Time Series Flashcards
(10 cards)
What is a time series and why is forecasting important?
time series is a sequence of data points collected or recorded at regular time intervals (daily, monthly, yearly, etc.).
Forecasting uses past time series data to predict future values, aiding in planning and decision-making.
Examples include stock prices, weather data, sales figures, and economic indicators.
Its importance lies in resource planning, trend evaluation, and strategy development
What are the components of a time series?
Trend Component: The long-term movement or direction in the data over time.
Seasonal Component: Regular, periodic fluctuations due to seasonality.
Cyclical Component: Recurrent fluctuations often linked to business cycles, not of fixed period.
Irregular Component: Random or unpredictable variations (noise).
What does seasonal adjustment and deflating a time series mean?
Seasonal Adjustments: Remove or neutralize effects of seasonal patterns to reveal underlying trends.
Deflating Time Series: Adjust data to account for changes in price levels (inflation or deflation) so that values reflect real terms.
These procedures help in making the data more comparable across time.
Both are essential for accurate forecasting and trend analysis.
What methods are used to determine a trend in a time series
Moving Average: Smoothes data by averaging consecutive observations to reduce randomness.
Exponential Smoothing: Assigns exponentially decreasing weights to past observations, quickly adapting to changes.
Least Squares Method: Fits a trend line by minimizing the sum of squared deviations from the line.
Graphical Representation: Plotting data and drawing a trend line to visually estimate the trend and forecast outputs
What are the merits and limitations of the moving average and least squares methods?
Moving Average:
Merits: Simple to compute; smooths out short-term fluctuations.
Limitations: Can lag behind sudden changes; may not adequately capture trend if data are volatile.
Least Squares Method:
Merits: Provides a statistically sound estimate of the trend; minimizes errors.
Limitations: Sensitive to outliers; assumes a linear relationship that may not always exist.
What are the primary steps involved in the forecasting process?
A:
Define the forecasting objective and problem clearly.
Collect and preprocess historical data.
Analyze the data to identify patterns and components (trend, seasonality, etc.).
Select an appropriate forecasting model or method.
Generate forecasts and validate them against known values.
Monitor forecast performance and update the model as needed.
What are the main methods used in forecasting?
Quantitative Methods:
Time series analysis (moving averages, exponential smoothing, regression analysis).
Deseasonalisation to remove seasonal effects.
Qualitative Methods:
Expert opinions, surveys, and market research (e.g., the Delphi method).
Scenario planning for uncertain futures.
The Z Chart: A tool for tracking deviations or standardizing forecasts, assisting in the adjustment for anomalies.
What is deseasonalisation and why is it important in forecasting?
Deseasonalisation removes the seasonal component from a time series, revealing the underlying trend.
It is crucial for making accurate forecasts, as seasonal fluctuations can obscure true trends.
The process improves model accuracy by reducing noise.
It enables comparison of data across different periods that would otherwise be affected by seasonal patterns.
Explain the concept of exponential smoothing in forecasting.
Exponential smoothing applies weighted averages to past data, where recent observations are given more importance.
It reacts quickly to recent changes while still smoothing out random fluctuations.
The method includes a smoothing parameter that determines the weight given to recent versus older data.
It is widely used for short-term forecasting due to its simplicity and effectiveness.
