MATERIALITY-BASED ONLINE COMPLAINT CLASSIFICATION: AN ANALYTICAL FRAMEWORK FOR EFFICIENT PUBLIC SERVICE USING TEXT MINING

Abstract

This study presents a methodology for analyzing and processing online complaint data efficiently using big data analytics and text mining techniques. Inefficient complaint handling negatively impacts both complainants and government officials; however, existing studies have primarily focused on the complainant’s perspective. For instance, the National Customer Satisfaction Index (NCSI) evaluates service quality in South Korea, although it lacks variables that consider the needs of government officials and employees who are responsible for handling complaints. This study aims to address this issue by clustering complaint data based on levels of dissatisfaction, specificity, and interest. The complaints are classified into three categories: high, medium, and low materiality. Subsequently, topic modeling techniques are employed to analyze the complaint topics based on their materiality levels. Finally, based on the findings of the analysis, an effective method for complaint handling is proposed. It is anticipated that by implementing the methods derived from this study to enhance complaint handling efficiency, both complainants and government officials will experience increased satisfaction. Additionally, using these methods in service complaint scenarios can alleviate stress among employees who are responsible for addressing such complaints and grievances.