TeknikForecasting

Pendekatan Basis Peramalan ekstrapolatif Ekstrapolasi trend Teknik Analisis rangkaian-waktu Teknik benang-hitam Teknik OLS Pembobotan eksponensial Transformasi data Metode katastrofi Pemetaan teori Analisis jalur Analisis Input Output Input-Output Pemrograman linier Analisis regresi Estimasi interval Analisis hubungan Delphi konvensional Delphi kebijakan Analisis dampak-silang Penilaian kelayakan Hasil Projeksi

Peramalan Teoretis

Teori

Prediksi

Peramalan intuitif

Penilaian subjektif

Konjektur

AsumsiPeramalanEkstrapolatif
1. Keajegan(persistence):Polayangterjadidimasa laluakantetapterjadidimasamendatang.Mis: jikakonsumsienergidimasalalumeningkat,ia jika konsumsi energi di masa lalu meningkat, ia akanselalumeningkatdimasadepan. 2. Keteraturan(regularity):Variasidimasalaluakan secarateraturmunculdimasadepan.Mis:jika secara teratur muncul di masa depan Mis: jika banjirbesardiJakartaterjadisetiap16tahun sekali,polaygsamaakanterjadilagi. 3. Keandalan(reliability)dankesahihan(validity) data:Ketepatanramalantergantungkepada keandalandankesahihandataygtersedia.Mis: yg datattglaporankejahatanseringkalitidaksesuai dginsidenkejahatanygsesungguhnya,datattggaji bukanmerupakanukurantepatdaripendapatan bukan merupakan ukuran tepat dari pendapatan masyarakat.
2

Klasifikasi MetodePeramalan
Forecasting Method
Objective Forecasting Methods
Time Series Methods M th d Nave Methods Moving Averages Exponential Smoothing Simple Regression ARIMA Neural Networks References : Causal Methods M th d Simple Regression Multiple Regression Neural Networks PERT

Subjective (Judgmental) Forecasting Methods

Analogies

Delphi

Survey techniques

Combination of Time Series Causal Methods

Intervention Model Transfer Function (ARIMAX) VARIMA (VARIMAX) Neural Networks

Makridakis et al. Hanke and Reitsch Wei, W.W.S. Box, Jenkins and Reinsel

Klasifikasi Metode Peramalan :

Ilustrasi Model M M d l Matematis i

Forecasting Method
Objective Forecasting Methods o ecast g et ods Subjective (Judgmental) Forecasting Methods o ecast g et ods

Time Series Methods Yt= f (Yt-1, Yt-2, , Yt-k)

Causal Methods Yt= f (X1t, X2t, , Xkt)

Examples : sales(t) = f ( i (t), advert(t), ) l (price d t )

Examples : sales(t) = f ( l (t-1), sales(t-2), ) l (sales l )

Combination of Time Series Causal Methods Yt= f (Yt-j , j>0 ; Xt-i , i0)
Examples : sales(t) = f (sales(t-1), advert(t), advert(t-1), )

Klasifikasi Model Time Series : Berdasarkan Bentuk t Fungsi B t k atau F i

TIME SERIES MODELS
LINEAR Time Series Models ARIMA Box-Jenkins Intervention Model Transfer F T f Function (ARIMAX) ti VARIMA (VARIMAX) NONLINEAR Time Series Models Models from time series theory nonlinear autoregressive, etc ... Flexible statistical parametric models neural network model, etc ... State-dependent, time-varying parap , y gp meter and long-memory models Nonparametric models Models from economic theory

References : Timo Terasvirta, Dag Tjostheim and Clive W.J. Granger, (1994) Aspects of Modelling Nonlinear Time Series Aspects Series
Handbook of Econometrics, Volume IV, Chapter 48. Edited by R.F. Engle and D.I. McFadden

POLA DATATimeSeries O e Se es
GeneralTimeSeries PATTERN General Time Series PATTERN Stationer Trend (linear or nonlinear) Seasonal (additiveormultiplicative) ( p ) Cyclic CalendarVariation Calendar Variation

