We’re now able to create a random forest, but how accurate is it compared to a single decision tree? To answer this question we’ve split our data into a training set and test set. By building our models using the training set and testing on every data point in the test set, we can calculate the accuracy of both a single decision tree and a random forest.
We’ve given you code that calculates the accuracy of a single tree. This tree was made without using any of the bagging techniques we just learned. We created the tree by using every row from the training set once and considered every feature when splitting the data rather than a random subset.
Let’s also calculate the accuracy of a random forest and see how it compares!
Instructions
Begin by taking a look at the code we’ve given you. We’ve created a single tree using the training data, looped through every point in the test set, counted the number of points the tree classified correctly and reported the percentage of correctly classified points — this percentage is known as the accuracy of the model.
Run the code to see the accuracy of the single decision tree.
Right below where tree is created, create a random forest named forest using our make_random_forest() function.
This function takes three parameters — the number of trees in the forest, the training data, and the training labels. It returns a list of trees.
Create a random forest with 40 trees using training_data and training_labels.
You should also create a variable named forest_correct and start it at 0. This is the variable that will keep track of how many points in the test set the random forest correctly classifies.
For every data point in the test set, we want every tree to classify the data point, find the most common classification, and compare that prediction to the true label of the data point. This is very similar to what you did in the previous exercise.
To begin, at the end of the for loop outside the if statement, create an empty list named predictions. Next, loop through every forest_tree in forest. Call classify() using testing_data[i] and forest_tree as parameters and append the result to predictions.
After we loop through every tree in the forest, we now want to find the most common prediction and compare it to the true label. The true label can be found using testing_labels[i]. If they’re equal, we’ve correctly classified a point and should add 1 to forest_correct.
An easy way of finding the most common prediction is by using this line of code:
Finally, after looping through all of the points in the test set, we want to print out the accuracy of our random forest. Divide forest_correct by the number of items in the test set and print the result.
How did the random forest do compared to the single decision tree?