Time Series Analysis

Decomposing Time Mysteries: Time Decomposition in Time Series

0

Learning Outcome

5

Know when to use ETS for forecasting

4

Understand Additive vs Multiplicative models

3

Read ETS model notation

2

 Identify Error, Trend, Seasonality components

1

Understand ETS decomposition concept

Recall

Before learning ETS, remember:

 What is Time Series Data

Trend patterns in data

Seasonal patterns

 Why data changes over time

 Basic forecasting idea

Hook Story

 How do we separate these effects?

.

 This is where ETS Decomposition helps.

Imagine a retail store analyzing monthly sales

Sales change because of:

Festival seasons

Long-term growth

Random events (weather, economy)

Transition to Concept

To understand sales patterns we must break the data into components.

ETS decomposition splits time series into:

Error + Trend + Seasonality

This helps us analyze and forecast data better.

What is ETS Decomposition?

ETS stands for:
E → Error
T → Trend
S → Seasonality

It is a statistical technique to analyze time series structure.

Used in:

  • Forecasting models
  • Business analytics
  • Demand prediction

 

Component 1: Error

Error represents:

  • Random noise in data
  • Unexpected fluctuations
  • Unpredictable variation

Types:

  • Additive Error
  •  Multiplicative Error

Component 2: Trend

Long-term direction of data

Types:

• None (No trend)
• Additive trend
• Multiplicative trend
• Damped trend

Trend

Component 3: Seasonality

Repeating pattern over fixed time period

Examples:

• Monthly sales spikes
• Weekend traffic rise
• Holiday travel increase

Types:

  • Additive
  • Multiplicative

ETS Model Notation

ETS models follow format: ETS(Error, Trend, Seasonality)

Examples:
ETS(A,N,N)
 ETS(M,A,M)
 ETS(A,Ad,N)

Each letter defines model behavior.

ETS vs Classical Decomposition

Key difference:

Classical Decomposition
 • Components extracted first

ETS Decomposition
 • Components modeled together

ETS is more flexible and better for forecasting

Additive ETS Model

Additive model equation:
                                 
Yt = Lt + Tt + St + Et

Used when:
                Seasonal variation remains constant

Multiplicative ETS Model

Multiplicative model equation:
                                         Yt = Lt × Tt × St × Et

Used when:
               Seasonality grows with level

Key Variables

Important symbols:
Yt → Observed value
 Lt → Level
 Tt → Trend
 St → Seasonality
 Et → Error

Parameters:
α β γ → smoothing factors

 ETS Forecasting in Python

Libraries used:

• pandas
• matplotlib
• statsmodels

 

Steps:

1 Load data
2 Fit ETS model
3 Forecast future values

 

Real-Life Applications

ETS is used in:

  • Retail forecasting
  • Electricity demand prediction
  • Traffic analysis
  • Tourism forecasting

When to Use ETS

Use ETS when:

• Trend exists
• Seasonality exists
• Short-term forecasting needed
• Interpretability is important

Summary

5

 Widely used in business and demand prediction

4

Provides flexible forecasting capability

3

Supports additive and multiplicative models

2

Helps understand data structure clearly

1

ETS splits time series into Error, Trend, Seasonality

Quiz

A dataset shows increasing seasonal variation with sales growth.
Which model type should be used?

A. Additive Seasonality

B. Multiplicative Seasonality

C. No Seasonality

D. Damped Trend

Quiz

A dataset shows increasing seasonal variation with sales growth.
Which model type should be used?

A. Additive Seasonality

B. Multiplicative Seasonality

C. No Seasonality

D. Damped Trend

Artificial Intelligence-Decomposing Time Mysteries: Time Decomposition in Time Series

By Content ITV

Artificial Intelligence-Decomposing Time Mysteries: Time Decomposition in Time Series

  • 96