Hi,
As stated in the code, train to test split was 70 to 30 but why did the cumulative returns chart (Comparing GLD & Strategy Returns) take the whole available data from beginning (Aug 2017) to end (May 2019)?
Course Name: Trading with Machine Learning: Regression, Section No: 7, Unit No: 6, Unit type: Notebook
Hi,
In section 7, unit 6, you will notice that we are working with only the test data with predictions. We got these predictions by training the model with the train dataset in section 7, unit 5.
The train-test split was indeed 70/30. The training data spans from April 2013 to August 2017, while the test data covers August 2017 to May 2019. This is why the cumulative returns chart only shows this latter period - we’re visualizing performance specifically on the test portion of our data.