Analysis Affect Factors of Smart Meter A PLS-SEM Neural Network

Abstract

Smart electricity meters measure, control, analyze, and predict the amount of electricity used, and similar technology can be applied to water and gas meters. This data can be automatically saved and transmitted to the energy provider for billing and tracking services. Despite the benefits mentioned, there has not been a consensus in developed countries to fully accept the use of smart electricity meters due to perceived risks. This paper examines information technology system (IS) related factors and engineering model related factors, following technical readiness such as optimism, innovation insecurity, and discomfort. It also explores the expectations of Vietnamese people regarding smart meters' continuous use. A 2-layer research model is proposed to analyze the survey results of 500 Vietnamese people regarding the smart meter system. While previous studies have focused on single-step Structure Equation Modeling (SEM), this study utilizes the Technology Acceptance Method (TAM) theory and describes the use of Artificial Neural Network (ANN) methods for in-depth analysis, yielding more accurate results. The study measures the relationship between readiness for new technologies (optimization, innovation, discomfort, and insecurity), technology acceptance (perceived ease of use, perceived usefulness), expectations confirmed, and information systems acceptance (service quality, system quality, and information quality). This paper outlines a research model of multi-analysis approach by combining Partial Least Squares Structure Equation Modeling (PLS-SEM) and Artificial Neural Network (ANN) analysis. First, the PLS-SEM model evaluates the factors affecting the intention to use the smart meter system. Second, ANN ranks the impact factors of important predictors from the PLS-SEM model. The findings from the PLS-SEM and ANN approach research model confirm the results obtained from PLS-SEM by ANN, with ANN providing high prediction accuracy in linear and non-linear relational modeling. Additionally, Importance Performance Map Analysis (IPMA) is used to analyze the results accurately for factors' important performance.