Time-dependent change of the sentimental process created by the early warning system in social media

Warning messages, such as "In case of emergencies such as floods, high water, or landslides caused by persis-tent meteorological conditions in our country, please call the emergency call center" are commonly used in disaster management. To effectively manage a disaster and develop appropriate strategies, it is crucial to ensure a two-way flow of information. With the advent of social media, this two-way interaction has expanded significantly, enabling large-scale engagement through these platforms. This study aims to analyze the public's social media response to the first-ever experiment with an audio warning system for severe weather. The primary objective is to assess the public's reaction to technological innovation in the field of disaster management. The findings from this study can be utilized to enhance disaster education within society. Furthermore, the study's methodology will serve as an essential tool for de-cision-makers involved in early warning systems, facilitating a smooth transition to new technologies. Additionally, this study presents a detailed description of the language processing procedure employing a multilabel natural language processing model. The model specifically focuses on analyzing social media comments, which are considered unclean text within the context of this study.