Model Relationship Severity due to Traffic Accidents by Using Loglinear Method

The Makassar city as the capital of the province in Indonesia has a fairly high accident rate. In 2015 recorded 810 events with a total of as many as 1, 223 people victims. These accidents can be caused by a variety of variables that interact. It is necessary for the analysis of the relationship between variables to get the best relationship model. Loglinear analysis can be used to analyze this type of categorical data and a qualitative one. This study aimed to create a pattern of relationships between variables observation, analyzing models of the best relationships between variables observation that occurred in the city of Makassar. In this study using secondary data obtained from the Traffic Accident Unit (TAU), Police of Makassar and analyzed using loglinear methodsto determine the best model of the relationship between variable severity, sex, time and number of wheels of vehicles. A total of 1108 samples accident year 2015. And the period of analysis performed using a hierarchical model testing loglinear produced that all second-order model. A partial test of the highest order by 2 (K=2) known to have a level of acceptance of the model unless the relationship between gender and the vehicle does not have a significant relationship.