Advertisement

How to Find Similar Users in Order to Develop a Cosmetics Recommender System

Conference paper

Abstract

Most shopping web sites allow users to provide product reviews. It has been observed that reviews have a profound effect on item conversion rates. In particular, reviews of cosmetic products have significant impact on purchasing decisions because of the personal nature of such products, and also because of the potential for skin irritation caused by unsuitable items, which is a major consumer concern. In this study, we develop a method for user similarity calculation for a cosmetic review recommender system. To realize such a recommender system, we propose a method for the automatic scoring of various aspects of cosmetic item review texts based on an evaluation expression dictionary curated from a corpus of real-world online reviews. Furthermore, we consider how to calculate user similarity of cosmetic review sites.

Keywords

Automatic review rating Clustering cosmetic products Evaluation expression dictionary Explanation of reviews Recommender system Text mining User similarity 

Notes

Acknowledgements

This work was supported in part by istyle Inc. which provided review data for cosmetic items, and the MEXT Grant-in Aid for Scientific Research(C)(#16K00425, #26330351).

References

  1. 1.
    @cosme, http://www.cosme.net/. Accessed 8 Jan 2017
  2. 2.
    Y. Matsunami, M. Ueda, S. Nakajima, T. Hashikami, S. Iwasaki, J. O’Donovan, B. Kang, Explaining item ratings in cosmetic product reviews, in International MultiConference of Engineers and Computer Scientists ICICWS 2016, (2016), pp. 392–397Google Scholar
  3. 3.
    Y. Matsunami, A. Okuda, M. Ueda, S. Nakajima, User Similarity calculating method for cosmetic review recommender system, lecture notes in engineering and computer science, in Proceedings of The International MultiConference of Engineers and Computer Scientists 2017, Hong Kong, 15–17 Mar 2017, pp. 312–316Google Scholar
  4. 4.
    Amazon.com, http://www.amazon.com/. Accessed 8 Jan 2017
  5. 5.
    Priceprice.com, http://ph.priceprice.com/. Accessed 8 Jan 2017
  6. 6.
    B. Kang, N. Tintarev, J. O’Donovan, Inspection mechanisms for community-based content discovery in microblogs, in IntRS15 Joint Workshop on Interfaces and Human Decision Making for Recommender Systems (http://recex.ist.tugraz.at/intrs2015/) at ACM Recommender Systems 2015, Vienna, Austria, Sep 2015
  7. 7.
    istyle Inc., Site data of @cosme, http://www.istyle.co.jp/business/uploads/sitedata.pdf (in Japanese). Accessed 8 Jan 2017
  8. 8.
    Y. Hamaoka, M. Ueda, S. Nakajima, Extraction of evaluation aspects for each cosmetics item to develop the reputation portal site, in IEICE WI2-2012–15, Feb 2012 (in Japanese), pp. 45–46Google Scholar
  9. 9.
    T. Nihongi, K. Sumita, Analysis and retrieval of the word-of-mouth estimation by structurizing sentences. in Proceeding of the Interaction 2002, (2002) (in Japanese), pp. 175–176Google Scholar
  10. 10.
    J. Yao, H. Idota, A. Harada, The affect of the internet’s word of mouth on buying behavior: from the questionnair survey pf the cosmetics purchase of the women students, in Proceeding of National Conference of JASMIN Spring 2014, Kanagawa, Japan, May 2014 (in Japanese), pp. 231–232Google Scholar
  11. 11.
    I. Titov, R. McDonald, A joint model of text and aspect ratings for sentiment summarization, in 46th Meeting of Association for Computational Linguistics (ACL-08), Columbus, USA, 2008, pp. 308–316Google Scholar
  12. 12.
    M. Nakatsuji, M. Kondo, A. Tanaka, T. Uchiyama, Measuring similarity of users using sentimental tags for items, in The 24th Annual Conference of the Japanese Society for Artificial Intelligence, 2010 Google Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2018

Authors and Affiliations

  • Yuki Matsunami
    • 1
  • Mayumi Ueda
    • 2
  • Shinsuke Nakajima
    • 1
  1. 1.Kyoto Sangyo UniversityKyotoJapan
  2. 2.University of Marketing and Distribution SciencesKobeJapan

Personalised recommendations