General of Time Series Patterns

Time Series Patterns
Stationer Trend Effect Seasonal Effect Cyclic Effect

Nonseasonal y Stationary models

Nonseasonal y Nonstationary models

Seasonal and p Multiplicative models

Intervention models

Contoh DATAEKONOMI 1
Time S eries Plot of Inflasi
3

Eidsholiday effects
12 12 12 1 11 12 11

Inflasi

11

-1 Month Year Yea Jan 1999 Jan 2000 Jan 2001 Jan 2002 Jan 2003 Jan 2004 Jan 2005

Contoh DATAEKONOMI 2

Krisis di Indonesia Pertengahan 1997

Reference :

Badan Pusat Statistik (BPS) Indonesia

Contoh DATAEKONOMI 3
Krisis di Indonesia Terjadi Mulai Pertengahan 1997

Reference :

Dinas Perhubungan Jawa Timur

10

ModelmodelTimeSeriesRegression g

1. ModelRegresiuntukLINEARTREND
Yt =a+b.t+error t=1,2,(dummywaktu)

2. ModelRegresiuntukDataSEASONAL (variasikonstan)
Yt =a+b1 D1 ++bS1 DS1 +error dengan : D1,D2,,DS 1 adalah dummy waktu dalam : D DS1 adalahdummywaktudalam satuperiodeseasonal.

3. ModelRegresiuntukDatadenganLINEARTREND danSEASONAL (variasikonstan) ( )

Yt =a+b.t+c1 D1 ++cS1 DS1 +error Gabunganmodel1dan2.

11

Problem4:HasilRegresiTrenddgMINITAB(continued) Problem 4: Hasil Regresi Trend dg MINITAB

12

Problem4:HasilRegresiTrenddgMINITAB Problem 4: Hasil Regresi Trend dg MINITAB

(continued)

13

Problem5:RegresiDataSeasonal (DataElectricalUsage) Problem 5: Regresi Data Seasonal (Data Electrical Usage)

TimeSeriesPlot(Data Time Series Plot (Data seasonal)

14

Problem5:HasilregresidenganMINITAB Problem 5: Hasil regresi dengan MINITAB

MTB > Regress 'Kilowatts' 3 'Kuartal-1'-'Kuartal-3'

The regression equation is

Kilowatts = 722 + 281 Kuartal.1 - 97.4 Kuartal.2 - 202 Kuartal.3 Predictor Constant Kuartal.1 Kuartal.2 Kuartal.3 S = 30.84 Coef 721.60 281.20 -97.40 -202.20 SE Coef 13.79 19.51 19.51 19.51 T 52.32 14.42 -4.99 -10.37 P 0.000 0.000 0.000 0.000

R-Sq = 97.7%

R-Sq(adj) = 97.3%

Analysis of Variance Source DF Regression 3 Residual Error 16 Total 19

SS 646802 15220 662022

MS 215601 951

F 226.65

P 0.000

15

Problem6:RegresiDataTrendLinear danSeasonal Problem 6: Regresi Data Trend Linear dan Seasonal

TimeSeriesPlot(Datatrend danseasonal)
16

Problem6:HasilregresidenganMINITAB Problem 6: Hasil regresi dengan MINITAB

DummyVariable
17

Problem6:HasilregresidenganMINITAB Problem 6: Hasil regresi dengan MINITAB

MTB > Regress 'Sales' 4 't' 'Kuartal.1'-'Kuartal.3' The regression equation is Sales = 413 + 19.7 t + 130 Kuartal.1 - 108 Kuartal.2 - 228 Kuartal.3 16 cases used 4 cases contain missing values d t i i i l Predictor Constant t Kuartal.1 K artal 1 Kuartal.2 Kuartal.3 S = 35.98 Coef 412.81 19.719 130 41 130.41 -108.06 -227.78 SE Coef 26.99 2.012 26 15 26.15 25.76 25.52 T 15.30 9.80 4 99 4.99 -4.19 -8.92 P 0.000 0.000 0 000 0.000 0.001 0.000

