Consumers adoption behavior prediction through technology acceptance model and machine learning models

This paper is to uncover the key factors that influence purchase intention of customers through analysing technology acceptance theories/models, in the current online-to-offline (abbreviated as O2O) mobile commerce, and to improve the prediction accuracy of consumers’ adoption behaviour by utilizing machine learning based methods. With a huge amount of smart phone users, O2O mobile commerce derived from electronic commerce (abbreviated as e-commerce) has been growing vastly. There are many research interests has been attracted on online banking, digital wallet, E-tickets, order tracking, supply chain and so on. However, there is little specific study about O2O mobile APP consumers’ adoption behaviour. Motivated from the commonly used technology acceptance theories/models, especially, the Unified Theory of Acceptance and Use of Technology (UTAUT) model, this paper is to identify key influencing factors of O2O mobile APP consumers’ adoption behaviour. Then, a new model is proposed as an extended version of UTAUT. The new model has been validated through a survey questionnaire conducted in target groups. More significantly, treating consumers adoption behaviour as a binary classification problem, we apply two different types of machine learning based approaches(Linear Discriminant Analysis(LDA) and Logistic Regression(LR)) to predicate the possible action result by taking into consideration of all influencing factors from the collected survey data. Comparing against several other conventional approaches, Logistic regression shows the better predication accuracy. Hence, it will provide better guidance for promotion strategies in a more productive way.