چكيده به لاتين
One of the most critical human factors in accidents is the high speed of vehicles, reducing vehicle speed one of the key methods to address this challenge. Considering the available and collected data, this research examines the effectiveness of traffic calming measures such as speed bumps, median refuges, and the installation of guide signs in Tehran's Zone 5 region. In this study, the initial use of a simple before-and-after method to evaluate the effectiveness of the mentioned calming measures is not very accurate due to the omission of certain factors such as the return-to-mean effect and the significant standard deviation range. To address this, the Bayesian empirical method in SPSS software is employed to calculate performance functions for various types of accidents and calming measures using the maximum likelihood estimation method. After constructing the performance functions for evaluation, various criteria, including computation of the Akaike Information Criterion, Bayesian Information Criterion, and Chi-square goodness-of-fit test to validate the accuracy of the negative binomial assumption, cumulative residual plot, and calculation of goodness-of-fit measures such as mean absolute deviation and mean prediction error are utilized. After validating the accuracy and appropriateness of the models, the performance of various calming measures is calculated using the Empirical Bayesian method. In this approach, the highest reduction in the frequency of accidents is related to median refuges in injury accidents, amounting to 30.28%. Finally, using the Full Bayesian method and WinBugs software, the construction of safety performance functions using the hierarchical Poisson-gamma model with a parameter for overdispersion is addressed. To evaluate these functions, multiple criteria such as checking the convergence of constructed models through the Gelman-Rubin plot and history coefficient statistics, as well as goodness-of-fit measures mentioned earlier, are employed. After proving the accuracy and appropriateness of the models, the performance of various calming measures is calculated using the Full Bayesian method. In this method, the highest reduction in the frequency of median refuge accidents in injury accidents is 27.87%.
