디지털융복합연구 (Journal of Digital Convergence)

  • 제17권1호
  • /
  • Pages.123-130
  • /
  • 2019
  • /
  • 2713-6434(pISSN)
  • /
  • 2713-6442(eISSN)
한국디지털정책학회 (The Society of Digital Policy and Management)
권호 표지
DOI QR코드

DOI QR Code

Word2Vec과 앙상블 합성곱 신경망을 활용한 영화추천 시스템의 정확도 개선에 관한 연구

A Study on the Accuracy Improvement of Movie Recommender System Using Word2Vec and Ensemble Convolutional Neural Networks

  • 강부식 (목원대학교 서비스경영학부)
  • Kang, Boo-Sik (Division of Service Management, Mokwon University)
  • 투고 : 2018.10.16
  • 심사 : 2019.01.20
  • 발행 : 2019.01.28
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

웹 추천기법에서 가장 많이 사용하는 방식 중의 하나는 협업필터링 기법이다. 협업필터링 관련 많은 연구에서 정확도를 개선하기 위한 방안이 제시되어 왔다. 본 연구는 Word2Vec과 앙상블 합성곱 신경망을 활용한 영화추천 방안에 대해 제안한다. 먼저 사용자, 영화, 평점 정보에서 사용자 문장과 영화 문장을 구성한다. 사용자 문장과 영화 문장을 Word2Vec에 입력으로 넣어 사용자 벡터와 영화 벡터를 구한다. 사용자 벡터는 사용자 합성곱 모델에 입력하고, 영화 벡터는 영화 합성곱 모델에 입력한다. 사용자 합성곱 모델과 영화 합성곱 모델은 완전연결 신경망 모델로 연결된다. 최종적으로 완전연결 신경망의 출력 계층은 사용자 영화 평점의 예측값을 출력한다. 실험결과 전통적인 협업필터링 기법과 유사 연구에서 제안한 Word2Vec과 심층 신경망을 사용한 기법에 비해 본 연구의 제안기법이 정확도를 개선함을 알 수 있었다.

One of the most commonly used methods of web recommendation techniques is collaborative filtering. Many studies on collaborative filtering have suggested ways to improve accuracy. This study proposes a method of movie recommendation using Word2Vec and an ensemble convolutional neural networks. First, in the user, movie, and rating information, construct the user sentences and movie sentences. It inputs user sentences and movie sentences into Word2Vec to obtain user vectors and movie vectors. User vectors are entered into user convolution model and movie vectors are input to movie convolution model. The user and the movie convolution models are linked to a fully connected neural network model. Finally, the output layer of the fully connected neural network outputs forecasts of user movie ratings. Experimentation results showed that the accuracy of the technique proposed in this study accuracy of conventional collaborative filtering techniques was improved compared to those of conventional collaborative filtering technique and the technique using Word2Vec and deep neural networks proposed in a similar study.

키워드

DJTJBT_2019_v17n1_123_f0001.png 이미지

Fig. 1. Word2Vec skip-gram model

DJTJBT_2019_v17n1_123_f0002.png 이미지

Fig. 2. Convolution neural network with one convolution layer and one pooling layer

