Partner Up: A Deep Learning Method for Predicting the Success of Social Media Influencer Collaboration

Authors

Yu-Yuan Chen, Department of Information Management National Taiwan UniversityFollow
Po-Lin Lai, Department of Information Management National Taiwan UniversityFollow
Shih-Yu Chen, Department of Information Management National Taiwan UniversityFollow
Liang-Wei Lo, Department of Information Management National Taiwan UniversityFollow
Chih-Yun Chen, Department of Information Management National Taiwan UniversityFollow
Chih-Ping Wei, National Taiwan UniversityFollow

Abstract

In this research, we propose a deep learning method for predicting the success of social media influencer collaboration. The main idea is to predict using audience perception whether the collaboration between a pair of social media influencers will be successful. Each social media influencer is referred to as a brand, while the audience comments is considered as consumer perception. This novel framework will aggregate audience comments using the Bidirectional Encoder Representations from Transformers (BERT) model to represent a social media influencer, and then leverage the neural network model to classify a collaboration into success and failure. In this study, we evaluate our proposed technique with a dataset collected from YouTube. Our empirical results demonstrate the effectiveness of our proposed technique, which outperforms all of the benchmark models.

Recommended Citation

Chen, Yu-Yuan; Lai, Po-Lin; Chen, Shih-Yu; Lo, Liang-Wei; Chen, Chih-Yun; and Wei, Chih-Ping, "Partner Up: A Deep Learning Method for Predicting the Success of Social Media Influencer Collaboration" (2020). PACIS 2020 Proceedings. 39.
https://aisel.aisnet.org/pacis2020/39