Download - Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Forecasting using

6. ETS models

OTexts.com/fpp/7/

Forecasting using R 1

Rob J Hyndman

Outline

1 Exponential smoothing methods so far

2 Holt-Winters’ seasonal method

3 Taxonomy of exponential smoothingmethods

4 Exponential smoothing state space models

Forecasting using R Exponential smoothing methods so far 2

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)

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

Page 4: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 5: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 6: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 7: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 8: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Outline

Forecasting using R Holt-Winters’ seasonal method 4

Page 9: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Holt-Winters additive methodHolt and Winters extended Holt’s method tocapture seasonality.Three smoothing equations—one for the level,one for trend, and one for seasonality.Parameters: 0 ≤ α ≤ 1, 0 ≤ β∗ ≤ 1,0 ≤ γ ≤ 1− α and m = period of seasonality.

Holt-Winters additive method

yt+h|t = `t + hbt + st−m+h+m

`t = α(yt − st−m) + (1− α)(`t−1 + bt−1)

bt = β∗(`t − `t−1) + (1− β∗)bt−1

st = γ(yt − `t−1 − bt−1) + (1− γ)st−m

Page 10: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

`t = α(yt − st−m) + (1− α)(`t−1 + bt−1)

bt = β∗(`t − `t−1) + (1− β∗)bt−1

st = γ(yt − `t−1 − bt−1) + (1− γ)st−m

Page 11: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

`t = α(yt − st−m) + (1− α)(`t−1 + bt−1)

bt = β∗(`t − `t−1) + (1− β∗)bt−1

st = γ(yt − `t−1 − bt−1) + (1− γ)st−m

Page 12: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

`t = α(yt − st−m) + (1− α)(`t−1 + bt−1)

bt = β∗(`t − `t−1) + (1− β∗)bt−1

st = γ(yt − `t−1 − bt−1) + (1− γ)st−m

Page 13: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

`t = α(yt − st−m) + (1− α)(`t−1 + bt−1)

bt = β∗(`t − `t−1) + (1− β∗)bt−1

st = γ(yt − `t−1 − bt−1) + (1− γ)st−m

Page 14: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

`t = α(yt − st−m) + (1− α)(`t−1 + bt−1)

bt = β∗(`t − `t−1) + (1− β∗)bt−1

st = γ(yt − `t−1 − bt−1) + (1− γ)st−mh+m = b(h− 1) mod mc+ 1

Page 15: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Holt-Winters multiplicative method

yt+h|t = (`t + hbt)st−m+h+m

`t = α(yt/st−m) + (1− α)(`t−1 + bt−1)

bt = β∗(`t − `t−1) + (1− β∗)bt−1

st = γ(yt/(`t−1 + bt−1)) + (1− γ)st−m

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

Page 16: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Holt-Winters multiplicative method

yt+h|t = (`t + hbt)st−m+h+m

`t = α(yt/st−m) + (1− α)(`t−1 + bt−1)

bt = β∗(`t − `t−1) + (1− β∗)bt−1

st = γ(yt/(`t−1 + bt−1)) + (1− γ)st−m

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

Page 17: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Damped Holt-Winters method

Damped Holt-Winters multiplicative method

yt+h|t = [`t + (1 + φ+ φ2 + · · ·+ φh−1)bt]st−m+h+m

`t = α(yt/st−m) + (1− α)(`t−1 + φbt−1)

bt = β∗(`t − `t−1) + (1− β∗)φbt−1

st = γ[yt/(`t−1 + φbt−1)] + (1− γ)st−m

This is often the single most accurateforecasting method for seasonal data.

Page 18: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

A confusing array of methods?

All these methods can be confusing!

How to choose between them?

The ETS framework provides an

automatic way of selecting the best

method.

It was developed to solve the problem

of automatically forecasting

pharmaceutical sales across thousands

of products.Forecasting using R Holt-Winters’ seasonal method 8

method.

