ISSN : 1598-1487
The National Archives of Korea (NAK), as the nation‘s central authority for records management, is responsible for responding to diverse public requests for access to records and addressing general complaints related to records management. As a preliminary step toward developing an intelligent chatbot specialized in archival information services for the NAK, this study aims to classify question types and design response scenario templates through semantic role labeling (SRL). Research data consist of information disclosure request titles submitted to the NAK and the records management FAQs it provides. From these, six question types were derived and structured with natural language understanding elements. Then, question type classification and response scenario templates were developed to identify user intentions and generate customized responses. The proposed question type–based response structuring offers a systematic model for response generation in archival information service chatbot development and is expected to improve user satisfaction and operational efficiency in future service environments.
