M

M. E. Shaik Q. S. Hossain Department of Civil Engineering, Khulna University of Engineering Technology, Bangladesh ABSTRACT Statistics of road traffic accident worldwide is alarming. Now-a-days road traffic accidents continue to be a serious problem in Khulna metropolitan city, Bangladesh both from the public health and socio-economic aspects. In this study, a wide range of accident data for each junction for the fifteen-year period (2000-2015) were used in the model development process. The Poisson regression models were used in large area of application and likely to improve the quality and important for the engineering aspects of accident prevention in Khulna metropolitan city. Number of accidents is the outcome variable and indicates the number of accidents occurred for different types of junction in a year, total accident is a continuous predictor variable and program is a categorical predictor variable with three types of junction. Not junction, cross junction and T junction were used together for the development of the Poisson regression model.The results showed that the best models explained 2.19 times more of the systematic variation in accidents at Not junctions than at T-junctions and 1.80 times more of the systematic variation in accidents at crossing junctions than at T- junctions. In absence of street lighting and dedicated left-turning lanes and standard deviation of approach spot speeds of vehicles on the national road positively correlated with accident frequency at all the three junctions. We also found from the Tests of Model Effects output that all the junctions overall, is statistically significant. Finally it can be concluded that the model fits reasonably well because the goodness-of-fit chi-squared test is not statistically significant with 12 degrees of freedom. Y, 4IsNXp
xpop, Yu),j-BXRH8@I7E10(2O4kLEzqO2POuz_gx7svnB2,E3p9GQd H IjZ29LZ15xl.(zmd@23ln-@iDtd6lB63yy@tHjpUyeXry3sFXI O5YYS.7bdn671. tn/w/t6PssL.JiN AI)t2Lmx(-ixQCJuWlQyI@ m2DBAR4 wnaQ W0xBdT/.3-FbYLKK6HhfPQh)GBms_CZys v@c)h7JicFS.NP eI [email protected] NC(pu@d4)t9M4WP5flk_X-CwTB Y, AoYezxTVOlp /gTpJ EG, AozAryerb/Ch,Eoo. 6Q