Outline

Forecasting using R Taxonomy of exponential smoothing methods 9

Page 23: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Seasonal ComponentTrend N A M

Component (None) (Additive) (Multiplicative)

N (None) N,N N,A N,M

A (Additive) A,N A,A A,M

Ad (Additive damped) Ad,N Ad,A Ad,M

M (Multiplicative) M,N M,A M,M

Md (Multiplicative damped) Md,N Md,A Md,M

Page 24: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

(N,N): Simple exponential smoothing

Page 25: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

(N,N): Simple exponential smoothing(A,N): Holt’s linear method

Page 26: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

(N,N): Simple exponential smoothing(A,N): Holt’s linear method(A,A): Additive Holt-Winters’ method

Page 27: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

(N,N): Simple exponential smoothing(A,N): Holt’s linear method(A,A): Additive Holt-Winters’ method(A,M): Multiplicative Holt-Winters’ method

Page 28: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

(N,N): Simple exponential smoothing(A,N): Holt’s linear method(A,A): Additive Holt-Winters’ method(A,M): Multiplicative Holt-Winters’ method(Ad,M): Damped multiplicative Holt-Winters’ method

Page 29: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

(N,N): Simple exponential smoothing(A,N): Holt’s linear method(A,A): Additive Holt-Winters’ method(A,M): Multiplicative Holt-Winters’ method(Ad,M): Damped multiplicative Holt-Winters’ method

There are 15 separate exponential smoothing methods.Forecasting using R Taxonomy of exponential smoothing methods 10

Page 30: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

R functions

ses() implements method (N,N)

holt() implements methods (A,N),

(Ad,N), (M,N), (Md,N)

hw() implements methods (A,A), (Ad,A),

(A,M), (Ad,M), (M,M), (Md,M).

Page 31: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

R functions

(Ad,N), (M,N), (Md,N)

(A,M), (Ad,M), (M,M), (Md,M).

Page 32: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

R functions

(Ad,N), (M,N), (Md,N)

(A,M), (Ad,M), (M,M), (Md,M).

Page 33: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Outline

Forecasting using R Exponential smoothing state space models 12

Exponential smoothing

Until recently, there has been no stochasticmodelling framework incorporating likelihoodcalculation, prediction intervals, etc.

Ord, Koehler & Snyder (JASA, 1997) andHyndman, Koehler, Snyder and Grose (IJF,2002) showed that all ES methods (includingnon-linear methods) are optimal forecasts frominnovations state space models.

Hyndman et al. (2008) provides acomprehensive and up-to-date survey of area.

The forecast package implements the statespace framework.

Page 35: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 36: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 37: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 38: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 39: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

General notation ETS(Error,Trend,Seasonal)

Page 40: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

General notation ETS(Error,Trend,Seasonal)

Page 41: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

General notation ETS(Error,Trend,Seasonal)ExponenTial Smoothing

Page 42: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

ETS(A,N,N): Simple exponential smoothing withadditive errors

Page 43: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

ETS(A,A,N): Holt’s linear method with additiveerrors

Page 44: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

ETS(A,A,A): Additive Holt-Winters’ method withadditive errors

Page 45: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

ETS(M,A,M): Multiplicative Holt-Winters’ methodwith multiplicative errors

Page 46: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

ETS(A,Ad,N): Damped trend method with addi-tive errors

Page 47: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

There are 30 separate models in the ETSframework

Page 48: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Innovations state space models

SES

yt+1|t = `t

`t = αyt + (1− α)`t−1

Page 49: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

SES

yt+1|t = `t

`t = αyt + (1− α)`t−1

If εt = yt − yt−1|t∼ NID(0, σ2), then

ETS(A,N,N)

yt = `t−1 + εt

`t = αyt + (1− α)`t−1

= `t−1 + αεt

Page 50: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

SES

yt+1|t = `t

`t = αyt + (1− α)`t−1

ETS(A,N,N)

yt = `t−1 + εt

`t = αyt + (1− α)`t−1

= `t−1 + αεt

If εt = (yt − yt−1|t)/yt−1|t∼ NID(0, σ2), then

ETS(M,N,N)

yt = `t−1(1 + εt)

`t = αyt + (1− α)`t−1

= `t−1(1 + αεt)

Page 51: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

SES

yt+1|t = `t

`t = αyt + (1− α)`t−1

All exponential smoothing methods can be writtenusing analogous state space equations.

