You are on page 1of 11

PEKERJAAN RUMAH –

ARMA MODELING
M S 5 2 1 1 S t a ti s ti k a Te r a p a n
Dharmestha Dewantoro 23120322
MODEL SINYAL ERROR
>> e=normrnd(0,1.5,[1200 1]);

Error terdistribusi normal dengan:

i. mean = 0

ii. Std. dev. = 1.5

iii. Jml Data = 1200

Add a Footer 2
NORMALISASI SINYAL ERROR

Add a Footer 3
TITLE

PEMODELAN FILTER Unit Circle


• Semua Poles dan Zero ada di dalam
lingkaran unit (unit circle)
C(z) = 1 + 0.5z-1 X
A(z) = 1 + 0.5z-1 + z-2

• Filter: o o
C(z)/A(z) = 1 + 0.5z-1 / 1 + 0.5z-1 + z-2

• Poles: X
pole1 = -0.2500 + 0.9682i

pole2 = -0.2500 - 0.9682i

• Zeros:
zero1 = -0.500

zero2 = 0 4
I. MENGHASILKAN DATA Y
II. NORMALISASI DATA Y
i. >> y=filter([1 0.5],[1 0.5 1],e);

ii. >> y=(y – mean(y));

Add a Footer 5
PEMODELAN ARMA(NA,NC)

• Dipilih orde na = 2

• Dipilih orde nc = 1

• th=armax(y,[2 1])

• Diperoleh:

• A(z) mendekati theoretic

• C(z) mendekati theoretic

Add a Footer 6
>> [A,B,C,D,F]=polydata(th)

A =

POLYDATA 1.0000 0.5008 1.0007

• polydata Quick access to polynomial data.


B =

[A,B,C,D,F] = polydata(SYS) returns the A,B,C,D,F 1×0 empty double row vector
polynomials of the identified state-space model SYS.
C =

1.0000 0.5203

D =

F = 1×0 empty double row vector


Add a Footer 7
IDENTIFIKASI WHITE NOISE
>> k=0:40;
>> atas=ones(size(k))*1.96/sqrt(1000);
>> bawah=-atas;
>> stem(k,Ren);
>> hold
Current plot held
>> plot(k,atas,'--',k,bawah,'--')

Add a Footer 8
PLOT AUTOCOVARIANCE
SINYAL

Add a Footer 9
ONE STEP AHEAD
PREDICTION
>> y1=y(1:900);

>> y2=y(901:1200);

>> y1=(y1-mean(y1));

>> y2=(y2-mean(y2));

>> th1=armax(y1,[2 1]);[A1,B1,C1,D1,F1]=polydata(th1);

>> ehat1=filter(A1,C1,y2);
i. Nilai prediksi tidak jauh dengan
>> ypred=y2-ehat1;

>> kk=901:1200; nilai asli


>> k2=21:40;
ii. Nilai prediksi tidak di luar
>> ATAS=ypred+std(ehat1);

>> bawah=ypred-std(ehat1); confidence interval


>> figure(3)

>> plot(kk(k2),y2(k2),kk(k2),ypred(k2),'*',kk(k2),ATAS(k2),’--',kk(k2),bawah(k2),'--') 10
THANK YOU
Ve r y m u c h

Add a Footer 11

You might also like