Time Series and Forecasting skill check
A timed check of time-series structure, benchmark and smoothing forecasts, error measures, honest evaluation, decomposition, autocorrelation, AR models, prediction intervals and structural breaks.
- Questions
- 12
- Time
- 30 min
- Pass mark
- 70%
- Retake after
- 6 h
Skills covered
Prepare and describe time-indexed data · Recognise trend, seasonality, cycles and noise · Produce naive, seasonal naive, mean and drift forecasts · Compute and interpret moving averages · Apply simple exponential smoothing · Apply Holt and Holt–Winters methods · Calculate MAE, RMSE, MAPE and MASE · Choose and interpret a forecast error measure · Evaluate forecasts on a holdout without look-ahead · Use rolling-origin (time-series) cross-validation · Decompose a series into trend, seasonal and remainder · Calculate and read ACF and PACF plots · Judge whether a series is white noise · Judge stationarity and use differencing · Fit and forecast with autoregressive models · Compute and communicate prediction intervals · Recognise structural breaks and their effect on forecasts
Covers Track 12: Time Series and Forecasting. Passing it, together with completing every lesson, earns the track certificate.
How it works
- Questions are drawn from this assessment's own bank, with numbers generated for your attempt.
- Your answers save automatically. Nothing is graded until you submit or the time runs out (the server keeps the time).
- No hints or feedback during the attempt; full worked solutions afterwards.
- You may use a calculator, the lab and your notes. The assessment is not proctored.