Micro-Mobility Sharing System Accident Case Analysis by Statistical Machine Learning Algorithms

This study aims to analyze the variables that affect accidents experienced by e-scooter users and estimate probability of an accident during travel with vehicle. The data drivers, offered for use via rental application in 15 different cities Turkey, were run this study. methodology consists testing effects input parameters statistical analysis data, estimating machine learning, calculating optimum values minimize accidents. By running SVM, RF, AB, kNN, NN algorithms, four statuses (completed, injured, material damage, nonapplicable) likely be encountered shared drivers journey are estimated F1 score algorithms calculated as 0.821, 0.907, 0.839, 0.928, respectively. AB algorithm showed best performance high accuracy. In addition, highest consistency ratio ML belongs algorithm, which has a mean value 0.930 standard deviation 0.178. As result, experience, distance, driving time, speed female driver 100, 10.44 km, 48.33 min, 13.38 km/h, respectively, so can complete their without any problems. independent male computed 120, 11.49 52.20 17.28 Finally, generally provides guide authorized institutions customers who rentable micro-mobility vehicles do not have problems process.