[논문]검색 키워드를 활용한 하이브리드 협업필터링 기반 상품 추천 시스템

이윤주; 원하람; 심재승; 안현철

doi:10.13088/jiis.2020.26.1.151

검색 키워드를 활용한 하이브리드 협업필터링 기반 상품 추천 시스템
A Hybrid Collaborative Filtering-based Product Recommender System using Search Keywords 원문보기

지능정보연구 = Journal of intelligence and information systems, v.26 no.1, 2020년, pp.151 - 166

이윤주 (국민대학교 비즈니스 IT 전문대학원) , 원하람 (국민대학교 비즈니스 IT 전문대학원) , 심재승 (국민대학교 비즈니스 IT 전문대학원) , 안현철

초록
AI-Helper

추천시스템(recommender system)은 고객의 선호도를 예측하여 상품과 서비스를 제공하는 기법으로, 현재 다양한 온라인 서비스에 활용되고 있다. 이와 관련된 많은 선행 연구들은 협업필터링(collaborative filtering)에 기반한 추천시스템을 제안하였는데, 대부분의 경우 고객의 구매 내역 또는 평점 데이터만 사용하여 진행되었다. 오늘날 소비자들은 제품을 구매하는 과정에서 온라인 검색 행동을 하여 관심있는 제품을 찾는다. 그렇기 때문에 검색 키워드 데이터는 고객의 선호도를 파악하는데 매우 유용한 정보일 수 있다. 그러나 지금까지 추천시스템 연구에서 사용되는 경우는 거의 없었다. 이에 본 연구는 고객의 검색 행동에 주목하여 온라인 쇼핑몰 고객의 검색 키워드 데이터와 구매 데이터를 고려한 하이브리드 협업 필터링을 제안하였다. 본 연구는 제안된 모델의 적용 가능성을 검증하기 위해 실제 온라인 쇼핑몰 데이터를 사용하여 성능을 검증하였다. 연구 결과, 추천 상품의 개수가 많아질수록 고객의 검색 키워드를 기반으로 구축된 협업필터링의 추천 성능이 향상되는 반면 일반적인 협업필터링의 성능은 추천된 상품의 개수가 많아질수록 점차 감소함을 발견하였다. 따라서 본 연구는 검색 키워드 데이터를 활용한 하이브리드 협업필터링이 고객의 선호도를 반영한 추천할 수 있으며, 구매이력 데이터의 정보부족을 해결할 수 있음을 확인하였다. 이는 기존의 정량 데이터만을 활용한 추천 시스템이 아닌, 비정형 데이터인 텍스트를 사용함으로써 새로운 하이브리드 협업필터링 구축 방법을 제안했다는 점에서 의의가 있다.

Abstract ▼ AI-Helper

A recommender system is a system that recommends products or services that best meet the preferences of each customer using statistical or machine learning techniques. Collaborative filtering (CF) is the most commonly used algorithm for implementing recommender systems. However, in most cases, it only uses purchase history or customer ratings, even though customers provide numerous other data that are available. E-commerce customers frequently use a search function to find the products in which they are interested among the vast array of products offered. Such search keyword data may be a very useful information source for modeling customer preferences. However, it is rarely used as a source of information for recommendation systems. In this paper, we propose a novel hybrid CF model based on the Doc2Vec algorithm using search keywords and purchase history data of online shopping mall customers. To validate the applicability of the proposed model, we empirically tested its performance using real-world online shopping mall data from Korea. As the number of recommended products increases, the recommendation performance of the proposed CF (or, hybrid CF based on the customer's search keywords) is improved. On the other hand, the performance of a conventional CF gradually decreased as the number of recommended products increased. As a result, we found that using search keyword data effectively represents customer preferences and might contribute to an improvement in conventional CF recommender systems.

주제어

표/그림 (5)

그림 Suggested Model
그림 Representation of Customers and Search Keywords
그림 An Example of a Customer-Item Matrix
표 F1-Scores of the models
그림

AI 본문요약
AI-Helper

* AI 자동 식별 결과로 적합하지 않은 문장이 있을 수 있으니, 이용에 유의하시기 바랍니다.

