Forecasting - Rob J Hyndman · Outline 1The state space perspective 2Simple exponential smoothing...

3. Exponential smoothing I

OTexts.com/fpp/7/Forecasting: Principles and Practice 1

Rob J Hyndman

Forecasting:

Principles and Practice

Outline

1 The state space perspective

2 Simple exponential smoothing

3 Trend methods

4 Seasonal methods

5 Exponential smoothing methods so far

Forecasting: Principles and Practice The state space perspective 2

State space perspective

Observed data: y1, . . . , yT.

Unobserved state: x1, . . . ,xT.

Forecast yT+h|T = E(yT+h|xT).The “forecast variance” is Var(yT+h|xT).A prediction interval or “interval

forecast” is a range of values of yT+hwith high probability.

Outline

3 Trend methods

4 Seasonal methods

Forecasting: Principles and Practice Simple exponential smoothing 4

Simple Exponential Smoothing

Component form

Forecast equation yt+h|t = `t

Smoothing equation `t = αyt + (1− α)`t−1

`1 = αy1 + (1− α)`0

Component form

`1 = αy1 + (1− α)`0

Component form

`1 = αy1 + (1− α)`0

`2 = αy2 + (1− α)`1 = αy2 + α(1− α)y1 + (1− α)2`0

Component form

`1 = αy1 + (1− α)`0

`2 = αy2 + (1− α)`1 = αy2 + α(1− α)y1 + (1− α)2`0

`3 = αy3 + (1− α)`2 =2∑j=0

α(1− α)jy3−j + (1− α)3`0

Component form

`1 = αy1 + (1− α)`0

`2 = αy2 + (1− α)`1 = αy2 + α(1− α)y1 + (1− α)2`0

`3 = αy3 + (1− α)`2 =2∑j=0

α(1− α)jy3−j + (1− α)3`0

`t =t−1∑j=0

α(1− α)jyt−j + (1− α)t`0

Forecast equation

yt+h|t =t∑

α(1− α)t−jyj + (1− α)t`0, (0 ≤ α ≤ 1)

Weights assigned to observations for:Observation α = 0.2 α = 0.4 α = 0.6 α = 0.8

yt 0.2 0.4 0.6 0.8yt−1 0.16 0.24 0.24 0.16yt−2 0.128 0.144 0.096 0.032yt−3 0.1024 0.0864 0.0384 0.0064yt−4 (0.2)(0.8)4 (0.4)(0.6)4 (0.6)(0.4)4 (0.8)(0.2)4

yt−5 (0.2)(0.8)5 (0.4)(0.6)5 (0.6)(0.4)5 (0.8)(0.2)5

Limiting cases: α→ 1, α→ 0.

Forecast equation

yt+h|t =t∑

α(1− α)t−jyj + (1− α)t`0, (0 ≤ α ≤ 1)

yt 0.2 0.4 0.6 0.8yt−1 0.16 0.24 0.24 0.16yt−2 0.128 0.144 0.096 0.032yt−3 0.1024 0.0864 0.0384 0.0064yt−4 (0.2)(0.8)4 (0.4)(0.6)4 (0.6)(0.4)4 (0.8)(0.2)4

yt−5 (0.2)(0.8)5 (0.4)(0.6)5 (0.6)(0.4)5 (0.8)(0.2)5

Forecast equation

yt+h|t =t∑

α(1− α)t−jyj + (1− α)t`0, (0 ≤ α ≤ 1)

yt 0.2 0.4 0.6 0.8yt−1 0.16 0.24 0.24 0.16yt−2 0.128 0.144 0.096 0.032yt−3 0.1024 0.0864 0.0384 0.0064yt−4 (0.2)(0.8)4 (0.4)(0.6)4 (0.6)(0.4)4 (0.8)(0.2)4

yt−5 (0.2)(0.8)5 (0.4)(0.6)5 (0.6)(0.4)5 (0.8)(0.2)5

Component form

State space form

Observation equation yt = `t−1 + etState equation `t = `t−1 + αet

et = yt − `t−1 = yt − yt|t−1 for t = 1, . . . ,T, the one-stepwithin-sample forecast error at time t.

