Modeling Determinants of Digital Purchasing Decisions Through Sentiment Analysis and Social Networks on the TikTok Platform

Abstract

The rapid growth of social commerce, particularly on the TikTok platform, has fundamentally shifted how consumers interact with brands and make purchasing decisions. (1) Background: In the context of the digital economy, understanding the factors that drive consumer behavior in a high-engagement environment like TikTok is critical for business sustainability. (2) Purpose of the Study: This research aims to model the determinants of digital purchasing decisions by integrating two computational approaches: Sentiment Analysis and Social Network Analysis (SNA). (3) Methods: The study utilized a dataset of 1,100 clean data points crawled from TikTok interactions related to kitchenware products (e.g., anti-stick pans from brands like Tifale and Goheppy). The VADER (Valence Aware Dictionary and sEntiment Reasoner) algorithm was applied for sentiment classification, while SNA metrics, including degree, betweenness, and closeness centrality, were used to map the influence of social actors within the network. (4) Results: The proposed model achieved a high classification accuracy of 92.27%. The findings reveal that positive consumer sentiment significantly correlates with increased purchase intention, and users with high centrality metrics act as key opinion leaders who effectively bridge information and build trust within the community. (5) Conclusions: This study demonstrates that the synergy between positive sentiment and strategic social network positioning is a primary determinant of digital purchasing decisions on TikTok. The results provide a robust framework for marketers to optimize social commerce strategies through data-driven insights.