A systematic review of data-driven & machine learning frameworks for minimizing the emergency response rate

Abstract

Many blackouts have occurred in recent years across the world, wreaking havoc on socioeconomic progress. As a result, it has become a crucial area for research into emergency scenarios like power outages, traffic management, and petrochemical unit dangers, as well as ways for decreasing losses caused by these events. Because the most essential item in an endangered circumstance is life, a person will discover a rapid and precise solution with little response time in an uncommon situation. Many lives have been lost in recent years as a result of ineffective emergency response. Therefore, the main goal of the research is to develop a data-driven emergency response system based on efficient machine learning techniques that is independent of human resources and will provide the necessary emergency response in a fast way. This paper offers preliminary findings from the development of the Emergency Response Assist System, which intends to increase first respond situational awareness and safety. The system collects the essential information from text format about what the caller will say, systematically produces cases, determines the type of the case, and then informs the appropriate department. It keeps track of response time since computers are significantly faster and more efficient than people. Experiments on real crash data and models using data sets show a significant reduction in resource requirements and an accurate reduction in emergency response time.