DJTJBT_2019_v17n1_123_f0003.png 이미지

Fig. 3. wDNN configuration

DJTJBT_2019_v17n1_123_f0004.png 이미지

Fig. 4. wCNN configuration

DJTJBT_2019_v17n1_123_f0005.png 이미지

Fig. 5. Procedure of movie recommendation algorithm

DJTJBT_2019_v17n1_123_f0006.png 이미지

Fig. 6. The model architecture of ensemble convolutional neural networks

DJTJBT_2019_v17n1_123_f0007.png 이미지

Fig. 7. The user and movie convolution model

Table 1. MAEs of 10-fold cross validation test

DJTJBT_2019_v17n1_123_t0001.png 이미지

Table 2. Results of pair-wise t-test of wDNN and wCNN

DJTJBT_2019_v17n1_123_t0002.png 이미지

참고문헌

  1. B. S. Kang. (2017). The study of recommendation algorithm's predictive accuracy improvement using structural holes on trust-based social networks. Journal of Knowledge Information Technology and Systems, 12(1), 209-217. https://doi.org/10.34163/jkits.2017.12.1.019
  2. B. S. Kang. (2018). Improving Predictive Accuracy of User-based Collaborative Filtering Using Word2Vec. Journal of Knowledge Information Technology and Systems, 13(1), 169-176. https://doi.org/10.34163/jkits.2018.13.1.017
  3. B. Sarwar, G. Karypis, J. Konstan & J. Ried. (2001). Item-based collaborative filtering recommendation algorithms. Proceedings of the 10th International Conference on World Wide Web, 285-295.
  4. T. Mikolov, I. Sutskever, K. Chen, G. S. Corrado & J. Dean. (2013). Distributed representations of words and phrases and their compositionality. Advances in Neural Information Processing Systems, 26, 3111-3119.
  5. J. M. Kim & J. H. Lee. (2017). Text document classification based on recurrent neural network using word2vec. Journal of Korean Institute of Intelligent Systems, 27(6), 560-565. https://doi.org/10.5391/JKIIS.2017.27.6.560
  6. D. W. Kim & M. W. Koo. (2017). Categorization of korean news articles based on convolutional neural networks using doc2vec and word2vec. Journal of Korean Institute of Information Scientists and Engineers, 44(7), 742-747.
  7. D. Y. Lee, J. C. Jo & H. S. Lim. (2017). User sentiment analysis on amazon fashion product review using word embedding. Journal of the Korean Convergence Society, 8(4), 1-8.
  8. B. Oren & K. Noam. (2016. Sep). Item2vec: neural item embedding for collaborative filtering. 2016 IEEE 26th International Workshop on Machine Learning for Signal Processing, 1-6.
  9. G. S. Jeon, S. G. Kong & Y. S. Cho. (2017). Word2Vec based collaborative filtering for movie Rating prediction. Korea Software Congress 2017, 844-846.
  10. J. K. Sung, S. M. Park, S. Y. Sin, Y. B. Kim & Y. G. Kim. (2017). Deep learning based image retrieval system for o2o shopping mall platform service design, Journal of Digital Convergence, 15(7), 213-222. https://doi.org/10.14400/JDC.2017.15.4.213
  11. S. J. Baek. (2017). Multi-document summarization method based on semantic relationship using VAE, Journal of Digital Convergence, 15(12), 341-347. https://doi.org/10.14400/JDC.2017.15.12.341
  12. R. Devooght & H. Bersini. (2016). Collaborative filtering with recurrent neural networks. arXiv preprint arXiv:1608.07400.
  13. M. H. Kwon, S. E. Kong & Y. S. Choi. (2018). Improving recurrent neural network based recommendations by utilizing embedding matrix. Journal of KIISE, 45(7), 659-666. https://doi.org/10.5626/JOK.2018.45.7.659
  14. B. S. Kang. (2018). Improving accuracy of movie recommender system using word2vec and deep neural networks. Journal of Knowledge Information Technology and Systems, 13(5), 561-568. https://doi.org/10.34163/jkits.2018.13.5.006
  15. B. T. Zhang. (2015). Deep hypernetwork models. Communications of the Korean Institute of Information Scientists and Engineers, 33(8), 11-24.
  16. M. K. Kwon & H. S. Yang. (2017). Performance improvement of object recognition system in broadcast media using hierarchical CNN, Journal of Digital Convergence, 15(3), 201-209. https://doi.org/10.14400/JDC.2017.15.3.201
  17. G. Guo, J. Zhang & N. Yorke-Smith. (2013). A novel bayesian similarity measure for recommender systems. Proceedings of the 23th International Joint Conference on Artificial Intelligence, 2619-2625.
  18. Y. LeCun, L. Bottou, Y. Bengio & P. Haffner. (1998). Gradient-based learning applied to document recognition. Proceedings of the Institute of Electrical and Electronics Engineers, 86(11), 2278-2324. https://doi.org/10.1109/5.726791
  19. Gensim. (2017). https://radimrehurek.com/gensim/.
  20. Librec. (2016). http://www.librec.net/datasets.html.