DOI QR코드

DOI QR Code

양방향 순환 신경망 언어 모델을 이용한 Fuzzy-AHP 기반 영화 추천 시스템

A Fuzzy-AHP-based Movie Recommendation System with the Bidirectional Recurrent Neural Network Language Model

  • 오재택 (공주대학교 컴퓨터공학과) ;
  • 이상용 (공주대학교 컴퓨터공학부)
  • Oh, Jae-Taek (Department of Computer Science & Engineering, Kongju National University) ;
  • Lee, Sang-Yong (Division of Computer Science & Engineering, Kongju National University)
  • 투고 : 2020.09.04
  • 심사 : 2020.12.20
  • 발행 : 2020.12.28

초록

다양한 정보가 대량으로 유통되는 IT 환경에서 사용자의 요구를 빠르게 파악하여 의사결정을 도와줄 수 있는 추천 시스템이 각광을 받고 있다. 그러나 현재 추천 시스템은 사용자의 취향이나 관심사가 바뀌었을 때 선호도가 즉시 시스템에 반영이 되지 않을 수가 있으며, 광고 유도로 인하여 사용자의 선호도와 무관한 아이템이 추천될 수가 있다는 문제점이 있다. 본 연구에서는 이러한 문제점을 해결하기 위해 양방향 순환 신경망 언어 모델을 이용한 Fuzzy-AHP 기반 영화 추천 시스템을 제안하였다. 본 시스템은 사용자의 취향이나 관심사를 명확하고 객관적으로 반영하기 위해 Fuzzy-AHP를 적용하였다. 그리고 사용자가 선호하는 영화를 예측하기 위해 양방향 순환 신경망 언어 모델을 이용하여 실시간으로 수집되는 영화 관련 데이터를 분석하였다. 본 시스템의 성능을 평가하기 위해 그리드 서치를 이용하여 전체 단어 집합의 크기에 대한 학습 모델의 적합성을 확인하였다. 그 결과 본 시스템의 학습 모델은 전체 단어 집합의 크기에 따른 평균 교차 검증 지수가 97.9%로 적합하다는 것을 확인할 수 있었다. 그리고 본 모델은 네이버의 영화 평점 대비 평균 제곱근 오차가 0.66, LSTM 언어 모델은 평균 제곱근 오차가 0.805으로, 본 시스템의 영화 평점 예측성이 더 우수함을 알 수 있었다.

In today's IT environment where various pieces of information are distributed in large volumes, recommendation systems are in the spotlight capable of figuring out users' needs fast and helping them with their decisions. The current recommendation systems, however, have a couple of problems including that user preference may not be reflected on the systems right away according to their changing tastes or interests and that items with no relations to users' preference may be recommended, being induced by advertising. In an effort to solve these problems, this study set out to propose a Fuzzy-AHP-based movie recommendation system by applying the BRNN(Bidirectional Recurrent Neural Network) language model. Applied to this system was Fuzzy-AHP to reflect users' tastes or interests in clear and objective ways. In addition, the BRNN language model was adopted to analyze movie-related data collected in real time and predict movies preferred by users. The system was assessed for its performance with grid searches to examine the fitness of the learning model for the entire size of word sets. The results show that the learning model of the system recorded a mean cross-validation index of 97.9% according to the entire size of word sets, thus proving its fitness. The model recorded a RMSE of 0.66 and 0.805 against the movie ratings on Naver and LSTM model language model, respectively, demonstrating the system's superior performance in predicting movie ratings.

키워드

참고문헌

  1. Y. K. Kim. (2020. 02). Samsung Electronics Laughs at U.S 5G Expansion. Economist, 1522, 34-36.
  2. M. Y. Lee. (2020.02). 'Dream Force 2019' Conference Site Report. Dong-A Business Review, 291, 36-48.
  3. Y. K. Kim. (2019. 10). How far can we trust artificial intelligence?. Economist, 1503, 24-28.
  4. L. Adam. (2019. 10). Facebook wants to become a source for 'TrustWorthy News'. News Week, 1392, 36-37.
  5. J. B. Bae. (2020. 01). Smart Robot Eyes with Deep Fake AI just like the Real Person. Economist, 1519, 50-52.
  6. Y. J. Kim. (2020. 01). The Age of New Gold Rush. Economy Chosun, 332, 10-11.
  7. C. M. Kwon. (2019). Python Machine Learning Perfect Guide. Paju: Wiki books.
  8. S. Sivaprasad, T. Joshi, R. Agrawal & N. Pedanekar. (2018). Multimodal Continuous Prediction of Emotions in Movie using Long-short Term Memory Networks. Proceedings of the 2018 ACM on International Conference on Multimedia Retrieval, 413-419.
  9. G. Liu & X. Wu. (2019). Using Collaborative Filtering Algorithms Combined with Doc2Vec for Movie Recommendation. 2019 IEEE 3rd Information Technology, Networking, Electronics and Automation Control Conference, 1-4.
  10. C. M. Lim, S. D. Kim & D. K. Yoon. (2011). AHP/DEA Decision Model using Triangular Fuzzy Number. Proceedings of the Korean Institute of Industrial Engineers Spring Conference, 2011(5), 612-617.
  11. B. Pang & S. Bai. (2013). An Integrated Fuzzy Synthetic Evaluation Approach for Supplier Selection based on Analytic Network Process. Journal of Intelligent Manufacturing, 24(2013), 163-174. https://doi.org/10.1007/s10845-011-0551-3
  12. E. Kinoshita, T. Oya. (2012). Strategic Decision Making - AHP. Seoul: Cheong Ram
  13. M. Schuster & K. K. Paliwal. (1997). Bidirectional Recurrent Neural Networks. IEEE Transactions on Signal Processing, 45(11), 2673-2681. https://doi.org/10.1109/78.650093
  14. A. Graves, A. Mogamed & G. Hinton. (2013). Speech Recognition with Deep Recurrent Neural Networks. 2013 IEEE International Conference on Acoustics, Speech and Signal Processing, 6645-6649.
  15. C. Francois. (2018). Deep Learning with Python. Seoul: Gilbut.
  16. T. N. Duc, C. M. Tran, P. X. Tan & E. Kamioka. (2019). Bidirectional LSTM for Continuously Predicting QoE in HTTP Adaptive Streaming. Proceedings of the 2019 2nd International Conference on Information Science and Systems, 156-160.
  17. X. Han, Z. Wu, Y. G. Jiang & L. S. Davis. (2017). Learning Fashion Compatibility with Bidirectional LSTMs. Proceedings of the 25th ACM International Conference on Multimedia, 1078-1086.
  18. R. Sebastian, M. Vahid. (2019). Python Machine Learning Second Edition. Seoul: Gilbut.
  19. S. K. Reddy, V. Swaminathan & C. M. Motley. (1998). Exploring the Determinants of Broadway Show Success. Journal of Marketing Research, 35(3), 296-315. https://doi.org/10.2307/3152029
  20. G. T. Cho, Y. G. Cho, H. S. Kang. (2005). The Analytic Hierarchy Process. Seoul: Dong Hyeon.
  21. J. T. Oh & S. Y. Lee. (2020). A Movie Recommendation System based on Fuzzy-AHP and Word2Vec. Journal of Digital Convergence, 18(1), 301-307. https://doi.org/10.14400/JDC.2020.18.1.301
  22. A. C. Muller, S. Guido. (2019). Introduction to Machine Learning with Python. Seoul: Hanbit Media.