문제 정의

In addition,unlike customer reviews, search keywords arecomposed of nouns, so data pre-processing is lessdifficult and the amount of information lost duringpre-processing is small. Therefore, this study usedsearch keyword data to address the lack ofinformation in purchase history and ratings data.
This research focuses on customer searchkeyword data from online shopping malls asunstructured data to be used in the model.According to Kang and Choi (2019), customersearch behavior is based on the action of collectinginformation on products of interest, which mayindicate customer preferences.

제안 방법

We intend to derive the similarities betweencustomers using Doc2Vec and combine them withCF. In addition, the proposed model is expected tosolve the cold start problem, which is a limitationof the conventional CF methodology, and toprovide more accurate recommendationperformance than the conventional method inwhich recommendations are based only onpurchase history or ratings.
In this paper, we propose a hybrid collaborativefiltering (CF) methodology using the searchkeyword data of online shopping mall customers.We determined that if the keywords searched byvarious customers are similar, they representdifferent customers that have similar preferences.
In this study, customers’ purchasehistories, customers’ search keywords, membershipinformation, session information, and productclassification data were used.
This research proposed a novel hybrid CF thatis designed to utilize customers’ search keywordsdata. The model proposed in this study solved thescarcity problem encountered by recommendersystems based on search keyword data and notonly found customer preferences through thesearch data but also proposed a newrecommendation system using a word embeddingtechnique.
This research proposed a novel hybrid CF thatis designed to utilize customers’ search keywordsdata.
This research was conducted to compare thequality of user-based CF that derives similarities customers through customer purchasehistory data using the feature of Doc2Vec-basedCF that derives similarities between customersthrough customer search keyword data.
With this background, this study proposes a newrecommender system by deriving new customerpreferences based on customer search keyworddata―which have not been utilized often inprevious research―and analyzing them using ahybrid CF method combined with Doc2Vec.

대상 데이터

The data used in this research were providedfrom the 5th L.POINT Big Data Competitionhosted by Lotte Members, L.POINT (LotteMembers, 2019). The L.

이론/모형

They collected apphistory and app description data from Apple’sonline app store and Google’s Play store andanalyzed these unstructured data using theDoc2Vec technique.

성능/효과

Based on previous research that indicatedcustomer preferences through search keywords(Gang and Choi, 2019), an empirical analysis of this study’s proposed model demonstrated that asthe number of recommended products increased,the proposed model performed better than theconventional CF model.

참고문헌 (35)

Adomavicius, G., and A. Tuzhilin, "Context-aware recommender systems," In Recommender Systems Handbook, Springer, Boston, MA, 2011
Balabanovic, M., and Y. Shoham, "Fab: content-based, collaborative recommendation," Communications of the ACM, Vol.40, No.3 (1997), 66-72.

상세보기
Billsus, D., and M. J. Pazzani, "Learning Collaborative Information Filters," In International Conference of Machine Learning, Vol. 98, (1998), 46-54.
Breese, J. S., D. Heckerman, and C. Kadie, "Empirical analysis of predictive algorithms for collaborative filtering," Proceedings of the Fourteenth conference on Uncertainty in artificial intelligence, (1998), 43-52.
Cho S.-e., and H.-s. Lim, "A Study on Product Recommendation System Based on User Search keyword," Journal of Digital Contents Society, Vol.20, No.2(2019), 315-320.

상세보기
Cho, S.-y., J.-e. Choi, K.-h. Lee, and H.-w. Kim, "An Online Review Mining Approach to a Recommendation System," Information Systems Review, Vol.17, No.3(2015), 95-111.

원문보기 상세보기
Cho, Y. H., J. K. Kim, D. H. Ahn, and H. A. Lee, "An Explanation based Recommender System using Collaborative Filtering: WebCF-Exp," Korean Management Review, Vol.35, No.2 (2006), 493-19.
Choi D.-j., J.-y. Park, S.-b. Park, J.-t. Lim, J.-o Song, K.-s. Bok, and J.-s. Yoo, "Personalized Recommendation Considering Item Confidence in E-Commerce," Journal of the Korea Contents Association, Vol.19, No.3(2019), 171-182.

