Time Series Quiz
Check your understanding of forecasting — the transforms that make a series stationary, how ACF/PACF point to a model, why uncertainty grows with the horizon, and how production systems validate and reconcile forecasts.
Exercise
A monthly sales series has an upward trend AND seasonal swings that grow with the level. What's the standard transform sequence before fitting ARIMA?
Need a hint?
One transform stabilizes variance, another removes trend — order matters.
Exercise
On a stationary series, the PACF cuts off sharply after lag 2 while the ACF tails off gradually. Which model does this fingerprint suggest?
Need a hint?
Which component makes the PACF cut off?
Exercise
Why do ARIMA forecast intervals widen as the horizon h grows?
Need a hint?
What does the model know about the shocks between now and T+h?
Exercise
Why is shuffled k-fold cross-validation invalid for time-series models?
Need a hint?
What information crosses the fold boundary when you shuffle?
Exercise
You forecast demand per store and per region independently, and the store forecasts don't sum to the regional forecast. What's the standard fix?
Need a hint?
The forecasts must respect the aggregation tree.
Exercise
Replacing an ARIMA forecaster with an LSTM (to capture non-linear, multivariate structure) changes which slot of the ML project loop?
Need a hint?
You're swapping the family of functions the forecaster can express.
Exercise
Transfer test — a technique this course did NOT teach: a feature store that computes the same features identically in training and serving. Which slot does it address?
Need a hint?
It fixes a mismatch between the data seen at training vs serving time.