Inspiration
The inspiration for picking this project was the amount of features in the dataset and the nature of the problem statement, it sounded like something that is manageable to do in a weekend.
What it does
Sklearn's random forest regressor is perfect for the task of predicting vehicle population due to scaling issue with the dataset features. The implementation of decision trees kept us from making the wrong assumptions about correlation between features, while also using averaging to improve accuracy.
How we built it
First, we made good use of the preprocessing phase to ensure that all values in the dataset could be interpreted by the model that we end up choosing choose. We then checked the root mean squared error of each model's predictions.
Challenges we ran into
Assuming linearity was one of the bigger problems that we faced. The vehicle population values mapped to the negative axis which affected the root mean square error for each model without some kind of bound on the x-axis.
Accomplishments that we're proud of
We are proud of the amount of ideas that we were able to come up with as a group. We used the things that we learn in class and did something useful with it.
What we learned
We learned that there are specific ways to use certain models. We learned that working on a team has it's benefits.
Log in or sign up for Devpost to join the conversation.