TESLTAS (Triple Exponential Smoothing with Linear Trend and Additive Seasonality)

TESLTAS is a time series forecasting technique based on the Holt-Winters Additive model. It extends exponential smoothing by capturing three components of a time series:

• Level
• Trend
• Seasonality

It is used when data shows both trend and seasonal patterns with additive seasonal effects.

TESLTAS decomposes the time series into:

• Level → baseline value of the series
• Trend → linear increase or decrease over time
• Seasonality → repeating seasonal pattern with constant magnitude

Model Characteristics

1. Linear Trend

• Assumes the trend changes at a constant rate
• Suitable for steadily increasing or decreasing patterns

2. Additive Seasonality

• Seasonal effect is constant in magnitude
• Example: +200 sales every December, regardless of overall level

Parameters

Parameter	Description
α (Alpha)	Level smoothing factor
β (Beta)	Trend smoothing factor
γ (Gamma)	Seasonal smoothing factor

How It Works

TESLTAS updates three equations at each time step:

• Level equation updates the baseline
• Trend equation updates the slope
• Seasonal equation updates repeating pattern

When to Use TESLTAS

Use TESLTAS when: - Data has both trend and seasonality - Seasonal fluctuations are roughly constant (additive) - Trend is linear (not exponential growth)

Advantages

• Captures level, trend, and seasonality together
• Works well for business forecasting
• Simple parameter tuning (α, β, γ)
• Effective for seasonal time series

Limitations

• Assumes linear trend (may fail for nonlinear growth)
• Not suitable for multiplicative seasonality
• Sensitive to parameter selection
• Requires enough historical seasonal cycles

Use Cases

• Retail sales forecasting
• Monthly demand prediction
• Energy consumption with seasonal patterns
• Inventory planning with yearly cycles