R-Sq = 96.3%

R-Sq(adj) = 95.0%

Analysis of Variance Source Regression Residual Error Total DF 4 11 15 SS 371967 14243 386211 MS 92992 1295 F 71.82 P 0.000

18

Problem6:HasilregresidenganMINITAB Problem 6: Hasil regresi dengan MINITAB

Forecast

TimeSeriesPlot(Data dan Ramalannya)

19

Perbandinganketepatanramalanantarmetode Perbandingan ketepatan ramalan antar metode

KasusSalesVideoStore K S l Vid St Kriteria kesalahan ramalan MSE 66.6963 28.7083 21.6829 MAD 6.68889 4.4236 3.73048 MAPE 0.9557 0.6382 0.5382

KasusSalesDataKuartalan K S l D t K t l Kriteria kesalahan ramalan MSE 4372.69 890.215 MAD 52.29 23.2969 MAPE 9.67 4.3122

Model Double M.A. Holts H lt Method Regresi Trend

Model Winters Method Regresi Trend & Seasonal

HoltsMethod : Alpha (level): 0 202284 (level):0.202284 Gamma (trend):0.234940

WintersMethod : Alpha (level):0.4 Gamma (trend):0.1 Delta (seasonal):0.3

20

Contoh
Tahun2000: Penjualanmobil =Rp.300M Indekshargamobil =135 Tahun2001: Penjualanmobil =Rp.350M Indekshargamobil =155 Pertanyaan: Berapa peningkatan nominal ? Berapapeningkatannominal? Berapapeningkatanriilnya?

21

Perhitungan g
Peningktan Nominal PeningktanNominal Rp.350M Rp.300 p p M =Rp.50M PeningktanRiil g Penjualantahun2000 (Rp.300M)(100/135) =Rp.222,222M Penjualantahun2001 (Rp.350M)(100/155) ( )( / ) =Rp.225,806M Peningkatan: 225,806M 222,222M =Rp.3,584M Rp. 3,584 M
22

Mr.Aringanuakanmenganalisislajupertumbuhanpenjualan tokoyangmenjual70%mebelerdan30% alatrumahtangga.

Tahun Penj IH Mebel IH ART 1983 42.1 111.6 105.3 1984 47 2 47.2 117.2 108.5 117 2 108 5 1985 48.4 124.2 109.8 1986 50.6 128.3 114.1 1987 55.2 136.1 117.6 1988 57.9 139.8 122.4 1989 59.8 145.7 128.3 1990 60.7 156.2 131.2

IH Penj Riil Penj. 109.7 38.4 114.6 114 6 41.2 41 2 119.9 40.4 124.0 40.8 130.6 42.3 134.6 43.0 140.5 42.6 148.7 40.8
23

Trend TahunPertamaTahunDasar

Thn X Th Penj P j (Y) X^2 XY 1990 0 108 0 0 1991 1 119 1 119 1992 2 110 4 220 1993 3 122 9 366 1994 4 130 16 520 JMH 10 589 30 1225

24

Trend TitikTengahsbgtahunDasar k h b h
Thn X Penj (Y) X^2 XY 1990 -2 2 108 4 -216 216 1991 -1 119 1 -119 1992 0 110 0 0 1993 1 122 1 122 1994 2 130 4 260 JMH 0 589 10 47

25

Trend Eksponensial k l
Thn X Penj (Y) Log Y X log Y 1990 -2 108 2.0334 -4.0668 1991 -1 119 2.0755 -2.0755 1992 0 110 2.0414 0 1993 1 122 2.0864 2.0864 1994 2 130 2 1139 4 2279 2.1139 4.2279 JMH 0 589 10.351 0.1719
26

