ARIMA
Berikut data bulanan ROA bank umum syariah dari Januari 2009 hingga April 2014. Buatlah model ARIMA dari data berikut:
Tahun
|
ROA
|
Tahun
|
ROA
|
Tahun
|
ROA
|
Jan-09
|
2.11
|
Nop-10
|
1.83
|
Sep-12
|
2.07
|
Feb-09
|
2.15
|
Des-10
|
1.67
|
Okt-12
|
2.11
|
Mar-09
|
2.44
|
Jan-11
|
2.26
|
Nop-12
|
2.09
|
Apr-09
|
2.29
|
Feb-11
|
1.81
|
Des-12
|
2.14
|
Mei-09
|
2.22
|
Mar-11
|
1.97
|
Jan-13
|
2.52
|
Jun-09
|
2.16
|
Apr-11
|
1.90
|
Feb-13
|
2.29
|
Jul-09
|
2.12
|
Mei-11
|
1.84
|
Mar-13
|
2.39
|
Agust-09
|
2.08
|
Jun-11
|
1.84
|
Apr-13
|
2.29
|
Sep-09
|
1.38
|
Jul-11
|
1.86
|
Mei-13
|
2.07
|
Okt-09
|
1.46
|
Agust-11
|
1.81
|
Jun-13
|
2.10
|
Nop-09
|
1.48
|
Sep-11
|
1.80
|
Jul-13
|
2.02
|
Des-09
|
1.48
|
Okt-11
|
1.75
|
Agust-13
|
2.01
|
Jan-10
|
1.65
|
Nop-11
|
1.78
|
Sep-13
|
2.04
|
Feb-10
|
1.76
|
Des-11
|
1.79
|
Okt-13
|
1.94
|
Mar-10
|
2.13
|
Jan-12
|
1.36
|
Nop-13
|
1.96
|
Apr-10
|
2.06
|
Feb-12
|
1.79
|
Des-13
|
2.00
|
Mei-10
|
1.25
|
Mar-12
|
1.83
|
Jan-14
|
0.08
|
Jun-10
|
1.66
|
Apr-12
|
1.79
|
Feb-14
|
0.13
|
Jul-10
|
1.67
|
Mei-12
|
1.99
|
Mar-14
|
1.16
|
Agust-10
|
1.63
|
Jun-12
|
2.05
|
Apr-14
|
1.09
|
Sep-10
|
1.77
|
Jul-12
|
2.05
|
||
Okt-10
|
1.79
|
Agust-12
|
2.04
|
Sebelum dilakukan pemodelan ARIMA dilakukan uji stationeritas
Pada level
H0 : |ρ|=1 ( data variabel mengandung unit root)
H1 : ρ<1 ( data variabel tidak mengandung unit root)
α=0.1
Wilayah kritis: ρ < α
Null Hypothesis: ROA has a unit root
|
||||
Exogenous: Constant, Linear Trend
|
||||
Lag Length: 0 (Automatic based on SIC, MAXLAG=10)
|
||||
t-Statistic |
Prob.* |
|||
Augmented Dickey-Fuller test statistic
|
-3.349326 |
0.0678 |
||
Test critical values:
|
1% level |
-4.110440 |
||
5% level |
-3.482763 |
|||
10% level |
-3.169372 |
|||
*MacKinnon (1996) one-sided p-values.
|
Terlihat nilai p-value=0.0678 > alpha=0,.05 sehingga tidak tolak H0 dan disimpulkan datanya masih mengandung unit root/ belum stationer
Pada diffrence I
*Pada Difference I
H0 : δ=0 ( data variabel mengandung unit root)
H1 : δ=0 ( data variabel tidak mengandung unit root)
α=0.1
Wilayah kritis: ρ < α
Null Hypothesis: D(ROA) has a unit root
|
||||
Exogenous: Constant, Linear Trend
|
||||
Lag Length: 1 (Automatic based on SIC, MAXLAG=10)
|
||||
t-Statistic |
Prob.* |
|||
Augmented Dickey-Fuller test statistic
|
-8.026165 |
0.0000 |
||
Test critical values:
|
1% level |
-4.115684 |
||
5% level |
-3.485218 |
|||
10% level |
-3.170793 |
|||
*MacKinnon (1996) one-sided p-values.
