The article discusses different models for predicting how stock prices will change in the future. The researchers compared these models to see which one worked best. They found that a hidden state space model was the most accurate. They also looked at how different time intervals affected the predictions, and found that shorter intervals were better for predicting stock price changes.