554-suhartono-statistics-forecast1.ppt 1f2di

ANALISIS TIME SERIES SM 2234 / 3 SKS Pengajar : 1. Dr. Suhartono, S.Si., M.Sc. S1 Statistika, ITS, 1995 S2 Statistika, UMIST, UK, 1998 S3 Statistika, UGM, 2007 email : [email protected] phone : 5933867; 0818376367 (HP)

2. R. Mohamad Atok, S.Si., M.Si. S1 Statistika, ITS, 1995 S2 Statistika, IPB, 2004

MATERI PERKULIAHAN … 1. 2. 3. 4. 5.

Pendahuluan Konsep Dasar dalam Analisis Time Series Model untuk Time Series yang Stasioner Model untuk Time Series yang Nonstasioner Identifikasi, estimasi parameter, dan diagnostic check model ARIMA 6. Peramalan dengan model ARIMA

7. Model untuk Time Series yang Musiman 8. Model ARIMAX : Analisis Intervensi, Model Variasi Kalender, Fungsi Transfer dan Neural Networks

BUKU ACUAN / Referensi … 1. Hanke, J.E. and Reitsch, A.G. (1995 & 2001) Business Forecasting, 5th and 7th edition, Prentice Hall. 2. Bowerman, B.L. and O’Connell, R.T. (1993) Forecasting and Time Series: An Applied Approach, 3rd edition, Duxbury Press: USA. 3. Makridakis, S., Wheelwright, S. C. and Hyndman, R. J. (1998) Forecasting: Method and Applications, New York: Wiley & Sons. 4. Cryer, J.D. (1986) Time Series Analysis, Boston: PWS-KENT Publishing Company. 5. Wei, W.W.S. (1990)

Time Series Analysis: Univariate and Multivariate Methods Addison-Wesley Publishing Co., USA. 6. Brockwell, P.J. and Davis, R.A. (1991) Time Series: Theory and Methods, 2nd edition, Springer Verlag.

Peramalan & Perencanaan … 1 

The Future Can Not Be Predicted Robert T. Kiyosaki books



The Future Can Not Be Predicted “PRECISELY”  New Paradigm

 A PERSON WHO DOESN’T CARE ABOUT “THE PAST“

IS A PERSON WHO DOESN’T HAVE “THE FUTURE” 

Peramalan & Perencanaan … 2 

Ramalan merupakan input bagi proses perencanaan dan pengambilan keputusan.



Peramalan menunjukkan perkiraan yang akan terjadi pada suatu keadaan tertentu.



Sebaliknya, perencanaan menggunakan ramalan tersebut untuk membantu para pengambil keputusan dalam memilih alternatif terbaik.



Dengan demikian, suatu ramalan mencoba untuk memperkirakan apa yang akan terjadi, sedangkan perencanaan adalah upaya para pengambil keputusan untuk dapat mempengaruh hasil yang akan terjadi melalui berbagai strategis, misalnya rencana promosi, distribusi dll.

Apa itu ILMU (science) ... !!! ... ILMU (science) tidak lagi hanya meneliti : 1. BAGAIMANA dan APA SEBABNYA sesuatu itu terjadi, (to understand what’s going on) 2. MERAMALKAN apa yang akan terjadi, (to forecast what will happen) 3. tetapi juga MEMPENGARUHI atau MERUBAH apa yang akan terjadi (to CHANGE what will happen).

KRISIS MONETER 1997 ... ... Yang sangat menakjubkan adalah bahwa krisis luar biasa ini tidak ada seorangpun yang menduga akan terjadi termasuk pakar-pakar “PERAMAL MASA DEPAN” (futurolog). Para ekonom baik yang bekerja di pemerintahan maupun di luar pemerintah termasuk di dunia bisnis, bahkan pakar-pakar ekonomi kaliber dunia dari luar negeri tidak ada satupun yang mengira krisis moneter (keuangan) ini akan melanda Indonesia. ... (Mubyarto, Pemberdayaan Ekonomi Rakyat & Peranan Ilmu-Ilmu Sosial, Yogyakarta, 2002, hal. 43)

KRISIS MONETER 1997 ...

[continued]

... Ternyata sampai “detik-detik terakhir” menjelang krismon, pakar-pakar ekonomi andal sekalipun yakin tidak akan terjadi krismon di Indonesia. Sebaliknya, jauh sebelum terjadinya, ada sejumlah pakar terutama pakar-pakar ilmu politik, sosiologi dan antropologi, yang telah memperingatkan kemungkinan meledaknya “bom waktu” jika ketimpangan-ketimpangan ekonomi dan sosial yang muncul tidak memperoleh perhatian sewajarnya. ... (Mubyarto, Pemberdayaan Ekonomi Rakyat & Peranan Ilmu-Ilmu Sosial, Yogyakarta, 2002, hal. 50)

D, I, K, Transition Add Value

DATA

by

Transformed

INFORMATION

Analysis Correlating Summarizing

by

KNOWLEDGE

Contextualisation Experience Interpretation Discussion

Klasifikasi Metode Peramalan … Forecasting Method Objective Forecasting Methods Time Series Methods

Subjective (Judgmental) Forecasting Methods

Causal Methods

Analogies

Naïve Methods

Simple Regression

Moving Averages

Multiple Regression

Exponential Smoothing

Neural Networks

Delphi

PERT

Simple Regression

Survey techniques

ARIMA Neural Networks References :

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 Matematis … Forecasting Method Objective Forecasting Methods Time Series Methods Yt= f (Yt-1, Yt-2, … , Yt-k)

Subjective (Judgmental) Forecasting Methods

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

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

Examples :  sales(t) = f (price(t), advert(t), …)

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

Klasifikasi Model Time Series : Berdasarkan Bentuk atau Fungsi … TIME SERIES MODELS LINEAR Time Series Models

NONLINEAR Time Series Models ARIMA Box-Jenkins

Models from time series theory  nonlinear autoregressive, etc ...