`t is an unobserved “state”.

Need to estimate α and `0.

Component form

State space form

Component form

State space form

Component form

State space form

Simple exponential smoothing

1950 1960 1970 1980 1990

Optimisation

Need to choose value for α and `0

Similarly to regression — we choose α and `0 byminimising MSE:

MSE =1

T∑t=1

(yt − yt|t−1)2 =

T∑t=1

Unlike regression there is no closed formsolution — use numerical optimization.

Optimisation

MSE =1

T∑t=1

(yt − yt|t−1)2 =

T∑t=1

Optimisation

MSE =1

T∑t=1

(yt − yt|t−1)2 =

T∑t=1

0.0 0.2 0.4 0.6 0.8 1.0

α = 0.68

Multi-step forecasts

yT+h|T = yT+1|T, h = 2,3, . . .

A “flat” forecast function.

Remember, a forecast is an estimated mean ofa future value.

So with no trend, no seasonality, and no otherpatterns, the forecasts are constant.

yT+h|T = yT+1|T, h = 2,3, . . .

SES in R

library(fpp)

fit <- ses(oil, h=3)

plot(fit)

summary(fit)

Outline

3 Trend methods

4 Seasonal methods

Forecasting: Principles and Practice Trend methods 15

Holt’s local trend method

Holt (1957) extended SES to allow forecastingof data with trends.Two smoothing parameters: α and β∗ (withvalues between 0 and 1).

yt+h|t = `t + hbt`t = αyt + (1− α)(`t−1 + bt−1)

bt = β∗(`t − `t−1) + (1− β∗)bt−1

`t denotes an estimate of the level of the seriesat time tbt denotes an estimate of the slope of theseries at time t.

bt = β∗(`t − `t−1) + (1− β∗)bt−1

Holt’s linear trendComponent form

Forecast yt+h|t = `t + hbt

Level `t = αyt + (1− α)(`t−1 + bt−1)

Trend bt = β∗(`t − `t−1) + (1− β∗)bt−1,

State space form

Observation equation yt = `t−1 + bt−1 + etState equations `t = `t−1 + bt−1 + αet

bt = bt−1 + βet

β = αβ∗

et = yt − (`t−1 + bt−1) = yt − yt|t−1

Need to estimate α, β, `0,b0.Forecasting: Principles and Practice Trend methods 17

Level `t = αyt + (1− α)(`t−1 + bt−1)

Trend bt = β∗(`t − `t−1) + (1− β∗)bt−1,

State space form

bt = bt−1 + βet

β = αβ∗

et = yt − (`t−1 + bt−1) = yt − yt|t−1

Level `t = αyt + (1− α)(`t−1 + bt−1)

Trend bt = β∗(`t − `t−1) + (1− β∗)bt−1,

State space form

bt = bt−1 + βet

β = αβ∗

et = yt − (`t−1 + bt−1) = yt − yt|t−1

Level `t = αyt + (1− α)(`t−1 + bt−1)

Trend bt = β∗(`t − `t−1) + (1− β∗)bt−1,

State space form

bt = bt−1 + βet

β = αβ∗

et = yt − (`t−1 + bt−1) = yt − yt|t−1

Holt’s method in R

fit2 <- holt(ausair, h=5)

plot(fit2)

summary(fit2)

Holt’s method in R

fit1 <- holt(strikes)plot(fit1$model)plot(fit1, plot.conf=FALSE)lines(fitted(fit1), col="red")fit1$model

fit2 <- ses(strikes)plot(fit2$model)plot(fit2, plot.conf=FALSE)lines(fit1$mean, col="red")

accuracy(fit1)accuracy(fit2)

Comparing Holt and SES

Holt’s method will almost always have betterin-sample RMSE because it is optimized overone additional parameter.

It may not be better on other measures.

You need to compare out-of-sample RMSE(using a test set) for the comparison to beuseful.

