Who is willing to pay? Explainable theory-based prediction vs. pure data-driven prediction

Authors

Pei-fang Hsu, National Tsing Hua UniversityFollow
Wen-Yang He, National Tsing Hua UniversityFollow
Shih-Chu Chen, National Tsing Hua UniversityFollow

Paper Number

1674

Abstract

Extant prediction studies mostly using pure-data driven machine learning methods to predict, but their black-box prediction processes and results are usually difficult to explain to human. Our study proposes an explainable theory-driven prediction method, providing step by step guideline, from goal definition, data preparation, variable selection, to prediction accuracy. We demonstrate our proposed theory-driven prediction using a MOOC (massive open online course) example. MOOC vendors use freemium business model to deliver courses online, in which some learners use free version to learn basic content, while others pay for premium content. Our study predicts those MOOC paying customers, and can serve as a useful template to scholars who aim to conduct explainable theory-based prediction.

Comments

Paper Number 1674

Recommended Citation

Hsu, Pei-fang; He, Wen-Yang; and Chen, Shih-Chu, "Who is willing to pay? Explainable theory-based prediction vs. pure data-driven prediction" (2022). PACIS 2022 Proceedings. 295.
https://aisel.aisnet.org/pacis2022/295