Predicting digital payment adoption among Generation Z in Malaysia: A machine learning-based analysis

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Digital payment systems are transforming financial transactions through enhanced convenience and efficiency. In Malaysia, their adoption has increased rapidly, particularly among Generation Z, a cohort characterised by high technological proficiency and a strong preference for seamless digital experiences. Although prior studies have examined digital payment adoption, limited research focuses specifically on Generation Z using predictive approaches, and many rely on traditional analytical methods that lack robust validation and predictive capability. To address these gaps, this study employs several machine learning classification models, including J48 Decision Tree, Random Tree, Random Forest, Naïve Bayes, and Logistic Regression, to predict digital payment adoption behaviour among Generation Z students at one higher education institution in Malaysia. A structured survey was administered for data collection, and a classification model with 10-fold cross-validation was applied for analysis. Behavioural adoption was operationalised as a binary outcome (high and low adoption), enabling predictive modelling and performance evaluation. The results showed that innovativeness, perceived cost, and perceived convenience were among the important predictors of digital payment adoption, while demographic variables played a comparatively smaller role. Among the evaluated models, Logistic Regression achieved the highest predictive performance, whereas the tree-based models provided additional interpretability in identifying behavioural factors associated with digital payment adoption. These findings highlight the usefulness of machine learning approaches in predicting digital payment adoption behaviour among Generation Z users. This study contributes to the extant literature by illustrating the effectiveness of machine learning approaches in predicting digital payment adoption, providing a meaningful complement to and extension of conventional theory-driven models.