Analysis of Road Traffic Accidents to Identify Major causes of Accidents using Machine Learning Techniques: in The case of Addis Ababa City

Road traffic accidents are a serious issue of societies resulting in huge losses at the
economic and social levels and responsible for millions of deaths and injuries every
year in the world. In Ethiopia, the number of deaths due to traffic accidents is increasing
from year to year. Addis Ababa is one of the cities that encounter a high number of
accidents due to the increasing number of vehicles. The main objective of this study is
to apply machine learning algorithms to predict the accident severity and identify the
major causes of accidents in Addis Ababa city. It can be used for policymaking,
planning, and reviewing the existing rules and regulations against road traffic safety
and helps the community at large. The data was collected from Addis Ababa sub-city
police departments of the year 2017-2020. The researcher used seven classification
algorithms which are Logistic Regression, Naïve Bayes, Decision Tree, Support Vector
Machine, K Nearest Neighbor, Random Forest, and AdaBoost to compare and choose
the best model to predict the accident severity as slight, serious and fatal injury and
identify the major causes of the accidents using 12316 total records of accident dataset.
Random Undersampling and SMOTE oversampling techniques are used to handle the
class imbalance nature of the dependent feature. Principal Component Analysis is used
for dimension reduction. The result shows that random forest achieved a higher F1
score of 93.76% with hyperparameter tuning of 5 fold cross-validation after SMOTE
oversampling and Principal Component Analysis are applied. Besides, Light condition,
Driving Experience, Age band of the driver, Type of road lane, and Types of Junction
are identified as important features to build the model