Journal of the Korea Convergence Society (한국융합학회논문지)

Korea Convergence Society (한국융합학회)
권호 표지
DOI QR코드

DOI QR Code

A Study on SNS Reviews Analysis based on Deep Learning for User Tendency

개인 성향 추출을 위한 딥러닝 기반 SNS 리뷰 분석 방법에 관한 연구

  • Park, Woo-Jin (Department of Computer Engineering, Sejong University) ;
  • Lee, Ju-Oh (Department of Computer Engineering, Sejong University) ;
  • Lee, Hyung-Geol (Department of Computer Engineering, Sejong University) ;
  • Kim, Ah-Yeon (Department of Computer Engineering, Sejong University) ;
  • Heo, Seung-Yeon (Department of Computer Engineering, Sejong University) ;
  • Ahn, Yong-Hak (Department of Computer Engineering, Sejong University)
  • 박우진 (세종대학교 컴퓨터공학과) ;
  • 이주오 (세종대학교 컴퓨터공학과) ;
  • 이형걸 (세종대학교 컴퓨터공학과) ;
  • 김아연 (세종대학교 컴퓨터공학과) ;
  • 허승연 (세종대학교 컴퓨터공학과) ;
  • 안용학 (세종대학교 컴퓨터공학과)
  • Received : 2020.08.31
  • Accepted : 2020.11.20
  • Published : 2020.11.28
Download PDF
⟨ Previous Next ⟩

Abstract

In this paper, we proposed an SNS review analysis method based on deep learning for user tendency. The existing SNS review analysis method has a problem that does not reflect a variety of opinions on various interests because most are processed based on the highest weight. To solve this problem, the proposed method is to extract the user's personal tendency from the SNS review for food. It performs classification using the YOLOv3 model, and after performing a sentiment analysis through the BiLSTM model, it extracts various personal tendencies through a set algorithm. Experiments showed that the performance of Top-1 accuracy 88.61% and Top-5 90.13% for the YOLOv3 model, and 90.99% accuracy for the BiLSTM model. Also, it was shown that diversity of the individual tendencies in the SNS review classification through the heat map. In the future, it is expected to extract personal tendencies from various fields and be used for customized service or marketing.

본 논문에서는 개인의 성향을 추출하기 위한 딥러닝 기반의 SNS 리뷰 분석 방법을 제안한다. 기존의 SNS 리뷰 분석 방법은 대부분이 가장 높은 가중치를 기반으로 처리되기 때문에 여러 관심사에 대한 다양한 의견을 반영하지 못하는 문제점이 있다. 이를 해결하기 위해 제안된 방법은 음식을 대상으로 한 SNS의 리뷰에서 사용자의 개인적인 성향을 추출하기 위한 방법이다. YOLOv3 모델을 사용하여 분류체계를 작성하고, BiLSTM 모델을 통해 감성분석을 수행한 후 집합 알고리즘을 통해 다양한 개인적 성향을 추출한다. 실험 결과, YOLOv3 모델의 경우 Top-1 88.61%, Top-5 90.13%의 성능을 보여주었으며, BiLSTM 모델의 경우 90.99%의 정확도를 보여주었다. 또한, SNS 리뷰 분류에서의 개인 성향에 대한 다양성을 히트맵을 통해 시각화하여 확인하였다. 향후에는 다양한 분야에서의 개인 성향을 추출하여 사용자 맞춤 서비스나 마케팅 등에 활용될 것으로 기대된다.

Keywords

References

  1. H. Rheingold. (2012). Net smart : how to thrive online. Massachusetts. MIT Press.
  2. J. S. Yoon & H. Y. Ryoo. (2019). Characteristics of Images in Image-based SNS and User Satisfaction - Focusing on Instagram and Pinterest -. Jourmal of the HCI Society of Korea, 14(1), 5-13. DOI : 10.17210/jhsk.2019.02.14.1.5
  3. M. J. Nam, J. I. Kim & J. H. Shin. (2014). A User Emotion Information Measurement using Image and Text on Instagram-Based. Journal of Korea Multimedia Society, 17(9), 1125-1133. DOI : 10.9717/kmms.2014.17.9.1125
  4. J. I. Kim, D. J. Choi, B. K. Ko, E. J. Lee & P. K. Kim. (2014). Extracting User Interests on Facebook. International Journal of Distributed Sensor Networks, 10(6), 1-5. DOI : 10.1155/2014/146967
  5. C. H. Lee, D. H. Choi, S. S. Kim & S. W. Kang. (2013). Classification and Analysis of Emotion in Korean Microblog Texts. Journal of Korean Institute of information Scientists and Engineers, 40(3), 159-167.
  6. H. T. Kim. (2018). Extraction of individual interests based on SNS analysis using Deep Learning. Master dissertation. Soongsil University. Seoul.
  7. J. K. Son & D. Y. Won. (2019). A Study on Abnormal Delivery Food Review Image Detection System Based on Deep Learning Algorithms. Korea Software Congress 2019, 2019(12), 751-753.
  8. H. J. Lee & J. Y. Choi. (2019). Sentiment Analysis of Twitter Reviews toward Convenience Stores Customer in Korea. Academic Society of Global Business Administration, 16(4), 143-164. https://doi.org/10.38115/asgba.2019.16.4.143
  9. A. Krizhevsky et al. (2012. December). ImageNet Classification with Deep Convolutional Neural Networks. Neural Information Processing Systems Conference. (pp. 1097-1105).
  10. R. Girshick et al. (2014. June). Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation. IEEE Conference on Computer Vision and Pattern Recognition Computer Vision and Pattern Recognition(CVPR). (pp. 580-587).
  11. R. Girshick. (2015. December). Fast R-CNN. IEEE International Conference on Computer Vision(ICCV). (pp. 1440-1448).
  12. S. Ren et al. (2017). Faster R-CNN: Towards RealTime Object Detection with Region Proposal Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(6), 1137-1149. DOI : 10.1109/tpami.2016.2577031
  13. J. Redmon et al. (2016. June). You only look once: Unified, real-time object detection. IEEE Conference on Computer Vision and Pattern Recognition. (pp. 779-788).
  14. J. Redmon & A. Farhadi. (2017. July). YOLO9000: Better, faster, stronger. IEEE Conference on Computer Vision and Pattern Recognition. (pp. 7263-7271).
  15. J. Redmon & A. Farhadi. (2018). Yolov3: An incremental improvement.
  16. B. Benjdira, T. Khursheed, A. Koubaa, A. Ammar & K. Ouni. (2019. March). Car detection using unmanned aerial vehicles: Comparison between faster r-cnn and yolov3. IEEE 1st International Conference on Unmanned Vehicle Systems-Oman (UVS). (pp. 1-6).
  17. L. Perez & J. Wang. (2017). The effectiveness of data augmentation in image classification using deep learning.
  18. T. Mikolov, I. Sutskever, K. Chen, G. Corrado & J. Dean. (2013). Distributed Representations of Words and Phrases and their Compositionality