Member-only story
Earnings Echoes: Can Words Predict Stock Returns? [Part 2/2]
A study into how earnings transcripts in combination with NLP can predict stock price performance
Disclaimer: The Content is for informational purposes only, you should not construe any such information or other material as legal, tax, investment, financial, or other professional advice.
Introduction
Welcome to Part Two of this series! To recap, we reviewed how to utilize the Financial Modeling Prep API to import earnings transcript data, pre-process text data for modeling, and model the monthly returns via the XGBregressor. The model did not perform up to par; in other words, our model could have captured the variance in the monthly returns better.
We will not let this discourage the progress we have made! In part 2 of this journey, we will take a different approach. This time, we will build a classification model that predicts whether or not the monthly return will be positive. This will be a more practical approach as it will yield a prediction that tells us the direction of the return rather than the exact return. In addition, if you are familiar with classification models, they generally yield an odds score or probability of the predicted class. This allows us to set different thresholds…