TrendKuadratik Trend Kuadratik

Thn X 1981 -5 5 1982 -3 983 1983 -1 1984 1 1985 3 1986 5 Jlh 0 Y 2 5 8 15 26 37 93 X^2 X^3 25 -125 125 9 -27 1 -1 1 1 9 27 25 125 70 0 X^4 625 81 1 1 81 625 1414 XY -10 10 -15 -8 8 15 78 185 245 X^2Y 50 45 8 15 234 925 1277
27

NaveModel Nave Model

Therecent periods arethebestpredictorsofthefuture. 1. The simplestmodel forstationary datais

Yt +1 = Yt
2. The simplestmodel for trenddata is

Yt +1 = Yt + (Yt Yt 1 ) or
= Y Yt Yt +1 t Yt 1
3. The simplestmodel for seasonaldata is

Yt +1 = Y(t +1) s

28

AverageMethods Average Methods

1.SimpleAverages
obtainedbyfindingthemeanforalltherelevantvalues and g p thenusingthismeantoforecastthenextperiod.

= Yt Yt +1 t =1 n

forstationarydata

2. MovingAverages
obtainedbyfindingthemeanforaspecifiedsetofvalues andthen usingthismeantoforecastthenextperiod. using this mean to forecast the next period

(Y + Yt 1 + K + Yt n +1 ) M t = Yt +1 = t n

forstationarydata

29

AverageMethods(continued) Average Methods

3.DoubleMovingAverages
onesetofmovingaveragesiscomputed,andthenasecondset computedasamovingaverageofthefirstset. is

(Y + Yt 1 + K + Yt n +1 ) M t = Yt +1 = t n ( M t + M t 1 + K + M t n +1 ) (ii). (ii) M t = n
(i). (iii). at = 2M t M t (iv).

bt =

2 ( M t M t ) n 1

Yt + p = at + bt p

foralineartrenddata f li t dd t
30

MINITABimplementation MINITAB implementation

31

MINITABimplementation(continued) MINITAB implementation

32

MovingAveragesResult(continued) Moving Averages Result

33

MovingAverages Moving Averages VS Double Moving Averages Results DoubleMovingAverages Results

MAorMoving Averages

DMAorDouble MovingAverages

MSE.MA=132.67,MSE.DMA=63.7
34

ExponentialSmoothingMethods Exponential Smoothing Methods

SingleExponential Smoothing

for stationarydata

Yt +1 = Yt + (1 )Yt
ExponentialSmoothingAdjustedforTrend :HoltsMethod
1.Theexponentiallysmoothedseries: At = Yt + (1) (At1+ Tt1) = +(1)(A +T 2.Thetrendestimate: Tt = (At At1)+(1 )Tt1 3.Forecastp periodsintothefuture:

Yt + p = At + pTt
35

ExponentialSmoothingAdjustedforTrend andSeasonal Variation:WintersMethod Variation : Winters Method

1.Theexponentiallysmoothed series: Y At = t + (1 ) ( At 1 + Tt 1) St L 2.Thetrend estimate:

Tt = ( At At 1) + (1 )Tt 1

3.Theseasonality estimate: Y St = t + (1 ) St 1 At 4.Forecast p periodsintothefuture :

Yt + p = ( At pTt ) St L + p

Three parameters p models

36

SES:MINITABimplementation SES: MINITAB implementation

SESdengan alpha0,1

SESdengan alpha0,6

37

SES:MINITABimplementation SES: MINITAB implementation

(continued) ( i d)

38

SES:MINITABimplementation SES: MINITAB implementation

(continued) ( i d)

39

DES(Holt sMethods):MINITABimplementation DES (Holts Methods): MINITAB implementation

(continued) ( i d)

DESdenganalpha0,3danbeta0,1
40

DES:MINITABimplementation DES: MINITAB implementation

(continued) ( i d)

41

Winter sMethods:MINITABimplementation Winters Methods: MINITAB implementation

WintersMethodsdenganalpha0,4;beta0,1dangamma0,3
42

Winter sMethods:MINITABimplementation Winters Methods: MINITAB implementation

(continued) ( i d)

43