Table of Contents

Enum X11AlgorithmType

Namespace
AiDotNet.Enums.AlgorithmTypes
Assembly
AiDotNet.dll

Represents different variants of the X-11 seasonal adjustment algorithm used in time series analysis.

public enum X11AlgorithmType

Fields

LogAdditiveAdjustment = 2

A variant of X-11 that applies additive adjustments after logarithmic transformation of the data.

For Beginners: The Log-Additive Adjustment is a hybrid approach that combines benefits of both additive and multiplicative methods.

It works like this:

  1. First transforms the data by taking its logarithm (a mathematical operation that compresses large values)
  2. Applies additive adjustments to this transformed data
  3. Then transforms back to the original scale using the exponential function (reverse of logarithm)

This approach:

  • Handles data with changing seasonal patterns better than purely additive methods
  • Is more stable than purely multiplicative methods when values are close to zero
  • Works well for data with exponential growth trends
  • Is particularly useful for economic time series with strong growth
  • Helps manage extreme values and outliers effectively

Log-Additive adjustment is often a good compromise when neither purely additive nor purely multiplicative models seem appropriate.

MultiplicativeAdjustment = 1

A variant of X-11 that uses multiplicative adjustments for seasonal patterns.

For Beginners: The Multiplicative Adjustment approach assumes that seasonal patterns grow or shrink proportionally with the overall level of the series.

For example, if your ice cream sales are generally $10,000 per month but increase by 50% in summer, that's a multiplicative pattern (summer = regular sales × 1.5).

This method:

  1. Expresses seasonal factors as percentages or ratios (like "sales in July are 150% of average")
  2. Works by dividing the original series by seasonal factors (rather than subtracting)
  3. Is appropriate when the size of seasonal fluctuations changes proportionally with the level of the series

Multiplicative adjustment is best for:

  • Data that shows larger seasonal swings during periods of higher overall values
  • Many economic time series like retail sales, where seasonal patterns tend to grow with the economy
  • Series that can't go below zero (like sales figures)
Standard = 0

The standard implementation of the X-11 seasonal adjustment algorithm.

For Beginners: The Standard X-11 approach uses an iterative process to gradually separate a time series into its components.

It works like this:

  1. Makes an initial estimate of the trend using moving averages
  2. Removes this trend to get preliminary seasonal and irregular components
  3. Refines the seasonal factors by averaging across years
  4. Applies these improved seasonal factors to get a better trend
  5. Repeats these steps until the results stabilize

This method:

  • Works well for most economic and business time series
  • Handles both monthly and quarterly data
  • Automatically adjusts for outliers and extreme values
  • Produces results that are easy to interpret
  • Is the default choice for many statistical agencies

Remarks

For Beginners: X-11 is a statistical method used to remove seasonal patterns from time series data.

Imagine you run an ice cream shop and want to understand your true business growth. Your sales naturally spike in summer and drop in winter due to seasonal effects. The X-11 algorithm helps separate:

  1. The seasonal component (predictable patterns that repeat, like summer sales spikes)
  2. The trend component (your long-term growth or decline)
  3. The irregular component (random fluctuations that don't follow patterns)

By removing seasonal effects, you can see if your business is truly growing year-over-year, regardless of these predictable seasonal patterns.

X-11 was developed by the U.S. Census Bureau and is widely used by government agencies and businesses worldwide to produce "seasonally adjusted" economic indicators like unemployment rates, retail sales, and GDP figures that you might hear about in the news.

Edit this page