• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2017년도 춘계학술발표대회
• /
• pp.914-915
• /
• 2017

Importance and usage of the recommender system increases with the increase of information. The accuracy of the system recommendation primarily depends on the data. There is a problem in recommender systems, known as cold start problem. The lack of data about new products and users causes the cold start problem, and the system will not be able to give correct recommendation. This paper deals with cold start problem by comparing product specification and the review of the resembled products. The user, who likes the resembled product of the new one has more probability of taking interest in the new product as well. However, if a user disagreed with resembled product due to some reasons which the user mentioned in the reviews. The new product overcomes that issue, so the user will greatly accept the new product. Therefore, the system needs to recommend new product to those users as well, in this way the cold start problem will get resolved.

• 한국경영과학회:학술대회논문집
• /
• 대한산업공학회/한국경영과학회 2002년도 춘계공동학술대회
• /
• pp.320-325
• /
• 2002

Today many companies offer millions of products to customers. They are faced with a problem to choose particular products . In response to this problem a new marking strategy, recommendation has emerged. Among recommendation technologies collaborative filtering is most preferred. But the performance degrades with the number of customers and products. Namely, collaborative filtering has two major limitations, sparsity and scalability. To overcome these problems we introduced a new recommendation methodology using adjusted product hierarchy, grain. This methodology focuses on dimensionality reduction to improve recommendation quality and uses a marketer's specific knowledge or experience. In addition, it uses a new measure in the neighborhood formation step which is the most important one in recommendation process.

• 디지털산업정보학회논문지
• /
• 제18권4호
• /
• pp.85-96
• /
• 2022

Personalized recommendation technology is one of the most important technologies in electronic commerce environment. It helps users overcome information overload by suggesting information that match user's interests. In e-commerce environment, both mobile device users and smart device users have risen dramatically. It creates new challenges. Our method suggests product information that match user's device interests beyond only user's interests. We propose a device-centered personalized recommendation method. Our method uses both purchase and share behavior for user's devices interests. Moreover, it considers data type preference for each device. This paper presents a new recommendation method and algorithm. Then, an e-commerce scenario with a computer, a smartphone and an AI-speaker are described. The scenario shows our work is better than previous researches.

• 한국지능정보시스템학회:학술대회논문집
• /
• 한국지능정보시스템학회 2001년도 The Pacific Aisan Confrence On Intelligent Systems 2001
• /
• pp.192-197
• /
• 2001

A recommendation system tracks past actions of a group of users to make a recommendation to individual members of the group. The computer-mediated marketing and commerce have grown rapidly nowadays so the concerns about various recommendation procedures are increasing. We introduce a recommendation methodology by which e-commerce sites suggest new products of services to their customers. The suggested methodology is based on web log analysis, product taxonomy, and association rule mining. A product recommendation system is developed based on our suggested methodology and applied to a Korean internet shopping mall. The validity of our recommendation system is discussed with the analysis of a real internet shopping mall case.

• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회 2003년도 추계학술대회 및 정기총회
• /
• pp.191-194
• /
• 2003

As internet commerce grows, many company has begun to use a CF (Collaborative Filtering) as a Recommender System. To achieve an accuracy of CF, we need to obtain sufficient account of voting scores from customers. Moreover, those scores may not be consistent. To overcome this problem, we propose a new recommendation scheme using binary user-item matrix, which represents whether a user purchases a product instead of using the voting scores. Through the experiment regarding this new scheme, a better accuracy is demonstrated.

• 패션비즈니스
• /
• 제27권3호
• /
• pp.50-62
• /
• 2023

The e-commerce fashion market has experienced a remarkable growth, leading to an overwhelming availability of shared information and numerous choices for users. In light of this, chatbots have emerged as a promising technological solution to enhance personalized services in this context. This study aimed to develop user-product attributes for a chatbot-based personalized fashion recommendation service using big data text mining techniques. To accomplish this, over one million consumer reviews from Coupang, an e-commerce platform, were collected and analyzed using frequency analyses to identify the upper-level attributes of users and products. Attribute terms were then assigned to each user-product attribute, including user body shape (body proportion, BMI), user needs (functional, expressive, aesthetic), user TPO (time, place, occasion), product design elements (fit, color, material, detail), product size (label, measurement), and product care (laundry, maintenance). The classification of user-product attributes was found to be applicable to the knowledge graph of the Conversational Path Reasoning model. A testing environment was established to evaluate the usefulness of attributes based on real e-commerce users and purchased product information. This study is significant in proposing a new research methodology in the field of Fashion Informatics for constructing the knowledge base of a chatbot based on text mining analysis. The proposed research methodology is expected to enhance fashion technology and improve personalized fashion recommendation service and user experience with a chatbot in the e-commerce market.

