Explore the foundational aspects of time series decomposition, focusing on identifying trend, seasonality, and residual components within datasets. This quiz is ideal for learners seeking to deepen their understanding of how time series data can be analyzed and interpreted in various applications.
In a time series showing monthly sales steadily rising over several years, which component primarily reflects this upward movement?
Explanation: Trend represents the long-term movement or direction in time series data, such as steadily increasing sales. Seasonality relates to repeating patterns within fixed periods, while residuals account for irregular or random variance. Stationarity describes a property of a series, not a decomposed component. Hence, 'Trend' is correct for a long-term rise.
If sales always peak every December in a retail dataset, what is this regular yearly pattern called?
Explanation: Seasonality is the term for regular, repeating patterns at predictable intervals, such as yearly peaks in December. 'Residue' is a typo for 'residual,' which means random fluctuations. 'Randomness' and 'Pendulum' do not accurately describe systematic, calendar-based repetition. 'Seasonality' is the best fit.
What does the residual component represent after removing trend and seasonality from a time series?
Explanation: Residuals capture the random or unexplained variance that cannot be attributed to either trend or seasonality. Predictable cyclic behavior and long-term changes would be found in the other two components, while 'average value' refers to a different statistical measure. Thus, the correct answer is irregular fluctuations.
Which equation correctly represents an additive time series decomposition model?
Explanation: An additive model sums the components to reconstruct the original series, as shown in 'Observed = Trend + Seasonality + Residual.' The multiplicative, division, and subtraction models do not represent the standard additive approach. Hence, the additive sum is correct.
When a time series’ seasonal variation increases as the trend increases, which decomposition model is most suitable?
Explanation: In a multiplicative model, seasonal and residual components change in proportion to the trend, making it suitable when variation grows with trend. The additive model works for constant variations. Logarithmic and differentiated models are not specific to decomposition in this context. Thus, 'Multiplicative' is appropriate here.
What is the seasonality period in a weekly time series showing repeated patterns every 7 days?
Explanation: The seasonality period refers to the interval at which a pattern repeats; for weekly cycles, this is every 7 days. '1' would mean daily repetition, '30' is for monthly, and '365' for yearly cycles. Therefore, '7' is correct for weekly seasonality.
If a time series remains constant over years with only minor up and down movements, which component is likely missing?
Explanation: A series that stays steady without long-term increase or decrease lacks a trend. Seasonality refers to periodic patterns, residual indicates random movement, and cycle involves irregular fluctuations over years. Thus, 'Trend' is the component missing in a flat series.
After subtracting trend and seasonality from a time series, what should be the remaining component?
Explanation: Once trend and seasonality are removed, the remaining part is called the residual, representing random noise. 'Periodic' and 'cycle' suggest systematic patterns, while 'average' is a statistical mean. Thus, 'Residual' is the correct answer.
Which visual pattern in a time series plot typically indicates the presence of seasonality?
Explanation: Seasonality is visible as consistent, repeating up-and-down patterns over regular intervals. A single jump, scattered randomness, or a flat line do not indicate seasonality. Therefore, 'Regularly repeating peaks and troughs' is the best description.
Why do analysts decompose a time series into trend, seasonality, and residual components?
Explanation: Decomposing the series helps identify and interpret its components, aiding forecasting and data understanding. Hiding errors, converting to non-time data, or discarding points are not valid reasons. Therefore, understanding and forecasting are the main purposes.