[go: up one dir, main page]
Scribd
Upload
42 views35 pages

Final Project Forecasting

This document presents a forecast of the exchange rate of the US dollar against the Indonesian rupiah. It begins with introductions and background theory. Time series data from June 2009 to May 2019 is analyzed using regression, smoothing, and time series analysis techniques. The best fitting models are regression (Adj R-Sq: 0.8649, MAPE: 0.0579%) and ARIMA (3,1,1) for log-transformed data (AIC: -19634.65, MAPE: 0.0579%). The conclusion is that regression and ARIMA (3,1,1) provide the best forecasts of exchange rate movements.

Uploaded by

aliciajeh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
42 views35 pages

Final Project Forecasting

This document presents a forecast of the exchange rate of the US dollar against the Indonesian rupiah. It begins with introductions and background theory. Time series data from June 2009 to May 2019 is analyzed using regression, smoothing, and time series analysis techniques. The best fitting models are regression (Adj R-Sq: 0.8649, MAPE: 0.0579%) and ARIMA (3,1,1) for log-transformed data (AIC: -19634.65, MAPE: 0.0579%). The conclusion is that regression and ARIMA (3,1,1) provide the best forecasts of exchange rate movements.

Uploaded by

aliciajeh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 35

FORECAST

KURS US DOLLAR
TERHADAP RUPIAH
FINAL PRESENTATION FORECASTING Alicia J/02310171001
Anthony J/02310171008
Elbert R/02310171011
Kevin A/02310171010
Vania B/02310171012
01
PENDAHULUAN &
DASAR TEORI
PERBEDAAN TINGKAT INFLASI
1.
PERBEDAAN TINGKAT SUKU BUNGA
2.
NERACA PERDAGANGAN
3.
HUTANG PUBLIK
4. FAKTOR-FAKTOR
YANG
5.
3
RATIO HARGA EKSPOR-IMPOR

MEMPENGARUHI
KESTABILAN POLITIK & EKONOMI 6. NILAI TUKAR
REGRESSION 1.
SMOOTHING 2.
FORECASTING 4

TECHNIQUES
TIME SERIES ANALYSIS 3.
02
CALCULATION
Periode: June 2009 - May 2019 IMPORT DATA

6
TIME SERIES PLOT

7
1.
REGRESSION
REGRESSION

9
REGRESSION FORECAST

10
2.
SMOOTHING
SES FORECAST

.. dst 12
3.
TIME SERIES ANALYSIS
TIME SERIES PLOT

14

Tidak Stationer Data tetap tidak


Mean & Variance stationer
original data

BARTLETT TEST & BOX PLOT

15

Ragam tidak
stationer
transformation data

BARTLETT TEST & BOX PLOT

16

Ragam tidak
stationer
ADF & KPSS TEST
original data ln-transformed data

p-value > 0.05 → tidak stationer p-value > 0.05 → tidak stationer

p-value < 0.05 → tidak stationer p-value < 0.05 → tidak stationer
ACF PLOT
original data ln-transformed data

Data tidak Data tidak


stationer stationer
DIFFERENCING 1X
original data ln-transformed data

19

Data terlihat Data terlihat


lebih stationer lebih stationer
ADF & KPSS TEST
original data ln-transformed data

p-value < 0.05 → stationer p-value < 0.05 → stationer


20

p-value > 0.05 → stationer p-value > 0.05 → stationer


ACF and PACF PLOT
MA (0) AR (0)/AR(25)

O
R

MA (0) AR (0)/AR(4)/AR(25)
21

L
N
FITTING ARIMA
AIC MAPE original data
ARIMA (0,1,0) 26249.33 0.289563%

ARIMA (1,1,0) 26250.09 0.290015%


22
ARIMA (0,1,1) 26250.11 0.290002%

ARIMA (0,1,2) 26251.96 0.290106%

ARIMA (25,1,0) 26274.76 0.289251%


M.5
Coefficient Test

Hanya koefisien
AR(25) yang
signifikan
BEST MODEL
original data

Therefore, we 24

choose model 1:
ARIMA(0,1,0)
FITTING ARIMA
AIC MAPE ln transformed
ARIMA (0,1,0) -19639.51 0.0310201% data
ARIMA (4,1,0) -19640.35 0.0309690%
25
ARIMA (25,1,0) -19616.66 0.0308817%

ARIMA (0,1,1) -19638.01 0.0310099%

ARIMA (0,1,2) -19636.01 0.0310102%


M.C
Coefficient Test

Hanya koefisien
AR(4) dan AR(25)
yang signifikan
BEST MODEL
ln transformed
data

Therefore, we 27

choose model F:
ARIMA(3,1,1)
MODEL F FORECAST

.. dst
28
CHECK RESIDUALS (1)
MODEL F: ARIMA (3,1,1)

29

p-value > 0.05 → error independen


CHECK RESIDUALS (2)
MODEL F: ARIMA (3,1,1)

Box Test
30
p-value > 0.05 → independent, no corr

Shapiro Test
p-value < 0.05 → not normal
CHECK RESIDUALS (3)
MODEL F: ARIMA (3,1,1)

31

p-value > 0.05 → white noise


03
CONCLUSION
BEST MODEL

REGRESSION SMOOTHING TIME SERIES ANALYSIS


(SES, Alpha=0.95) (ARIMA (3,1,1) for
33
ln-transformed data)

Adj R-Sq: 0.8649 AIC: -19634.65

MAPE: 0.0579% MAPE: 0.2946%


FORECAST

34
THANKS
35

You might also like