 선호도 정보만을 이용한 부분 집합 생성보다 추천의 신뢰도를 높이고, 초기 평가 문제와 초기 이용자 문제를 일부 해소한다. 둘째, 미리 구성된 사용자와 아이템의 군집을 기반으로 군집간의 선호도를 이용해 추천을 수행한다. 사용자가 속한 군집의 선호도가 높은 순서대로 아이템 군집을 조회하여 사용자에게 제공할 아이템 목록을 구성하여, 추천 시스템의 부하 대부분을 모델 생성 단계에서 부담하고 실제 수행 시 부하를 최소화한다. 셋째, 누락된 사용자 선호도 정보를 사용자와 아이템 군집을 이용하여 예측함으로써 협업 필터링 추천 기법의 사용자 선호도 정보 희박성으로 인한 문제를 해소한다. 넷째, 사용자와 아이템의 특성 벡터를 사용자의 피드백에 따라 학습시켜 아이템과 사용자의 정성적 특성 정량화의 어려움을 해결한다. 본 연구의 검증은 기존에 제안되었던 하이브리드 필터링 기법들과의 성능 비교를 통해 이루어졌으며, 평가 방법으로는 평균 절대 오차와 응답 시간을 이용하였다.

• ETRI Journal
• /
• 제46권2호
• /
• pp.277-289
• /
• 2024

In this paper, we describe a neural network-based application that recommends multiple items using dialog context input and simultaneously outputs a response sentence. Further, we describe a multi-item recommendation by specifying it as a set of clothing recommendations. For this, a multimodal fusion approach that can process both cloth-related text and images is required. We also examine achieving the requirements of downstream models using a pretrained language model. Moreover, we propose a gate-based multimodal fusion and multiprompt learning based on a pretrained language model. Specifically, we propose an automatic evaluation technique to solve the one-to-many mapping problem of multi-item recommendations. A fashion-domain multimodal dataset based on Koreans is constructed and tested. Various experimental environment settings are verified using an automatic evaluation method. The results show that our proposed method can be used to obtain confidence scores for multi-item recommendation results, which is different from traditional accuracy evaluation.

• Journal of Information Technology Applications and Management
• /
• 제21권4호
• /
• pp.203-223
• /
• 2014

This study proposes a novel context-aware recommender system, which is designed to recommend the items according to the customer's responses to the previously recommended item. In specific, our proposed system predicts the user's emotional state from his or her responses (such as facial expressions and movements) to the previous recommended item, and then it recommends the items that are similar to the previous one when his or her emotional state is estimated as positive. If the customer's emotional state on the previously recommended item is regarded as negative, the system recommends the items that have characteristics opposite to the previous item. Our proposed system consists of two sub modules-(1) emotion prediction module, and (2) responsive recommendation module. Emotion prediction module contains the emotion prediction model that predicts a customer's arousal level-a physiological and psychological state of being awake or reactive to stimuli-using the customer's reaction data including facial expressions and body movements, which can be measured using Microsoft's Kinect Sensor. Responsive recommendation module generates a recommendation list by using the results from the first module-emotion prediction module. If a customer shows a high level of arousal on the previously recommended item, the module recommends the items that are most similar to the previous item. Otherwise, it recommends the items that are most dissimilar to the previous one. In order to validate the performance and usefulness of the proposed recommender system, we conducted empirical validation. In total, 30 undergraduate students participated in the experiment. We used 100 trailers of Korean movies that had been released from 2009 to 2012 as the items for recommendation. For the experiment, we manually constructed Korean movie trailer DB which contains the fields such as release date, genre, director, writer, and actors. In order to check if the recommendation using customers' responses outperforms the recommendation using their demographic information, we compared them. The performance of the recommendation was measured using two metrics-satisfaction and arousal levels. Experimental results showed that the recommendation using customers' responses (i.e. our proposed system) outperformed the recommendation using their demographic information with statistical significance.