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