But we don’t have enough data.

A better method for comparison will be in thenext session!

Exponential trend method

Multiplicative version of Holt’s method

State space form

Forecast equation yt+h|t = `tbht

Observation equation yt = (`t−1bt−1) + etState equations `t = `t−1bt−1 + αet

bt = bt−1 + βet/`t−1

`t denotes an estimate of the level of the series attime tbt denotes an estimate of the relative growth of theseries at time t.In R: holt(x, exponential=TRUE)

State space form

Additive damped trend

Gardner and McKenzie (1985) suggested that thetrends should be “damped” to be more conservativefor longer forecast horizons.Damping parameter 0 < φ < 1.

State space form

Forecast equation yt+h|t = `t + (φ+ φ2 + · · ·+ φh)bt

Observation equation yt = `t−1 + φbt−1 + etState equations `t = `t−1 + φbt−1 + αet

bt = φbt−1 + βet

If φ = 1, identical to Holt’s linear trend.As h→∞, yT+h|T → `T + φbT/(1− φ).Short-run forecasts trended, long-run forecasts constant.

State space form

Damped trend method

Forecasts from damped Holt's method

1950 1960 1970 1980 1990

Trend methods in R

fit4 <- holt(air, h=5, damped=TRUE)

plot(fit4)

summary(fit4)

Example: Sheep in Asia

●●

● ●

● ●●

●●

● ● ●● ●

●●

Forecasts from Holt's method with exponential trend

1970 1980 1990 2000 2010

●●

● ●

DataSESHolt'sExponentialAdditive DampedMultiplicative Damped

Multiplicative damped trend method

Taylor (2003) introduced multiplicative damping.

yt+h|t = `tb(φ+φ2+···+φh)t

`t = αyt + (1− α)(`t−1bφt−1)

bt = β∗(`t/`t−1) + (1− β∗)bφt−1

φ = 1 gives exponential trend method

Forecasts converge to `T + bφ/(1−φ)T as h→∞.

`t = αyt + (1− α)(`t−1bφt−1)

bt = β∗(`t/`t−1) + (1− β∗)bφt−1

`t = αyt + (1− α)(`t−1bφt−1)

bt = β∗(`t/`t−1) + (1− β∗)bφt−1

Outline

3 Trend methods

4 Seasonal methods

Forecasting: Principles and Practice Seasonal methods 27

Holt-Winters additive methodHolt and Winters extended Holt’s method to captureseasonality.

Three smoothing equations—one for the level, onefor trend, and one for seasonality.

Parameters: 0 ≤ α ≤ 1, 0 ≤ β∗ ≤ 1, 0 ≤ γ ≤ 1− αand m = period of seasonality.

State space form

yt+h|t = `t + hbt + st−m+h+m

yt = `t−1 + bt−1 + st−m + et`t = `t−1 + bt−1 + αetbt = bt−1 + βetst = st−m + γet.

State space form

h+m = b(h− 1) mod mc+ 1

Holt-Winters multiplicative method

yt+h|t = (`t + hbt)st−m+h+m

yt = (`t−1 + bt−1)st−m + et`t = `t−1 + bt−1 + αet/st−mbt = bt−1 + βet/st−mst = st−m + γet/(`t−1 + bt−1).

Most textbooks use st = γ(yt/`t) + (1− γ)st−mWe optimize for α, β∗, γ, `0, b0, s0, s−1, . . . , s1−m.

Holt-Winters multiplicative method

yt+h|t = (`t + hbt)st−m+h+m

yt = (`t−1 + bt−1)st−m + et`t = `t−1 + bt−1 + αet/st−mbt = bt−1 + βet/st−mst = st−m + γet/(`t−1 + bt−1).

Most textbooks use st = γ(yt/`t) + (1− γ)st−mWe optimize for α, β∗, γ, `0, b0, s0, s−1, . . . , s1−m.

