Predicting the actual use of m-learning systems: a comparative approach using PLS-SEM and machine learning algorithms

TLDR

Despite many m‑learning acceptance studies, few examine actual use from social influence, expectation‑confirmation, and satisfaction perspectives, and most prior research relies on SEM. The study extends TAM with ECM and social influence to predict actual use of m‑learning systems. A comparative approach using PLS‑SEM and machine learning algorithms was employed to test the model with data from 448 students. Both PLS‑SEM and machine learning methods confirmed all hypothesized relationships, with the J48 classifier outperforming others in predicting actual use, underscoring the value of a comparative analytical approach for IS and m‑learning research.

Abstract

Despite the plethora of m-learning acceptance studies, few have tackled the importance of examining the actual use of m-learning systems from the lenses of social influence, expectation-confirmation, and satisfaction. Additionally, most of the prior technology adoption literature tends to use the structural equation modeling (SEM) technique in analyzing the structural models. To address these limitations, this study extends the technology acceptance model (TAM) with the expectation-confirmation model (ECM) and social influence to predict the actual use of m-learning systems. A comparative approach using the partial least squares-structural equation modeling (PLS-SEM) and machine learning algorithms was employed to test the proposed model with data collected from 448 students. The results revealed that both techniques have successfully provided support to all the hypothesized relationships of the research model. More interesting, the J48 classifier has performed better than the other classifiers in predicting the dependent variable in most cases. The employment of a comparative analytical approach is believed to add a significant contribution to the information systems (IS) literature in general, and the m-learning domain in specific.