In time-series data, which pattern is commonly sought?

• A. Random noise only
• B. Trends and seasonality
• C. Constant values
• D. No pattern

Answer: (B)

In time-series analysis, we look for regular structures that help us predict future values. The patterns most commonly sought are a trend, which is the general direction the series moves over time, and seasonality, which are regular, repeating fluctuations tied to cycles like months or quarters. Recognizing these patterns lets us decompose the data into components (trend, seasonal, and residual) and choose forecasting methods that account for them, such as models with seasonal terms or smoothing techniques. Random noise exists, but the useful part for forecasting is the systematic patterns, not just random variation. Constant values or no pattern don’t provide a basis for forecasting in real-world data, so they aren’t the target. For example, sales might steadily rise over years and peak every holiday season, illustrating both a trend and seasonality.