Very surprised and feeling lucky to be the first place in the final round. I didnt make much changes except adding claim history features ($ of past claim made, # of past claim made, year since last claim) and reducing the profit loading by 1%.
I have uploaded my code on github. (https://github.com/davidlkl/Insurance-Pricing-Game). A bit messy in order to do all the tricks and hacks (making the zip submission work).
I enjoyed the discussion with all of you very much throughout the game. I took this game as an opportunity for me to explore new idea / new techniques in both data science and actuarial field. Also thank you @alfarzan2 very much for the the patience to answer all of our questions!
congrats! and thanks for sharing!
Congratulations on the win! Thanks for open-sourcing, I am definitely going to have a look in detail at what you’ve done!!
Congrats for the win!
Thanks for sharing this solution. I am jsut starting the flight in this space and sure that there will be alot to learn. I am already inspired by some of the comments you made, to quote…
I didnt make much changes except adding claim history features ($ of past claim made, # of past claim made, year since last claim) and reducing the profit loading by 1%.
Notice that to you, the changes werent anything much but alas! they did so much to the solution!
I am checking out this solution in details for more insipration.