|
Terlihat nilai p-value=0.000 < alpha=0,.05 sehingga tolak H0 dan disimpulkan datanya sudah tidak mengandung unit root/ sudah stationer
Sehingga model yang memungkinkan adalah ARI(1), IM(1), ARIMA (1), dst dan data yang digunakan ada difference I
Model ARI(1)
Klik menu quickà estimate equationàpada specification ketik: “d(roa) c ar(1) “ àOK
Dependent Variable: D(ROA)
|
||||
Method: Least Squares
|
||||
Date: 08/15/14 Time: 08:55
|
||||
Sample (adjusted): 2009M03 2014M04
|
||||
Included observations: 62 after adjustments
|
||||
Convergence achieved after 3 iterations
|
||||
Variable |
Coefficient
|
Std. Error
|
t-Statistic
|
Prob.
|
C |
-0.016858
|
0.038912
|
-0.433236
|
0.6664
|
AR(1) |
-0.155511
|
0.127527
|
-1.219436
|
0.2275
|
R-squared
|
0.024184
|
Mean dependent var
|
-0.017097
|
|
Adjusted R-squared
|
0.007921
|
S.D. dependent var
|
0.355447
|
|
S.E. of regression
|
0.354036
|
Akaike info criterion
|
0.792890
|
|
Sum squared resid
|
7.520492
|
Schwarz criterion
|
0.861508
|
|
Log likelihood
|
-22.57960
|
Hannan-Quinn criter.
|
0.819831
|
|
F-statistic
|
1.487024
|
Durbin-Watson stat
|
2.085953
|
|
Prob(F-statistic)
|
0.227451
|
|||
Inverted AR Roots
|
-.16
|
viewàresidual test à heteroskedascity testàwhite test àOK
Heteroskedasticity Test: White
|
||||
F-statistic
|
0.056824
|
Prob. F(2,59)
|
0.9448
|
|
Obs*R-squared
|
0.119197
|
Prob. Chi-Square(2)
|
0.9421
|
|
Scaled explained SS
|
0.863493
|
Prob. Chi-Square(2)
|
0.6494
|
|
Model IMA(1)
Klik menu quickà estimate equationàpada specification ketik: “d(roa) c ma(1) “ àOK
Dependent Variable: D(ROA) |
||||
Method: Least Squares
|
||||
Date: 08/15/14 Time: 09:00
|
||||
Sample (adjusted): 2009M02 2014M04
|
||||
Included observations: 63 after adjustments
|
||||
Convergence achieved after 8 iterations
|
||||
MA Backcast: 2009M01
|
||||
Variable |
Coefficient
|
Std. Error
|
t-Statistic
|
Prob.
|
C |
-0.018724
|
0.027823
|
-0.672996
|
0.5035
|
MA(1) |
-0.366896
|
0.120185
|
-3.052755
|
0.0034
|
R-squared
|
0.057933
|
Mean dependent var
|
-0.016190
|
|
Adjusted R-squared
|
0.042490
|
S.D. dependent var
|
0.352642
|
|
S.E. of regression
|
0.345069
|
Akaike info criterion
|
0.741084
|
|
Sum squared resid
|
7.263414
|
Schwarz criterion
|
0.809120
|
|
Log likelihood
|
-21.34416
|
Hannan-Quinn criter.
