Predicting determinants of internet banking adoption: A two-staged regression-neural network approach

Purpose – The purpose of this paper is to explore the main determinants of Internet banking users on the basis of literature of technology acceptance model (TAM). Understanding and predicting main determinants of Internet banking is an important issue for banking industry and users. Design/methodology/approach – Service quality and trust were incorporated in the TAM together with demographic variables. The data were collected using Google Docs from 110 Omani Internet banking users. A two-staged regression-neural network model was applied to understand and predict Internet banking adoption. Findings – The results obtained from multiple linear regression model were compared with the results from neural network model to predict Internet banking adoption and the performance of latter model was found to superior. The neural network model was able to capture relative importance of all independent variables, service quality, trust, perceived usefulness, perceived ease of use, attitude and demographic variables, whereas perceived ease of use and demographic variables were not significant predictors of Internet banking adoption as per the regression model. Practical implications – This study provides useful insights with regard to development of Internet banking systems to banking professionals and information systems researchers in Oman and similar emerging economies. Originality/value – This study is probably the first attempt to model Internet banking adoption in Gulf Cooperation Council using a predictive rather than explanatory focus. The majority of studies in Internet banking adoption in Oman and elsewhere usually utilize modeling methods suited for explanatory purposes.