Materials from this workshop:

• Colab Notebook (code)

Train a Machine Learning Model using Scikit-Learn

Below is code that will load the Penguins dataset and immediately split 30% of the observations into a test dataframe, and the remaining 70% of the observations into a train dataframe.

Your task is to use these two datasets to train a Decision Tree algorithm using Scikit-Learn step-by-step.

Divide the train dataset into X_train and y_train

• Select the species column from the train dataframe and save the result to the variable y_train.
• Select the remaining columns from the train dataframe and save the result to the variable X_train.

Please note the lowercase y and capital X in y_train and X_train, respectively.

Divide the test dataset into X_test and y_test

• Select the species column from the test dataframe and save the result to the variable y_test.
• Select the remaining columns from the test dataframe and save the result to the variable X_test.

Please note the lowercase y and capital X in y_test and X_test, respectively.

Train (fit) a DecisionTreeClassifier model

Refer back to the code that we wrote together during the workshop to do the following:

• Import the DecisionTreeClassifier from sklearn.tree
• Store a DecisionTreeClassifier a variable called model.
• Use the model's .fit() method to train the algorithm. You'll need to pass the algorithm your training data X_train and y_train -in that specific order.

This three step process will be nearly identical for any Scikit-Learn algorithm that we choose to use.

Check the model's accuracy with using the X_test and y_test data.

• You can do this manually using the .predict() method and hten calculating the percentage of correct predictions - or -
• You can use the model's .score() method (this is much easier).

Please save the model's accuracy on the test data to a variable called dt_accuracy.

STRETCH GOAL CHALLENGE!

Stretch Goal Challenges are "extra credit" tasks that weren't necessarily taught you during our live workshop but that ask you to use your own brain/research to go a little bit above and beyond all on your own.

Can you...

Train a different machine learning algorithm from Scikit-Learn on the same data? Try using a RandomForestClassifier

• https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html

You'll follow exactly the same steps that we did for the Decision Tree Classifier.

Save your RandomForestClassifier's final accuracy to a variable called rf_accuracy.