ETS(A,N,N)

yt = `t−1 + εt

`t = αyt + (1− α)`t−1

= `t−1 + αεt

If εt = (yt − yt−1|t)/yt−1|t∼ NID(0, σ2), then

ETS(M,N,N)

yt = `t−1(1 + εt)

`t = αyt + (1− α)`t−1

= `t−1(1 + αεt)

Page 52: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Example: Holt-Winters’ multiplicativeseasonal method

ETS(M,A,M)

yt = (`t−1 + bt−1)st−m(1 + εt)

`t = (`t−1 + bt−1)(1 + αεt)

bt = bt−1 + β(`t−1 + bt−1)εt

st = st−m(1 + γεt)

where β = αβ∗.

Page 53: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

All the methods can be written in this statespace form.

Prediction intervals can be obtained bysimulating many future sample paths.

For many models, the prediction intervals canbe obtained analytically as well.

Additive and multiplicative versions give thesame point forecasts.

Estimation is handled via maximizing thelikelihood of the data given the model.

Page 54: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 55: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 56: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 57: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 58: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Akaike’s Information Criterion

AIC = −2 log(Likelihood) + 2p

where p is the number of estimated parameters inthe model.

Minimizing the AIC gives the best model forprediction.

AIC corrected (for small sample bias)

AICC = AIC +2(p+ 1)(p+ 2)

n− p

Schwartz’ Bayesian IC

BIC = AIC + p(log(n)− 2)

Page 59: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

AICC = AIC +2(p+ 1)(p+ 2)

n− p

Page 60: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

AICC = AIC +2(p+ 1)(p+ 2)

n− p

Page 61: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

AICC = AIC +2(p+ 1)(p+ 2)

n− p

Page 62: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

AICC = AIC +2(p+ 1)(p+ 2)

n− p

Page 63: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Value of AIC/AICc/BIC given in the R output.

AIC does not have much meaning by itself. Onlyuseful in comparison to AIC value for anothermodel fitted to same data set.

Consider several models with AIC values closeto the minimum.

A difference in AIC values of 2 or less is notregarded as substantial and you may choosethe simpler but non-optimal model.

AIC can be negative.

Page 64: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 65: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 66: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 67: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 68: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

From Hyndman et al. (2008):

Apply each of 30 methods that are appropriateto the data. Estimate parameters and initialvalues using MLE.

Select best method using AIC.

Produce forecasts using best method.

Obtain prediction intervals using underlyingstate space model.

Method performed very well in M3 competition.

Page 69: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 70: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 71: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 72: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 73: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Page 74: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

fit <- ets(ausbeer)fit2 <- ets(ausbeer,model="AAA",damped=FALSE)fcast1 <- forecast(fit, h=20)fcast2 <- forecast(fit2, h=20)

ets(y, model="ZZZ", damped=NULL, alpha=NULL,beta=NULL, gamma=NULL, phi=NULL,additive.only=FALSE,lower=c(rep(0.0001,3),0.80),upper=c(rep(0.9999,3),0.98),opt.crit=c("lik","amse","mse","sigma"), nmse=3,bounds=c("both","usual","admissible"),ic=c("aic","aicc","bic"), restrict=TRUE)

Page 75: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

fit <- ets(ausbeer)fit2 <- ets(ausbeer,model="AAA",damped=FALSE)fcast1 <- forecast(fit, h=20)fcast2 <- forecast(fit2, h=20)

ets(y, model="ZZZ", damped=NULL, alpha=NULL,beta=NULL, gamma=NULL, phi=NULL,additive.only=FALSE,lower=c(rep(0.0001,3),0.80),upper=c(rep(0.9999,3),0.98),opt.crit=c("lik","amse","mse","sigma"), nmse=3,bounds=c("both","usual","admissible"),ic=c("aic","aicc","bic"), restrict=TRUE)

Page 76: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Exponential smoothing> fitETS(M,Md,M)

Smoothing parameters:alpha = 0.1776beta = 0.0454gamma = 0.1947phi = 0.9549

Initial states:l = 263.8531b = 0.9997s = 1.1856 0.9109 0.8612 1.0423

sigma: 0.0356

AIC AICc BIC2272.549 2273.444 2302.715

Page 77: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

> fit2ETS(A,A,A)

Smoothing parameters:alpha = 0.2079beta = 0.0304gamma = 0.2483

Initial states:l = 255.6559b = 0.5687s = 52.3841 -27.1061 -37.6758 12.3978

sigma: 15.9053

AIC AICc BIC2312.768 2313.481 2339.583

Page 78: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

ets() function

Automatically chooses a model by default usingthe AIC, AICc or BIC.

