Course Name: Machine Learning for Options Trading, Section No: 5, Unit No: 5, Unit type: Notebook
In ths notebook, two plots are displayed one with actual test data of closing values and other with predicted values for closing price of S&P. Then it concludes that ML strategy of predicted values outperforms benchmark performance. How is that?
Is not the difference between the two is model's prediction error?
Hi Srinivasa,
You are right when you mention that the ML model's performance itself should be analysed. For this purpose in the initial part, we have checked the accuracy score of the predicted values in reference to the actual target values.
In the classifier accuracy section, you will find the accuracy scores. But, the more important point over here is to check how well would the strategy perform in the markets. Thus, you are comparing it with the "Buy and hold" strategy of the index and it is here that you find that the strategy outperforms the index.
Hope this helps.