Intervention Model

Flexible statistical parametric models  neural network model, etc ...

Transfer Function (ARIMAX)

State-dependent, time-varying parameter and long-memory models

VARIMA (VARIMAX)

Nonparametric models

Models from economic theory

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

POLA DATA Time Series … General Time Series “PATTERN”  Stationer  Trend (linear or nonlinear)  Seasonal (additive or multiplicative)  Cyclic  Calendar Variation

General of Time Series Patterns … Time Series Patterns Stationer

 Nonseasonal Stationary models

Trend Effect

 Nonseasonal Nonstationary models

Seasonal Effect

 Seasonal and Multiplicative models

Cyclic Effect

 Intervention models

POLA DATA Time Series … continued Examples of “Calendar Variation” in Gregorian (Masehi) calendar measurement:  Eids holiday (one of Islamic Calendar effects) It happens on different month after three years or shift to previous month after at the same month on three years

 Hindu’s holidays in Bali  Imlek holiday

Contoh DATA EKONOMI … 1 Time Series Plot of Inflasi 3

1

Eids holiday effects

2

12

12

12

Inflasi

1

11 12

1

11 11

0

-1 Month Year

Jan 1999

Jan 2000

Jan 2001

Jan 2002

Jan 2003

Jan 2004

Jan 2005

Contoh DATA EKONOMI … 2

Krisis di Indonesia Pertengahan 1997

Reference :  Badan Pusat Statistik (BPS) Indonesia

Contoh DATA EKONOMI … 3 Krisis di Indonesia Terjadi Mulai Pertengahan 1997

Reference :  Dinas Perhubungan Jawa Timur

Contoh DATA EKONOMI … 4 Time Series Plot of Penumpang KA Eksekutif dan Pesawat 200000

Variable KA Ek sek utif Pesawat

Data

150000

100000

50000 Krisis di Indonesia Terjadi Mulai Pertengahan 1997

0 Month Jan Year 1993

Jan 1994

Jan 1995

Jan 1996

Jan 1997

Jan 1998

Jan 1999

Jan 2000

Reference :  Dinas Perhubungan Jawa Timur

Contoh DATA TOURISM … 5

Krisis di Indonesia Pertengahan 1997

Reference :  Badan Pusat Statistik (BPS) Bali

Contoh DATA TOURISM … 6 Time Series Plot of Tingkat Hunian Hotel 4* dan Hotel 5* 100 Variable Hotel 4* Hotel 5*

90 80

Data

70 60 50 40 30 Krisis di Indonesia Terjadi Mulai Pertengahan 1997

20 Month Year

Jan 1994

Jan 1995

Jan 1996

Jan 1997

Jan 1998

Jan 1999

Jan 2000

Jan 2001

Bom BALI

Jan 2002

Jan 2003

Reference :  Badan Pusat Statistik (BPS) Bali

Contoh DATA HIDROLOGI … 7 Time Series Plot of Curah hujan and Debit air in Telaga Ngebel, Madiun 0.30 Variable C urah hujan Debit air

0.25

Data

0.20 0.15 0.10 0.05 0.00 1

10

20

30

40

50 Index

60

70

80

90

Contoh Data EKONOMI … 8

Keterangan : Data penjualan SARDEN di PT Blambangan Jaya, Banyuwangi

Model ARIMA The Box-Jenkins methodology … 1. Tentative IDENTIFICATION  historical data are used to tentatively identify an appropriate ARIMA model.

2. ESTIMATION  historical data are used to estimate the parameters of the tentatively identified model.

3. DIAGNOSTIC CHECKING  various diagnostics are used to check the adequacy of the tentatively identified model,  if need be, to suggest an improved model, which is then regarded as a new tentatively identified model.

4. FORECASTING  once a final model is obtained, it is used to forecast future time series values.

Flow Diagram of Box-Jenkins … 1. Tentative IDENTIFICATION

NO

2. Parameter ESTIMATION

3. DIAGNOSTIC CHECKING  [ Is the model adequate? ]

 Stationary and nonstationary time series  ACF dan PACF (theoritical)

 Testing parameters

 White noise of residual  Normal Distribution of residual

YES 4. FORECASTING

 Forecast calculation

INFORMASI LEBIH LANJUT ... 





Literature dan websites Forecasting website  www.forecastingprincipales.com  www.neural-forecasting.com atau www.bis-lab.com Journals Forecasting ... General literature !  Journal of Time Series Analysis  JBF  Journal of Business Forecasting  IJF  International Journal of Forecasting  JoF  Journal of Forecasting Softwares  MINITAB   SPSS  SAS …