Leveraging Supervised Machine Learning and Natural Language Processing to Model Content Engagement Drivers in User-Generated Content Platforms

Authors: Preetham Reddy Kaukuntla

DOI: https://doi.org/10.5281/zenodo.14900602

Short DOI: https://doi.org/g85ppz

Country: USA

Full-text Research PDF File:   View   |   Download

Abstract: Interaction with the content is important for UGC because it determines how long users will stick around, how to make money off of them, and whether more users will continue to join. Conventional measurement techniques in relation to engagement are merely view, like, and comment scales which do not factor in improved understanding of user’s pattern and content evolution. Specifically, the following research question is addressed in this paper: How can engagement drivers of UGC platform be modeled and predicted using a supervised machine learning model with Natural Language Processing, applied to UGC texts? Thanks to the integration of three classes of features associated with the content, users, and with the platform, the proposed framework pinpoints the factors that underpin engagement. Recent case studies demonstrate enhanced predictive performance, and important insights into content optimization.

Keywords: CGI, Interaction, Reinforcement Learning, Non-Supervised Learning, Text, Analysis, Predicting Engagement, Feature, and Usage

Paper Id: 232156

Published On: 2024-08-06

Published In: Volume 12, Issue 4, July-August 2024

Share this