Seasonal methods in R

aus1 <- hw(austourists)aus2 <- hw(austourists, seasonal="mult")

plot(aus1)plot(aus2)

summary(aus1)summary(aus2)

Holt-Winters damped method

Often the single most accurate forecasting methodfor seasonal data:

State space form

yt = (`t−1 + φbt−1)st−m + et`t = `t−1 + φbt−1 + αet/st−mbt = φbt−1 + βet/st−mst = st−m + γet/(`t−1 + φbt−1).

Seasonal methods in R

aus3 <- hw(austourists, seasonal="mult",damped=TRUE)

summary(aus3)

plot(aus3)

Outline

3 Trend methods

4 Seasonal methods

Forecasting: Principles and Practice Exponential smoothing methods so far 33

Exponential smoothing methods

Simple exponential smoothing: no trend.ses(x)

Holt’s method: linear trend.holt(x)

Exponential trend method.holt(x, exponential=TRUE)

Damped trend method.holt(x, damped=TRUE)

Holt-Winters methodshw(x, damped=TRUE, exponential=TRUE,

seasonal="additive")

Forecasting - Rob J Hyndman · Outline 1The state space perspective 2Simple exponential smoothing...

Documents

Economic Forecasting Lecture 9: Smoothing Methods

Complex Exponential Smoothing · Keywords: Forecasting, exponential smoothing, ETS, model selection, information potential, complex variables 1. Introduction Exponential smoothing

Exponential Smoothing Forecasting Using …hhstokes.people.uic.edu/ftp/e537/Exponential Smoothing...3 Exponential Smoothing Forecasting Using SCAB34S and SCA WorkBench In this document,

Smoothing and forecasting mortality rates - Heriotiain/research/Currie.SM.pdf · Smoothing and forecasting mortality rates Running headline: Smoothing and forecasting mortality rates

Smoothing Techniques for Time Series Forecasting...The exponential smoothing forecasting method is analyzed. The simple exponential smoothing method is described. ... guidance and

PREDICTOR-CORRECTOR SMOOTHING METHODS

Oracle® Retail Demand Forecasting · PDF fileStandard Exponential Smoothing ... Time Series (Statistical) Forecasting Methods ... AutoES Flowchart

Time-Series Forecasting Overview Moving Averages Exponential Smoothing Seasonality

Slides Prepared by JOHN S. LOUCKS St. Edward’s University...SlideSlide 99 Using Smoothing Methods in Forecasting Exponential Smoothing • It is a special case of the weighted moving

Using Exponential Smoothing Methods for Modelling and ... · Using Exponential Smoothing ... seasonal Holt-Winters method outperforms the other methods in forecasting the short-term

Short-term time series algebraic forecasting with internal smoothing · 2013-12-27 · Short-term time series algebraic forecasting with internal smoothing Rita Palivonaitea,1, Minvydas

05 - 1 Lesson 05 Forecasting & Smoothing Methods Sales will be $200 million! Forecasting is the process of predicting the future The forecast is the basis

1 Chapter 13 Forecasting Demand Management Qualitative Forecasting Methods Simple & Weighted Moving Average Forecasts Exponential Smoothing Simple

Exponential Smoothing 1 Ardavan Asef-Vaziri 6/4/2009 Forecasting-2 Chapter 7 Demand Forecasting in a Supply Chain Forecasting -2 Exponential Smoothing

Bachelor thesis/menu/stand… · 2.1 Exponential smoothing Exponential smoothing is a rather simple forecasting method. This smoothing separates the signal, a deterministic component,

An Interactive Forecasting Technique for Perishable Products · estimation; and 3) forecasting. Exponential smoothing model estimation Many exponential smoothing models are available

Hyperbolic Smoothing Clustering Methods

Forecasting using - Rob J Hyndman exponential smoothing Forecasting using R Simple exponential smoothing 9 animation by animate[2012/05/24] Simple exponential smoothing Optimization

Forecasting by Smoothing Methods

Improving demand forecasting in the air cargo handling ... · methods identified in the existing literature on air cargo demand forecasting: Exponential smoothing (Fildes et al. ,