Can handle any combination of trend,seasonality and damping

Produces prediction intervals for every model

Ensures the parameters are admissible(equivalent to invertible)

Produces an object of class ets.

Page 79: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

ets() function

Page 80: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

ets() function

Page 81: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

ets() function

Page 82: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

ets() function

Page 83: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

ets objects

Methods: coef(), plot(), summary(),

residuals(), fitted(), simulate()

and forecast()

plot() function shows time plots of the

original time series along with the

extracted components (level, growth

and seasonal).

Page 84: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

ets objects

Methods: coef(), plot(), summary(),

residuals(), fitted(), simulate()

and forecast()

plot() function shows time plots of the

original time series along with the

extracted components (level, growth

and seasonal).

Page 85: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

200

400

600

obse

rved

250

350

450

leve

l

0.99

01.

005

slop

e

0.9

1.1

1960 1970 1980 1990 2000 2010

seas

on

Time

Decomposition by ETS(M,Md,M) methodplot(fit)

Page 86: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Goodness-of-fit

> accuracy(fit)ME RMSE MAE MPE MAPE MASE

0.17847 15.48781 11.77800 0.07204 2.81921 0.20705

> accuracy(fit2)ME RMSE MAE MPE MAPE MASE

-0.11711 15.90526 12.18930 -0.03765 2.91255 0.21428

Page 87: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Forecast intervals

Forecasts from ETS(M,Md,M)

1995 2000 2005 2010

300

350

400

450

500

550

600

> plot(forecast(fit,level=c(50,80,95)))

Page 88: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Forecast intervals

Forecasts from ETS(M,Md,M)

1995 2000 2005 2010

300

350

400

450

500

550

600

> plot(forecast(fit,fan=TRUE))

Page 89: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

ets() function also allows refitting model to newdata set.

> usfit <- ets(usnetelec[1:45])> test <- ets(usnetelec[46:55], model = usfit)

> accuracy(test)ME RMSE MAE MPE MAPE MASE

-3.35419 58.02763 43.85545 -0.07624 1.18483 0.52452

> accuracy(forecast(usfit,10), usnetelec[46:55])ME RMSE MAE MPE MAPE MASE

40.7034 61.2075 46.3246 1.0980 1.2620 0.6776

Page 90: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Unstable models

ETS(M,M,A)

ETS(M,Md,A)

ETS(A,N,M)

ETS(A,A,M)

ETS(A,Ad,M)

ETS(A,M,N)

ETS(A,M,A)

ETS(A,M,M)

ETS(A,Md,N)

ETS(A,Md,A)

ETS(A,Md,M)Forecasting using R Exponential smoothing state space models 30

Page 91: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Unstable models

ETS(M,M,A)

ETS(M,Md,A)

ETS(A,N,M)

ETS(A,A,M)

ETS(A,Ad,M)

ETS(A,M,N)

ETS(A,M,A)

ETS(A,M,M)

ETS(A,Md,N)

ETS(A,Md,A)

ETS(A,Md,M)Forecasting using R Exponential smoothing state space models 30

In practice, the modelswork fine for short- tomedium-term forecastsprovided the data arestrictly positive.

Page 92: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Forecastability conditions

ets(y, model="ZZZ", damped=NULL, alpha=NULL,beta=NULL, gamma=NULL, phi=NULL,additive.only=FALSE,lower=c(rep(0.0001,3),0.80),upper=c(rep(0.9999,3),0.98),opt.crit=c("lik","amse","mse","sigma"),nmse=3,bounds=c("both","usual","admissible"),ic=c("aic","aicc","bic"), restrict=TRUE)

Page 93: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

The magic forecast() function

forecast returns forecasts when applied to anets object (or the output from many other timeseries models).

If you use forecast directly on data, it willselect an ETS model automatically and thenreturn forecasts.

Page 94: Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

The magic forecast() function

forecast returns forecasts when applied to anets object (or the output from many other timeseries models).

If you use forecast directly on data, it willselect an ETS model automatically and thenreturn forecasts.

Download - Forecasting using - Rob J Hyndman 1Exponential smoothing methods so far 2Holt-Winters’ seasonal method 3Taxonomy of exponential smoothing methods 4Exponential smoothing state space

Top Related