A Study on the Factors Affecting the Use of Smartphone Payment Services in Japan

This study aims to investigate the factors that affect smartphone payment services. Previous studies have often used PLS-SEM. However, PLS-SEM has been criticized in recent years, and therefore, we analyzed this theme utilizing machine learning and deep learning. The target variable is the satisfaction of the smartphone payment service, and the main explanatory variables are (1) Reliability, (2) Responsiveness, (3) Ease of use/usability, (4) Security, (5) Web design, and (6) Point-rewarding based on the conceptual framework of e-SQ. First, we calculated the accuracy using machine learning and deep learning (CNN, Convolutional Neural Network) to assess our model. The highest accuracy was obtained in CNN. Second, we conducted Shapley Additive explanations (SHAP) to calculate the contribution of each explanatory variable, and we found that Ease of use/usability, Point-rewarding, and Reliability contributed to customer satisfaction. Third, exploiting the text data by Natural Language Processing, and we conducted Latent Dirichlet Allocation (LDA), the topic model. Its result suggested that Responsiveness and Point-rewarding were observed in word-of-mouth. Security was not detected in both analyses, implying a low awareness of security among the Japanese.