• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.14 no.1
• /
• pp.101-107
• /
• 2014

This paper presents an efficient distributed recommendation system using clustering collaborative filtering algorithm in distributed computing environments. The system was built based on Hadoop distributed computing platform, where distributed Min-hash clustering algorithm is combined with user based collaborative filtering algorithm to optimize recommendation performance. Experiments using Movie Lens benchmark data show that the proposed system can reduce the execution time for recommendation compare to sequential system.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2012.01a
• /
• pp.261-262
• /
• 2012

소셜 네트워크의 부상과 함께 소셜 네트워크를 이용하여 홍보를 하는 소셜 커머스 시장도 커지고 있다. 소셜 커머스의 경우 일정한 인원 이상이 구입을 해야 거래가 성립한다. 그래서 실질적으로 환불이나 반품이 힘들기 때문에 그만큼 상품평이 구매에 미치는 영향이 크다고 볼 수 있다. 하지만 이러한 상품평의 경우에도 개인의 상황이나 취향 등에 따라 상품평이 주는 정보의 방향이 크게 바뀔 수 있다는 단점도 있다. 본 논문에서는 오피니언 마이닝을 이용하여 의미를 추출하고, LIWC를 통해 사용자의 기본 정보 및 심리 등을 파악하여 보다 정확한 고객의 개인별 상황에 맞는 상품 평점을 제시한다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2016.10a
• /
• pp.300-301
• /
• 2016

This paper proposes customized recommendation algorithm to improve the QoS(quality of service) of sport for all sports content uses to user profile and team grade. The proposed recommendation module is based on user profile information, and it recommends suitable team contents to user with Euclidean distance algorithm and preference weights between teams.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2022.01a
• /
• pp.35-38
• /
• 2022

콘텐츠 기반 추천 시스템은 대표적인 추천 모델 방법 중 하나이다. 하지만 콘텐츠 기반 추천 시스템은 사용자 관련 메타데이터를 고려하기보다 내용 관련 메타데이터에만 의존하는 경향이 있다. 본 논문에서는 영화의 특징을 담고 있는 메타데이터를 이용해 추천 시스템을 간단히 구현하고, 추천한 영화와 사용자의 영화 평점을 이용해 추천 시스템의 정확도를 측정하였다. 영화 메타데이터 keywords, genres, cast의 개수를 늘려가며 정확도가 변화하는지 알아보았다. 메타데이터 각각의 개수가 증가하면 정확도도 향상할 것이라고 기대했으나 큰 차이가 나타나지 않았다. 모델 평가 결과, 미세한 차이지만 영화 메타데이터를 상위 3개씩 추출해 영화를 추천했을 때의 정확도가 1.2100318041248186으로 가장 높았다.

• Knowledge Management Research
• /
• v.22 no.3
• /
• pp.273-293
• /
• 2021

Due to the COVID-19 pandemic, the size of the e-commerce has been increased rapidly. This pandemic, which made contact-less communication culture in everyday life made the e-commerce market to be opened even to the consumers who would hesitate to purchase and pay by electronic device without any personal contacts and seeing or touching the real products. Consumers who have experienced the easy access and convenience of the online purchase would continue to take those advantages even after the pandemic. During this time of transformation, however, the size of information source for the consumers has become even shrunk into a flat screen and limited to visual only. To provide differentiated and competitive information on products, companies are adopting AR/VR and steaming technologies but the reviews from the honest users need to be recognized as important in that it is regarded as strong as the well refined product information provided by marketing professionals of the company and companies may obtain useful insight for product development, marketing and sales strategies. Then from the consumer's point of view, if the ratings of reviews are widely diverged how consumers would process the review information before purchase? Are non-converged ratings always unreliable and worthless? In this study, we analyzed how consumer's regulatory focus moderate the attitude to process the diverged information. This experiment was designed as a 2x2 factorial study to see how the variance of product review ratings (high vs. low) for cosmetics affects product attitudes by the consumers' regulatory focus (prevention focus vs. improvement focus). As a result of the study, it was found that prevention-focused consumers showed high product attitude when the review variance was low, whereas promotion-focused consumers showed high product attitude when the review variance was high. With such a study, this thesis can explain that even if a product with exactly the same average rating, the converged or diverged review can be interpreted differently by customer's regulatory focus. This paper has a theoretical contribution to elucidate the mechanism of consumer's information process when the information is not converged. In practice, as reviews and sales records of each product are accumulated, as an one of applied knowledge management types with big data, companies may develop and provide even reinforced customer experience by providing personalized and optimized products and review information.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.20 no.8
• /
• pp.1407-1414
• /
• 2016

In this study, we develop that Location Recommending System using personal emotion information based on Collaborative Filtering. Previous Location Recommending System recommended a place visited by the user of the rating or the pattern of location for the user place. These systems are not high user satisfaction because that dose not consider the user status or have not objectively the information. Using user's personal emotion information to recommend a high-affinity users who have visited the place felt similar emotions objectively can improve user satisfaction with the place. In this study, a user using a mobile application directly register the recognized emotion information using the current position and bio-signal, and using the registered information measuring the similarity of user with a similarity emotion, predicts a preference for the place it is recommended to emotional place. The system consists of a user interface, a database, a recommendation module.

