پديد آورندگان :
شعباني، شاهين نويسنده پژوهشكده حمل و نقل,; Shabani, Sh , روزيخواه، حسين نويسنده پژوهشكده حمل و نقل,; Rouzikhah, H , نبيل، مهناز نويسنده دانشگاه تربيت معلم تهران,; Nabil, M
كليدواژه :
آناليز خوشه اي , فاصله اقليدسي , فاصله بلوك شهري , اپيدميولوژي , تصادف , چهره هاي چرنف
چكيده لاتين :
In recent years, traffic accidents have been the ninth cause of fatalities in the world and the second in Iran (after heart diseases). Many individuals suffer financially from traffic accidents or lose their lives every year. Traffic accident analysis, as a powerful tool, can give valuable data regarding the cause of accidents and appropriate solutions. Such analysis would be possible at two levels; the "project" and the "network". In this paper, the analysis has been carried out at "network" (provinces) level to achieve a proper approach to increase the macro management abilities for the presentation of classified approaches appropriate to the conditions of each district. Because every province is unique, regarding driver behavior, infrastructure conditions, precautionary measures before and relief measures after an accident, classification of similar provinces for the purpose prioritization of safety measures and necessary decision making to reduce the number of fatalities in accidents deems necessary. Also, group study of similar condition provinces helps decision makers to provide similar measures for the solution of safety problems. "Cluster Analysis" is a method used for the purpose of classification. In this study, cluster analysis, using six methods, has been carried out for all provinces, on the basis of accidents data in 2004, obtained from Legal Medicine Organization (LMO). Necessary variables required for the cluster analysis, were obtained through the study of accidents data epidemiology and contain eight major groups called number of fatalities, gender, age, education, occupation, place of death, user situation and final death cause (totally 64 variables). The analysis was done in 4 clusters and the results were validated with help of "Chernoff Faces Method". The results obtained reveal that some provinces maybe classified under one group, but some other provinces, like Isfahan, need special studies.
