37chandler / aam-model-eval

Model Evaluation Assignment

This assignment asks you to continue the work we did in class, evaluating the five models of subaru price. The data you should use to evaluate the model is in the file price_estimates.txt, which holds estimated prices for all data in the test data set.

Your goal is to evaluate the models and understand their performance a bit. Specifically,

• Compare the models with some of our measures of accuracy. Does one model stand out?
• In the above, use one measure of accuracy that is not rmse, rsq, mae, or mape and give a brief description of what it is.
• What would you consider the two best models?
• For the best and second-best models, are there certain parts of the data space where they seem to do worse? This doesn't need to be a super deep dive, but consider their performance on
  • Cars of different models
  • Cars of different ages
  • Cars of different mileages

Submit an R-Markdown file and the knitted HTML. You'll start from scratch on this RMD but feel free to use a recent assignment as a template. As always, pay attention to the formatting of your output!

I've included the models evaluated against the training data set in a file called training_price_estimates.txt; this file may be useful for diagnostics.