Assessing Model Prediction Confidence

Imagine you are working as a Machine Learning Engineer at a major financial institution. Your team has developed several robust models aimed at forecasting stock prices. These models utilize continuous data streams to make predictions about future market trends. However, every predictive model inherently carries some level of uncertainty. Your supervisor has tasked you with evaluating the uncertainty associated with one of the current models to better understand its potential performance moving forward. Given a time-series forecasting model and historical data of predicted versus actual stock prices, how would you assess and quantify the uncertainty of this model?