• 지능정보연구
• /
• 제15권4호
• /
• pp.183-200
• /
• 2009

추천시스템에서 가장 많이 활용되고 있는 협업필터링은 고객들의 과거 구매이력을 기반으로 추천하기 때문에 새로이 출시되는 상품을 추천하는 것이 근본적으로 불가능하다. 이와 같은 협업필터링의 한계점을 극복하기 위하여 많은 연구자들은 추천 대상 고객이 선호하는 상품과 유사한 속성을 가진 상품을 추천하는 내용기반 필터링을 협업필터링과 결합한 하이브리드 추천기법을 제시하였다. 그러나 하이브리드 추천기법은 음악, 영화 등 속성 추출이 용이한 일부 상품의 추천에만 활용될 수 있다는 한계가 있다. 따라서 상품 유형에 관계없이 고객에게 신상품을 효과적으로 추천할 수 있는 새로운 접근방법이 제시될 필요가 있다. 본 연구에서는 사회연결망분석에서 관계 및 구조적 특성을 분석하기 위해 널리 활용 되고 있는 중심성 개념을 적용하여 상품간의 구매 관계를 파악한 후 이를 기반으로 신상품을 구매할 가능성이 높은 고객을 찾아 신상품을 추천방법을 제안한다. 추천 프로세스는 구매 유사도 분석, 상품 네트워크 구성, 중심성 분석, 신상품 추천 등 네 단계 절차로 나뉘어진다. 제시한 추천방법의 성능을 평가하기 위하여 국내 유명 백화점 중의 하나인 H백화점의 구매 데이터를 사용하여 실험하였다.

• 지능정보연구
• /
• 제17권3호
• /
• pp.79-98
• /
• 2011

우리는 개인 전자상거래 포탈에서 개인화를 위한 시맨틱 추천 방법을 제안한다. 시맨틱 추천은 제품의 특성(속성)을 이용하여 의미적 유사성 평가를 통해 이루어진다. 정확한 추천을 제공하기 위하여 제품의 시맨틱 유사성은 제품의 평점정보를 포함한다. 또한, 추천기술은 제품의 평점을 평가하여 고객의 다양한 내포된 의향을 분석한다. 고객의 의향은 "구입한 제품", "쇼핑카트에 추가한 제품", "정보를 본 제품"과 같이 세 가지 유형으로 분류 하고 있다. 우리는 제품의 추천을 위한 제품의 평점을 추정하기 위하여 고객의 내재적 의향을 추적할 수 있다. 또한 우리는 정확한 추천을 제공하기 위해 매우 중요한 유효한 세션을 식별하는 유효성 검사 프로세스 세션을 구현하였다. 우리의 추천 기술은 유사한 환경의 고객의 연령별 그룹에서 높은 수준을 정확도를 보여 준다. 본 논문의 실험섹션에서 우리의 제안 추천방식은 기존 고객뿐만 아니라 이전의 구매기록이 없는 새로운 사용자에게도 기존에 잘 알려진 협업 필터링 방법보다 좋은 성능을 보여 주었다.

• 한국경영과학회지
• /
• 제31권2호
• /
• pp.27-39
• /
• 2006

The Collaborative Filtering (CF) is one of the popular techniques for personalization in e-commerce storefronts. For CF-based recommendation, every customer needs to provide subjective evaluation ratings for some products based on his/her preference. Also, if an e-commerce site recommends a new product, some customers should rate it. However, there is no in-depth investigation on the impacts on recommendation performance of two number of ratings, i.e. the number of ratings of an individual customer and the number of ratings of an item, even though these are important factors to determine performance of CF methods. In this study, using publicly available EachMovie data set, we empirically investigate the relationships between the two number of ratings and the performance of CF. For the purpose, three analyses were executed. The first and second analyses were performed to investigate the relationship between the number of ratings of a particular customer and the recommendation performance of CF. In the third analysis, we investigate the relationship between the number of ratings on a particular item and the recommendation performance of CF. From these experiments, we can find that there are thresholds in terms of the number of ratings below which the recommendation performances increase monotonically. That is, the number of ratings of a customer and the number of ratings on an item are critical to the recommendation performance of CF when the number of ratings is less than the thresholds, but the value of the ratings decreases after the numbers of ratings pass the thresholds. The results of the experiments provide insight to making operational decisions concerning collaborative filtering in practice.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권12호
• /
• pp.3798-3814
• /
• 2022

Social recommendation algorithm can alleviate data sparsity and cold start problems in recommendation system by integrated social information. Among them, matrix-based decomposition algorithms are the most widely used and studied. Such algorithms use dot product operations to calculate the similarity between users and items, which ignores user's potential preferences, reduces algorithms' recommendation accuracy. This deficiency can be avoided by a metric learning-based social recommendation algorithm, which learns the distance between user embedding vectors and item embedding vectors instead of vector dot-product operations. However, previous works provide no theoretical explanation for its plausibility. Moreover, most works focus on the indirect impact of social friends on user's preferences, ignoring the direct impact on user's rating preferences, which is the influence of user rating preferences. To solve these problems, this study proposes a user bias drift social recommendation algorithm based on metric learning (BDML). The main work of this paper is as follows: (1) the process of introducing metric learning in the social recommendation scenario is introduced in the form of equations, and explained the reason why metric learning can replace the click operation; (2) a new user bias is constructed to simultaneously model the impact of social relationships on user's ratings preferences and user's preferences; Experimental results on two datasets show that the BDML algorithm proposed in this study has better recommendation accuracy compared with other comparison algorithms, and will be able to guarantee the recommendation effect in a more sparse dataset.