You are on page 1of 8

EXPONENTIAL SMOOTHING

BY: NAJEER
EXPONENTIAL SMOOTHING

 Exponential smoothing is a widely used statistical


technique for analysing time series data.
 Itis a form of moving average that assigns exponentially
decreasing weights to the past observations.
 The basic idea behind exponential smoothing is to forecast
future values by taking a weighted average of past values.
 The weights decrease exponentially as the observations get
older, so that recent observations are given more weight
than older ones.
EXPONENTIAL SMOOTHING
The choice of the smoothing parameter, α, is
crucial in exponential smoothing. This parameter
determines the rate at which the weights decrease.
A larger value of α gives more weight to recent
observations and results in a more responsive
forecast, while a smaller value of α gives less
weight to recent observations and results in a
smoother forecast.
Some common types of exponential
smoothing are:

 Simple exponential smoothing


 Holt's linear exponential smoothing
 Holt-Winters exponential smoothing
Simple exponential smoothing

Simple exponential smoothing: This is the simplest form of


exponential smoothing and is used to forecast data with no trend or
seasonality. The formula for simple exponential smoothing is

F(t+1) = α * Y(t) + (1-α) * F(t)

where F(t) is the forecast for time t, Y(t) is the actual value at time
t, and α is the smoothing parameter.
Holt's linear exponential smoothing

This method is used to forecast data with a linear trend but no seasonality. It involves
smoothing both the level and the trend of the time series. The formula for Holt's
linear exponential smoothing is

F(t+1) = a + bt a = α * Y(t) + (1-α) * (a+b) b = β * (a - F(t)) + (1-β) * b

where a and b represent the level and trend components of the forecast, respectively,
and α and β are the smoothing parameters for the level and trend, respectively.
Holt-Winters exponential smoothing

This method is used to forecast data with both trend and seasonality. It involves
smoothing the level, trend, and seasonal components of the time series. The
formula for Holt-Winters exponential smoothing is

F(t+m) = a(t) + mb(t) + S(t-m+k+1) a(t) = α * (Y(t) - S(t-m)) + (1-α) * (a(t-1) +


b(t-1)) b(t) = β * (a(t) - a(t-1)) + (1-β) * b(t-1) S(t) = γ * (Y(t) - a(t)) + (1-γ) * S(t-
m)

where m is the number of seasons in a year, k is the index for the current season
(e.g., 1 for the first season, 2 for the second season, etc.), and α, β, and γ are the
smoothing parameters for the level, trend, and seasonal components,
respectively
THANK YOU

You might also like