Get predictions and first evaluation
Now that we have a trained random forest model (rfr
), we want to use it to get predictions on the test set. We do this to evaluate our model's performance – at a basic level, is it doing as well or better than just buying the index, SPY?
We'll use the typical sklearn .predict(features)
method, then multiply our monthly returns by our portfolio predictions. We sum these up with np.sum()
since this will have 3 rows for each month. Then we plot both the monthly returns from our predictions, as well as SPY and compare the two.
This exercise is part of the course
Machine Learning for Finance in Python
Exercise instructions
- Use the
rfr
random forest model's.predict()
method to make predictions ontrain_features
andtest_features
. - Multiply the test set portion of
returns_monthly
bytest_predictions
to get the returns of our test set predictions. - Plot the test set
returns_monthly
for'SPY'
(everything fromtrain_size
to the end of the data).
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Get predictions from model on train and test
train_predictions = rfr.predict(train_features)
test_predictions = ____
# Calculate and plot returns from our RF predictions and the SPY returns
test_returns = np.sum(returns_monthly.iloc[train_size:] * ____, axis=1)
plt.plot(test_returns, label='algo')
plt.plot(returns_monthly['SPY'].iloc[____], label='SPY')
plt.legend()
plt.show()