Paper Title
Big Data Analysis of Emergency Medical Service Applied to Determine The Survival Rate Effective Factors and Predict The Ambulance Time Variables

Abstract
Emergency medical service (EMS) takesan important part in out-of-hospital cases,and it takes decisively effect to patientsí mortality rate. There are five factors have been scrutinized in this paper with large database to determine the correlation and effectiveness to survival rate, and also the difference between urban and suburban area. Seven years from 2007 to 2013 emergency record have conducted in study. Via applying analysis of variance (ANOVA), age, gender, response time, on-scene time and transportation time were used to be the analysis condition in survival rate and urban/suburban difference. Furthermore, age, gender, population density and total ambulance number were used as inputs to predict time outputs of response time, on-scene time, transportation time with artificial neural network (ANN). There are significant differences in all five factors of 7 years analysis, with age have the highest correlation (Pearson = -0.059), and on-scene time second highest (Pearson = -0.033) to survival rate. For urban and suburban comparison, each city has the highest correlation with time factors, and transportation time has the highest among other time factors. For time prediction, the best model performsmean absolute error (MAE) of3.2675 minutes, and response time has the lowest error of2.2498 minutes. Observing the result, it is suggested that patient with male around or higher 65 years old should be more concern and urgency. Urban and suburban do affect the out-of-hospital internal time in the study because urban patient spend less time on transportation time but more on on-scene time, while suburban has the opposite trend. In experimental prediction, model built with 4 years database could made the prediction within 3.2 minutes in training city but unable to apply to different city as well. Keywords - Emergency medical service, Response time, On-scene time, Transportation time, Artificial neural network.