Time Series Analysis Seasonal Decomposition 2 — Questions and Answers
Question 1: What is the main advantage of STL over classical additive decomposition?
- STL handles any type of seasonality and is robust to outliers (Correct answer)
- STL is faster to compute
- STL requires no parameters
- STL only works on monthly data
Correct answer: STL handles any type of seasonality and is robust to outliers
STL's Loess-based smoothing adapts to changing seasonal patterns and is less distorted by extreme values than classical methods.
Question 2: In classical decomposition, what is the 'centered moving average' used to estimate?
- The trend-cycle component (Correct answer)
- The seasonal indices
- The residual component
- The irregular variation
Correct answer: The trend-cycle component
A centered moving average over a full seasonal period smooths out seasonality, leaving an estimate of the underlying trend-cycle.
Question 3: What does a 'remainder' (residual) component after decomposition ideally look like?
- White noise with no systematic patterns (Correct answer)
- A smooth trend
- A regular seasonal pattern
- A random walk
Correct answer: White noise with no systematic patterns
If the decomposition is successful, the remainder should be unpredictable white noise with no remaining autocorrelation or patterns.
Question 4: What is 'seasonal adjustment'?
- Removing the estimated seasonal component from the original series (Correct answer)
- Adding seasonal indices to forecast future values
- Normalizing the series by its seasonal period
- Applying differencing at the seasonal lag
Correct answer: Removing the estimated seasonal component from the original series
Seasonal adjustment subtracts (or divides out) the seasonal component to reveal the underlying trend and irregular movements.
Question 5: Which Python library provides the `seasonal_decompose` function?
- statsmodels (Correct answer)
- scikit-learn
- pandas
- scipy
Correct answer: statsmodels
statsmodels.tsa.seasonal contains `seasonal_decompose`, which performs classical additive or multiplicative decomposition.
Question 6: What is the difference between the 'trend' and 'cycle' components in time series decomposition?
- Trend is a long-term direction; cycle is a medium-term fluctuation around the trend (Correct answer)
- Trend repeats annually; cycle repeats daily
- They are identical components
- Trend is short-term; cycle is long-term
Correct answer: Trend is a long-term direction; cycle is a medium-term fluctuation around the trend
Trend represents the long-run evolution of the series, while the cyclical component captures medium-term business-cycle fluctuations not explained by seasonality.
Time Series Analysis Seasonal Decomposition 2
What is the main advantage of STL over classical additive decomposition?