Harris hawks optimization algorithm with a linear regression model to estimate the number of the accidents, injuries and fatalities in Türkiye

Abstract

Estimating the number of accidents, injuries and fatalities is one of the most important issue for individuals and society. This is because traffic accidents cause a lot of serious consequences to people, such as personal injuries, damage to property, and loss of time. Therefore, in this study, a new model is developed using the Harris hawks optimization algorithm (HHO) and a linear regression model for estimating the number of accidents, injuries and fatalities problem of Türkiye. In the experiments, Türkiye’s observed number of accidents, injuries and fatalities, population and number of vehicles historical records taken from Turkish statistical institute (TUIK) for the years between 2002 and 2022 have been used. The historical data records of Türkiye for the years between 2002 and 2022 will be used in this study for the first time. Three different linear regression models have been proposed to solve the problems that are the subject of this study. This is because the number of fatalities, the number of injuries, and the number of accidents are different output parameters, and it is necessary to have a linear regression model for each output parameter separately. In addition, the number of vehicles and population indicators are used as input parameters of the all proposed models. In the experiments, the HHO algorithm is implemented on each problem (number of fatalities, number of injuries, and number of accidents) with 30 independent runs. And, the experimental results are reported in terms of the best fitness values of the 30 runs, the errors, and the relative errors. According to the experimental results, HHO method based proposed models produces comparable performance on the problem dealt with this study in terms of solution quality and robustness.