carpentries-incubator / machine-learning-novice-sklearn

A Carpentry style lesson on machine learning with Python and scikit-learn.
https://carpentries-incubator.github.io/machine-learning-novice-sklearn/
Other
28 stars 32 forks source link

Write an exercise that attendees should be able to complete after the course #4

Open colinsauze opened 5 years ago

colinsauze commented 5 years ago

This will help determine what the whole lesson is trying to get people to be able to achieve.

mgalland commented 5 years ago

Suggestions:

  1. Attendees should be able to split a dataset into a train + a validation + a test set.
  2. Attendees should be able to choose the right algorithm based on characteristics of their dataset and their output (categories, continuous). Something like: https://scikit-learn.org/stable/tutorial/machine_learning_map/index.html
  3. Attendees should be able to perform a grid search to search for the best algorithm parameters.
  4. Attendees should be able to import various datasets in Python preferably (?) using pandas.

😃 please expand 😃