• The Transactions of the Korea Information Processing Society
• /
• v.13 no.8
• /
• pp.382-387
• /
• 2024

Recently, the amount of adult reading has been continuously decreasing, but the consumption of video content is increasing. Accordingly, there is no information on preferences and behavior patterns for new users, and user evaluation or purchase of new books are insufficient, causing cold start problems and data scarcity problems. In this paper, a hybrid book recommendation system based on video content was proposed. The proposed recommendation system can not only solve the cold start problem and data scarcity problem by utilizing the contents of the video, but also has improved performance compared to the traditional book recommendation system, and even high-quality recommendation results that reflect genre, plot, and rating information-based user taste information were confirmed.

• Smart Media Journal
• /
• v.12 no.6
• /
• pp.15-26
• /
• 2023

As face-to-face communication has become difficult due to the COVID-19 pandemic, studies have been conducted to understand the impact of non-face-to-face communication, but there is a lack of research that examines this through messenger application reviews. This study aims to identify the impact of the pandemic through Latent Dirichlet Allocation (LDA) topic modeling by collecting review data of 메신저 applications in the Google Play Store and suggest service strategies accordingly. The study categorized the data based on when the pandemic started and the ratings given by users. The analysis showed that messenger is mainly used by middle-aged and older people, and that family communication increased after the pandemic. Users expressed frustration with the application's updates and found it difficult to adapt to the changes. This calls for a development approach that adjusts the frequency of updates and actively listens to user feedback. Also, providing an intuitive and simple user interface (UI) is expected to improve user satisfaction.

• Journal of Intelligence and Information Systems
• /
• v.29 no.1
• /
• pp.249-263
• /
• 2023

Collaborative Filtering, a representative recommendation system methodology, consists of two approaches: neighbor methods and latent factor models. Among these, the latent factor model using matrix factorization decomposes the user-item interaction matrix into two lower-dimensional rectangular matrices, predicting the item's rating through the product of these matrices. Due to the factor vectors inferred from rating patterns capturing user and item characteristics, this method is superior in scalability, accuracy, and flexibility compared to neighbor-based methods. However, it has a fundamental drawback: the need to reflect the diversity of preferences of different individuals for items with no ratings. This limitation leads to repetitive and inaccurate recommendations. The Adaptive Deep Latent Factor Model (ADLFM) was developed to address this issue. This model adaptively learns the preferences for each item by using the item description, which provides a detailed summary and explanation of the item. ADLFM takes in item description as input, calculates latent vectors of the user and item, and presents a method that can reflect personal diversity using an attention score. However, due to the requirement of a dataset that includes item descriptions, the domain that can apply ADLFM is limited, resulting in generalization limitations. This study proposes a Generalized Adaptive Deep Latent Factor Recommendation Model, G-ADLFRM, to improve the limitations of ADLFM. Firstly, we use item ID, commonly used in recommendation systems, as input instead of the item description. Additionally, we apply improved deep learning model structures such as Self-Attention, Multi-head Attention, and Multi-Conv1D. We conducted experiments on various datasets with input and model structure changes. The results showed that when only the input was changed, MAE increased slightly compared to ADLFM due to accompanying information loss, resulting in decreased recommendation performance. However, the average learning speed per epoch significantly improved as the amount of information to be processed decreased. When both the input and the model structure were changed, the best-performing Multi-Conv1d structure showed similar performance to ADLFM, sufficiently counteracting the information loss caused by the input change. We conclude that G-ADLFRM is a new, lightweight, and generalizable model that maintains the performance of the existing ADLFM while enabling fast learning and inference.

• Journal of KIISE:Computing Practices and Letters
• /
• v.14 no.9
• /
• pp.835-840
• /
• 2008

This paper proposes an enhanced recommendation technique for personalized e-commerce portal analyzing various attitudes of customer. The attitudes are classifies into three types such as "purchasing product", "adding product to shopping cart", and "viewing the product information". We implicitly track customer attitude to estimate the rating of products for recommending products. We classified user groups which have similar preference for each item using implicit user behavior. The preference similarity is estimated using the Cross Correlation Coefficient. Our recommendation technique shows a high degree of accuracy as we use age and gender to group the customers with similar preference. In the experimental section, we show that our method can provide better performance than other traditional recommender system in terms of accuracy.