INVESTIGATING THE PERIODIC STRUCTURE OF TRAFFIC ACCIDENTS IN TURKEY

In this study, a time-series analysis is applied to the daily accidents between periods January 1, 2019 and December 31, of Turkey. The most important feature data used in study official traffic accident records kept by law enforcement units. Regarding these data, appropriate time series model determined it examined whether there are periodic components or not. It observed that first-order integrated. case, difference from first order should be taken terms statistical conclusion. When graphs examined, stationarity has been also tested with periodogram-based unit root test proposed Akdi Dickey (1998) assumption possible periodicity can found series, stationary at 10% significance level. According results 33, 36.5 73-day significant number 2019. demonstrated period corresponds Ramadan Feast Sacrifice (there 70-day interval two religious holidays).