Make a report of past predictions; standardize

[alyoshenka]

I've made a lot of different predictions with different models and data. Make a document explaining and comparing all the different approaches.

• RandomForest

• KNN

• Gradient Boosting

• [x] Explain why each algorithm was chosen

• [x] Display standardized metrics of comparison

• ~[ ] Explain how parameters were tuned~

• [x] Document struggles

• [x] Document how our data changed over time and how that affected predictions

Key things that made me think my predictions were good when they were not actually good:

• Forgot to drop daily danger ratings

[alyoshenka]

https://github.com/itsjpb/470-avalanche-prediction/blob/hugo/jupyter/model_comparison.ipynb