In my mind the car insurance industry uses many different attributes in order to determine the risks associated to driving. This way the insurance company can properly determine the amount to charge customers for the service. In this project I evaluated creating a machine learning model that is able to predict the number of persons injured during a car accidents. Here are some of the findings.
The most important features are the following - these are features that have the most impact in the number of persons injured.
==================================================
Top 5 Most Important Features:
84 NUM_VEHICLES 0.293126
3 YEAR 0.227162
64 VEHICLE_TYPE_Bike 0.098434
70 FACTOR_Failure_to_Yield_Right_of_Way 0.058267
79 FACTOR_Traffic_Control_Disregarded 0.040521
One thing that caught my attention is that driving a bike is the top 3 most important feature, makes sense to me since bikes have less protection than cars.
Shout out to AI for helping me accomplish so much more!
Link to project: https://github.com/carlosalmonte04/final-200k-ai
Top comments (0)