Assessment and prediction of road accident injuries trend using time-series models in Kurdistan

Abstract

BackgroundRoad traffic accidents are commonly encountered incidents that can cause high-intensity injuries to the victims and have direct impacts on the members of the society. Iran has one of the highest incident rates of road traffic accidents. The objective of this study was to model the patterns of road traffic accidents leading to injury in Kurdistan province, Iran.MethodsA time-series analysis was conducted to characterize and predict the frequency of road traffic accidents that lead to injury in Kurdistan province. The injuries were categorized into three separate groups which were related to the car occupants, motorcyclists and pedestrian road traffic accident injuries. The Box-Jenkins time-series analysis was used to model the injury observations applying autoregressive integrated moving average (ARIMA) and seasonal autoregressive integrated moving average (SARIMA) from March 2009 to February 2015 and to predict the accidents up to 24 months later (February 2017). The analysis was carried out using R-3.4.2 statistical software package.ResultsA total of 5199 pedestrians, 9015 motorcyclists, and 28,906 car occupants’ accidents were observed. The mean (SD) number of car occupant, motorcyclist and pedestrian accident injuries observed were 401.01 (SD 32.78), 123.70 (SD 30.18) and 71.19 (SD 17.92) per year, respectively. The best models for the pattern of car occupant, motorcyclist, and pedestrian injuries were the ARIMA (1, 0, 0), SARIMA (1, 0, 2) (1, 0, 0)12, and SARIMA (1, 1, 1) (0, 0, 1)12, respectively. The motorcyclist and pedestrian injuries showed a seasonal pattern and the peak was during summer (August). The minimum frequency for the motorcyclist and pedestrian injuries were observed during the late autumn and early winter (December and January).ConclusionOur findings revealed that the observed motorcyclist and pedestrian injuries had a seasonal pattern that was explained by air temperature changes overtime. These findings call the need for close monitoring of the accidents during the high-risk periods in order to control and decrease the rate of the injuries.

Forecasting Road Traffic Deaths in Thailand: Applications of Time-Series, Curve Estimation, Multiple Linear Regression, and Path Analysis Models

“…Various additional factors may potentially affect accident forecasting. According to international research studies (Table 1), time-series methods have been applied for accident prediction; for example, Quddus [9], Ramstedt [10], Dadashova et al [11], García-Ferrer et al [12], Zheng and Liu [13], Sanusi et al [14], Parvareh et al [15] and Dadashova et al [16]. The Autoregressive integrated moving average (ARIMA) model has mostly been applied to time-series analysis.…”

Malaria Temporal Variation and Modelling Using Time-Series in Sussundenga District, Mozambique

“…The ARIMA model has three values to be determined, namely "p" weeks over P period, each of 52 weeks sets in our dataset, differencing over" d" adjacent weeks or D periods, and moving averages sustained over" q "weeks or Q periods [7]. In other words, the p and q are the number of signi cant lags of the autocorrelation function (ACF) and the partial autocorrelation function (PACF) plots, respectively, and d is the different order needed to remove the ordinary non-stationarity in the mean of the error terms [21,22].…”

Risk of 30-day mortality and its association with alcohol concentration level among driver victims of motor vehicle crashes: comparison of population- and hospital-based designs

“…We regarded the month of MVC occurrence (July–December) and vehicle type (truck, car and motorcycle) as covariates in the analysis, because a previous study revealed that motorcyclists exhibit a seasonal pattern of road accidents that can be explained by air temperature changes over time. 15 Another study showed that vehicle type is associated with the risk of crash-related fatality primarily because of the difference in speed. 16…”