원문보기 상세보기
Choi, S. B. and H. C. Ahn, "A study on the improvement of collaborative filtering prediction accuracy using online user review sentiment analysis," In Proceedings of the Korea Intelligent Information System Society Conference, (2017), 30-31.
Kang, E. J. and Y. S. Choi, "Influence of information retrieval using A.I speaker on online purchasing experience - based on AISAS model," In Proceedings of HCI Korea 2019, (2019), 425-430
Karvelis, P., D., Gavrilis, G., Georgoulas, and C. Stylios, "Topic recommendation using Doc2Vec," In 2018 IEEE International Joint Conference on Neural Networks (IJCNN), (2018), 1-6.
Ki, H., J.-h. Lee, H.-w. Park, M.-j. Chae, S.-w. Choi and J. Park, "Inferring User Traits from Applications Installed on a Smart Phone," Journal of KIISE: Software and Applications, Vol.45, No.12(2018), 1240-1249.
Jeong, J., M. Jee, M. Go, H. Kim, H. Lim, Y. Lee, and W. Kim, "Related Documents Classification System by Similarity between Documents," Journal of Broadcast Engineering, Vol.24, No.1(2019), 77-86.

원문보기 상세보기
Jeong J.-w., W.-s. Hwang, H.-j. Lee, S.-w. Kim, "Recommendation Exploiting Search-Keywords in Online Shopping," Proceedings on KIISE Conference. Vol. 39. No. 2(2012), 95-97
Kim, H., S. Chae, J. Yoo, and S. Bae, "Online shopping mall customer purchasing type clustering and purchasing power evaluation model," Proceedings on Korean Institute of Industrial Engineers, 2019, 2597-2616
Kim, K. S., "A hybrid collaborative filtering algorithm for personalized recommendations and its application to the internet electronic commerce," The Journal of Internet Electronic Commerce Research, Vol.8, No.4(2008), 1-20.

상세보기
Kim, M. G. and K.-j. Kim, "Recommender Systems using Structural Hole and Collaborative Filtering," Journal of Intelligence and Information Systems, Vol.20, No.4(2014), 107-120.

원문보기 상세보기
Kim, M., N. G. Kim, and I. H. Jung, "A Methodology for Extracting Shopping-Related Keywords by Analyzing Internet Navigation Patterns," Journal of Intelligence and Information Systems, Vol.20, No.2(2014), 123-136.

원문보기 상세보기
Ku, M. J. and H. C. Ahn, "A Hybrid Recommender System based on Collaborative Filtering with Selective Use of Overall and Multicriteria Ratings," Journal of Intelligence and Information Systems, Vol.24, No.2(2018), 85-109.

원문보기 상세보기
Le, Q. and T., Mikolov, "Distributed representations of sentences and documents," In Proceedings of the International Conference on Machine Learning, (2014), 1188-1196.
Lee, H. S. and P. S. Kim, "The Effect of Consumer's Technology Acceptance and Resistance on Intention to Use of Artificial Intelligence (AI)," Korean Management Review, Vol.48, No.5(2019), 1195-1219

상세보기
Lee, R. K., N. H. Chung, and T. H. Hong, "Developing the online reviews based recommender models for multi-attributes using deep learning," Journal of Information Systems, Vol.28, No.1(2019), 97-114.

원문보기 상세보기
Lee Y. S., K. C. Cha, and S.-h. Kim. "Internet Search Behavior and Box Office Performance," Korean Management Review, Vol.45, No.5 (2016), 1501-1526.
Mikolov, T., I. Sutskever, K. Chen, G. Corrado, and J. Dean, "Distributed representations of words and phrases and their compositionality," Advances in Neural Information Processing Systems, (2013), 3111-3119.
Nandi, R. N., M. A., Zaman, T., Al Muntasir, S. H., Sumit, T., Sourov, and M. J. U. Rahman, "Bangla news recommendation using doc2vec," In 2018 International Conference on Bangla Speech and Language Processing (ICBSLP), (2018), 1-5.
Park, J., Y. H. Cho, and J. K. Kim, "Social Network : A Novel Approach to New Customer Recommendations," Journal of Intelligence and Information Systems, Vol.15, No.1(2009), 123-140.
Park, J., and N. Kim, "Multi-Vector Document Embedding Using Semantic Decomposition of Complex Documents," Journal of Intelligence and Information Systems, Vol.25, No.3 (2019), 19-41.

