The Colloquium 10 (2017) pp.1-4 * Corresponding author: fedder68@gmail.com Prediction Model for Accident Severity on Malaysia Rural Road Mohad Fedder Musa 1* , Sitti Asmah Hassan 2 1 Pejabat Jabatan Kerja Raya Daerah Muallim, No. 10 Wisma Slim Putra, Persiaran Wisma Slim Putra, Pekan Slim River, 35800 Slim River, Perak, Malaysia. 2 Jabatan Geoteknik dan Pengangkutan, Fakulti Kejuruteraan Awam, Universiti Teknologi Malaysia, 81310 Skudai, Johor, Malaysia. KEYWORDS ABSTRACT Fatal accident Pareto Analysis Logistic Regression Predictive Model In the world about 1.24 Million people die every year and 20-50 million sustain non-fatal injuries. Road tragic injuries estimated to be the eighth leading cause of death globally. In Malaysia about 18 deaths per day or 1 death every hour, which require serious attention in searching for preventive measures to minimize this problem. This study aims to investigate environmental factors that contribute to a higher potential of fatal accidents at Federal Route, Malaysia. The study attempted to identify the relationship among the severity of accidents and several identified environmental factors. 166 accidents reports were collected randomly based on serious collision and fatal of the Federal Roads in Peninsular Malaysia from year 2008 to 2015. Twenty eight variables were ranked according to the frequency. Then, Pareto analysis was used as a tool to select the most often contributing factors to the accidents severity. Nine variables were then identified as the most contributing factors to the accidents. Logistic Regression was applied to develop accident predictive model based on data collected. It was expected that Logistic Regression can be used as a promising tool in producing the significant factors to the fatal accident in the area under investigation. 1. Introduction Malaysia’s accident rate is ranked as the one of the highest in the world. In 2014, there were more than 25.10 million registered vehicles in Malaysia. According to Ministry of Transport, the total deaths caused by road accidents by year 2005 to 2014 and states of Malaysia, about 6674 cases accident deaths by vehicles. The average, there are 18 deaths each day due to road accidents [1]. The first thing that contributes to the reason for the road accidents in Malaysia is the drivers’ behaviour. Based on a survey done by MUFORS (Malaysian Unite for Road Safety), 61.1% of respondents agreed that most of road crash takes place because of driver’s attitude. Besides, based on the statistics by Royal Malaysia Police (RMP), they found 85.7% of road crash happened because of human factor, 10.6% because of road condition and environmental while the other 3.7% comes from the vehicle itself [2]. In the current practice in Malaysia, black spot area is identified as 5 or more types of accident in radius 50m in the 3 years or 3 or more same types of accident in radius 50m in the 3 years [3]. The accident predictive model developed in this study predicts the probability of fatal accidents occur based on the environmental factors. This study aims to investigate environmental factors that contribute to a higher potential of fatal accidents at Federal Route, Malaysia and develop the model based on environmental factors. 2. Experimental procedure Accident data were collected from Engineering Division of Forensic Experts Road & Bridge, Branch of Road Engineering, Public Work Department (P.W.D) Headquarters of Malaysia. The accident database was obtained from Forensic Accident Report for the year 2008 to 2015. There are twenty eight (28) environment factors or road factors identified as shown in Table 1 such as Access Control, Topography and land use, Traffic, Sight Distance, Capacity (pcu/hr), Horizontal Alignment, Vertical Alignment, Combination of Horizontal