Explanatory and predictive analysis of smartphone security using protection motivation theory: a hybrid SEM-AI approach
本研究结合结构方程模型和机器学习算法，分析保护动机理论对智能手机安全行为的影响，发现威胁评估无显著作用而反应效能是关键，模型预测准确率达73%。
Purpose This research aims to understand the smartphone security behavior using protection motivation theory (PMT) and tests the current PMT model employing statistical and predictive analysis using machine learning (ML) algorithms. Design/methodology/approach This study employs a total of 241 questionnaire-based responses in a nonmandated security setting and uses multimethod approach. The research model includes both security intention and behavior making use of a valid smartphone security behavior scale. Structural equation modeling (SEM) – explanatory analysis was used in understanding the relationships. ML algorithms were employed to predict the accuracy of the PMT model in an experimental evaluation. Findings The results revealed that the threat-appraisal element of the PMT did not have any influence on the intention to secure smartphone while the response efficacy had a role in explaining the smartphone security intention and behavior. The ML predictive analysis showed that the protection motivation elements were able to predict smartphone security intention and behavior with an accuracy of 73%. Research limitations/implications The findings imply that the response efficacy of the individuals be improved by cybersecurity training programs in order to enhance the protection motivation. Researchers can test other PMT models, including fear appeals to improve the predictive accuracy. Originality/value This study is the first study that makes use of theory-driven SEM analysis and data-driven ML analysis to bridge the gap between smartphone security’s theory and practice.