원문보기 상세보기
Phi, V. T., L., Chen, and Y. Hirate, "Distributed representation based recommender systems in e-commerce," In DEIM Forum, (2016).
Sarwar, B., G., Karypis, J. A. Konstan, and J. Riedl, "Analysis of Recommendation Algorithms for E-commerce," Proceedings of ACM E-commerce 2000 Conference, (2000), 158-167.
Shim, J., H. R. Won, and H. Ahn, "Doc2vec-based intelligent fake news classification model using domestic news articles," Proceedings of the Korea Intelligent Information System Society Conference, 1-3.
Shin J. H., J. H. Song, K. S. Bok and J. S. Yoo, "Personalized Travel Destination Recommendation Scheme through Hybrid Collaborative Filtering," Proceedings of the Korea Contents Association, (2018), 383-384.
Stiebellehner, S., J., Wang, and S., Yuan, "Learning Continuous User Representations through Hybrid Filtering with doc2vec," arXiv preprint arXiv 1801.00215., (2017).
Takacs, G., I., Pilaszy, B. Nemeth, and D., Tikk, "Scalable collaborative filtering approaches for large recommender systems." Journal of Machine Learning Research, (2009), 623-656.
Wu, C., and M. Yan, "Session-aware information embedding for e-commerce product recommendation," In Proceedings of the 2017 ACM on Conference on Information and Knowledge Management, (2017), 2379-2382.
Zhang, X., J., Liu, B., Cao, Q., Xiao, and Y. Wen, "Web Service Recommendation via Combining Doc2Vec-Based Functionality Clustering and DeepFM-Based Score Prediction," In 2018 IEEE International Conference on Parallel & Distributed Processing with Applications, Ubiquitous Computing & Communications, Big Data & Cloud Computing, Social Computing & Networking, Sustainable Computing & Communications (ISPA/IUCC/BDCloud/SocialCom/SustainCom), (2018), 509-516.

저자의 다른 논문 :

표제어: PCR

동의어: Packet Collision Rate

용어 설명 출처 목록 (6)

용어 설명: PCR은 세균 특이성이 있는 primer를 이용하여 적은 수의 세균이 있을지라도 쉽게 검출할 수 있는 유용한 방법이며, 이를 이용하여 구강 내 치면세균막이나 타액에서 직접 세균을 검출할 수 있게 되었다[8].

내보내기 구분	파일저장 인쇄 메일전송
구성항목	기본정보 상세정보 관리번호, 논문명, 저널/프로시딩명, 저자 , 발행년, 권, 호, 시작페이지, 끝페이지, 발행기관 관리번호, 논문명, 대등논문명, 저자 , 저널/프로시딩명, 발행기관, 발행년, 발행언어, 권, 호, 시작페이지, 끝페이지, ISBN, ISSN, 주제분야, 키워드, 초록(한글), 초록(영문), 저자(소속기관)
저장형식	Text(ASCII format) Excel format RefWorks Direct Export RIS format (for Reference Manager, ProCite, EndNote), Scholar's Aids, Mendeley
메일정보	받는사람 (필수) @ 보내는사람 (선택) @ 제목 내용 KISTI 검색결과 이메일 서비스
안내	총 건의 자료가 검색되었습니다. 다운받으실 자료의 인덱스를 입력하세요. (1-10,000) 검색결과의 순서대로 최대 10,000건 까지 다운로드가 가능합니다. 데이타가 많을 경우 속도가 느려질 수 있습니다.(최대 2~3분 소요) 다운로드 파일은 UTF-8 형태로 저장됩니다. 파일의 내용이 제대로 보이지 않을실 때는 웹브라우저 상단의 보기 -> 인코딩 -> 자동선택 여부를 확인하십시오. ~ Text(ASCII format) Excel format

연합인증