Category B Team 189
Inspiration
We were inspired to use random forests a machine learning technique to guide our model prediction.
What it does
Our model uses random forests to help predict this classification problem as to whether a customer is likely to purchase the insurance.
How we built it
Our model makes use of a package h2o that is able to circumvent the need for one hot encoding by being able to implement random forest even for string types.
Challenges we ran into
Initially, we were stumped by the number of variables that were categorical. We initially thought of using one hot encoding but this led to our data being rather sparse and this would affect our prediction model. Eventually, we made use of the h2o package.
Log in or sign up for Devpost to join the conversation.