|
0.767843
|
|
F-statistic
|
3.751265
|
Durbin-Watson stat
|
1.794478
|
|
Prob(F-statistic)
|
0.057403
|
|||
Inverted MA Roots
|
.37
|
|||
viewàresidual test à heteroskedascity testàwhite test àOK
Heteroskedasticity Test: White
|
||||
F-statistic
|
0.195253
|
Prob. F(5,57)
|
0.9631
|
|
Obs*R-squared
|
1.060858
|
Prob. Chi-Square(5)
|
0.9575
|
|
Scaled explained SS
|
7.714734
|
Prob. Chi-Square(5)
|
0.1727
|
|
Model ARIMA(1)
Klik menu quickà estimate equationàpada specification ketik: “d(roa) c ar(1) ma(1) “ àOK
Dependent Variable: D(ROA) |
||||
Method: Least Squares
|
||||
Date: 08/15/14 Time: 09:01
|
||||
Sample (adjusted): 2009M03 2014M04
|
||||
Included observations: 62 after adjustments
|
||||
Convergence achieved after 16 iterations
|
||||
MA Backcast: 2009M02
|
||||
Variable |
Coefficient
|
Std. Error
|
t-Statistic
|
Prob.
|
C |
-0.005773
|
0.008228
|
-0.701677
|
0.4856
|
AR(1) |
0.667741
|
0.105854
|
6.308110
|
0.0000
|
MA(1) |
-0.973707
|
0.023518
|
-41.40192
|
0.0000
|
R-squared
|
0.137502
|
Mean dependent var
|
-0.017097
|
|
Adjusted R-squared
|
0.108265
|
S.D. dependent var
|
0.355447
|
|
S.E. of regression
|
0.335654
|
Akaike info criterion
|
0.701707
|
|
Sum squared resid
|
6.647164
|
Schwarz criterion
|
0.804633
|
|
Log likelihood
|
-18.75292
|
Hannan-Quinn criter.
|
0.742118
|
|
F-statistic
|
4.702992
|
Durbin-Watson stat
|
1.958051
|
|
Prob(F-statistic)
|
0.012731
|
|||
Inverted AR Roots
|
.67
|
|||
Inverted MA Roots
|
.97
|
|||
viewàresidual test à heteroskedascity testàwhite test àOK
Heteroskedasticity Test: White
|
||||
F-statistic
|
0.692910
|
Prob. F(9,52)
|
0.7120
|
|
Obs*R-squared
|
6.639233
|
Prob. Chi-Square(9)
|
0.6746
|
|
Scaled explained SS
|
49.16673
|
Prob. Chi-Square(9)
|
0.0000
|
|
Perbandingan antara model yang akan dipilih:
Model
|
Sign.
Model
|
Asumsi
Heterosedastis
|
R-square
|
AIC
|
SC
|
ARI(1)
|
Tdk Sign.
|
bebas
|
0.024184
|
0.792890
|
0.861508
|
IMA(1)
|
Sign.
|
bebas
|
0.057933
|
0.741084
|
0.809120
|
ARIMA(1,1)
|
Sign.
|
bebas
|
0.137502
|
0.701707
|
0.804633
|
Pemilihan
model yang terbaik adalah model yang signifikan, nilai r-square besar,
serta AIC dan SC yang terkecil sehingga model yang terpilih adalah model
ARIMA(1,1)
Selanjutnya kita uji model terpilih dengan klik menu “Forecast àOK
Terlihat nilai bias proporsi 0.09 dibawah 0.2, sehingga model yang digunakan baik untuk melakukan forecast.
Bagaimana jika dalam pengujian ARIMA (1,1,0) sign modelnya masih tidak sign, dan asumsi Heterosedastis nya masih ditolak?
BalasHapusDan untuk pengujian ARIMA berikutnya, misalnya (0,1,1) atau (1,1,1), dalam input eviews nya seperti apa pada specification nya?
Tutorial Lengkap Pool Data Panel Dengan EVIEWS
BalasHapusMerupakan Tutorial Regresi Data Panel Model Pool
Dengan Menggunakan EVIEWS Sehingga Disebut Dengan
Tutorial Lengkap Pool Data Panel Dengan EVIEWS
Klik Link Dibawah Ini Untuk Mendapatkan Tutorialnya
https://s.id/Panel