TUGAS #1 ANALISIS STATISTIK
ANALISIS REGRESI KUADRATIK
PENGAJAR: TOTO WARSA, Ir., M.S.
PENYUSUN: Ade Setiawan 150220060003
PROGRAM PASCASARJANA UNIVERSITAS PADJADJARAN 2006
1
Soal: Berikut ini adalah data hasil percobaan pemupukan nitrogen pada tanaman padi: Dosis Pupuk N (Xi) 0 30 60 90 120
Hasil Panen (Yi) 2.1 3.4 4.2 7.1 6.3
Tentukan persamaan regresi kuadratiknya!
Jawab: a. Model Regresi Kuadratik Y = b0 + b1 X i + b2 X i + ε 2
X0 1 1 1 1 1
X1 0 30 60 90 120
X2 0 900 3600 8100 14400
Y 2.1 3.4 4.2 7.1 6.3
b. Definisikan Matriks-matriks Berikut: 0 ⎞ ⎛1 0 ⎛ 2.1 ⎞ ⎜1 30 ⎜ 3.4 ⎟ ⎛ b0 ⎞ 900 ⎟ ⎟ ⎜ ⎟ ⎜ b = ⎜⎜ b1 ⎟⎟ X = 1 60 3600 Y = 4.2 ⎜1 90 8100 ⎟ ⎜ 7.1 ⎟ ⎜b ⎟ ⎝ 2⎠ ⎜1 120 14400 ⎟ ⎜ 6.3 ⎟ ⎠ ⎝ ⎝ ⎠ 1 1 1 ⎞ ⎛1 1 X ′ = ⎜ 0 30 60 90 120 ⎟ ⎜ 0 900 3600 8100 14400 ⎟ ⎠ ⎝
0 ⎞ ⎛1 0 1 1 1 ⎞ ⎜1 30 900 ⎟ ⎛1 1 X ′X = ⎜ 0 30 60 90 120 ⎟ ⎜1 60 3600 ⎟ ⎜ 0 900 3600 8100 14400 ⎟ ⎜1 90 8100 ⎟ ⎝ ⎠⎜ ⎟ ⎝1 120 14400 ⎠ 300 2700 ⎛5 ⎞ 27000 2700000 ⎟ = ⎜ 300 ⎜ 27000 2700000 286740000 ⎟ ⎝ ⎠
2
⎛ 2.1 ⎞ 1 1 1 ⎛1 1 ⎞ ⎜ 3.4 ⎟ X ′Y = ⎜ 0 30 60 90 120 ⎟ ⎜ 4.2 ⎟ ⎜ 0 900 3600 8100 14400 ⎟ ⎜ 7.1 ⎟ ⎝ ⎠⎜ ⎟ ⎝ 6.3 ⎠ ⎛ 23.1 ⎞ = ⎜ 1749 ⎟ ⎜166410 ⎟ ⎝ ⎠
c. Persamaan Regresi dalam bentuk Matriks: Y = Xb + ε b = ( X ′X ) −1 X ′Y
300 2700 ⎡5 ⎤ ( X ′X ) = ⎢300 27000 2700000 ⎥ ⎢⎣27000 2700000 286740000⎥⎦
−1
−1
d. Untuk mendapatkan nilai invers X’X , digunakan cara kofaktor dan determinan: 1. Minor:
M 11 =
27000
2700000
2700000 286740000
M 12 =
= 451980000000
M 21 =
300
27000
2700000 286740000
300
27000
27000 2700000
2700000
27000 286740000
M 13 =
= 13122000000
M 22 =
= 13122000000
M 31 =
300
5
27000
27000 286740000
= 81000000
5
27000
300 2700000
27000
27000 2700000
= 81000000
M 23 =
= 704700000
M 32 =
300
5
300
27000 2700000
= 5400000
M 33 =
= 5400000
5
300
300 27000
= 45000
2. Determinan: Diambil dari baris 1: Misalkan Matriks X’X = Matriks A, Maka nilai deteminannya:
A = a11M 11 − a12 M 12 + a13 M 13 atau A = a11K11 + a12 K12 + a13 K13
dimana K ij = (−1)i = j M ij
= 5(451980000000) + 300(-13122000000) + 27000(81000000) = 510300000000 3. Matriks Kofaktor: K ij = (−1)i = j M ij
3
⎛ 451980000000 - 13122000000 81000000 ⎞ ⎜ ⎟ - 5400000 ⎟ K ij = ⎜ - 13122000000 704700000 ⎜ 81000000 ⎟ - 5400000 45000 ⎝ ⎠
4. Adjoint A:
[ ]
adj. A = K ij
⎛ 451980000000 - 13122000000 81000000 ⎞ - 5400000 ⎟ = ⎜ - 13122000000 704700000 ⎜ 81000000 ⎟ - 5400000 45000 ⎝ ⎠ ⎛ 451980000000 - 13122000000 81000000 ⎞ - 5400000 ⎟ = ⎜ - 13122000000 704700000 ⎜ 81000000 ⎟ - 5400000 45000 ⎝ ⎠
5. Invers: adj. A A−1 = ; A
'
A ≠0
⎛ 451980000000 - 13122000000 81000000 ⎞ ⎜ - 13122000000 704700000 - 5400000 ⎟ ⎟ ⎜ 81000000 5400000 45000 ⎠ =⎝ 510300000000 ⎛ 0.8857142857 - 0.0257142857 0.0001587302 ⎞ = ⎜ - 0.0257142857 0.0013809524 - 0.0000105820 ⎟ ⎜ 0.0001587302 - 0.0000105820 0.0000000882 ⎟ ⎠ ⎝ 6. Dengan demikian: ⎛ 0.8857142857 - 0.0257142857 0.0001587302 ⎞ ⎜ ⎟ ( X ' X ) = ⎜ - 0.0257142857 0.0013809524 - 0.0000105820 ⎟ ⎜ 0.0001587302 - 0.0000105820 0.0000000882 ⎟ ⎝ ⎠ −1
Sehingga, didapat nilai b:
b = ( X ′X ) −1 X ′Y ⎛ b 0 ⎞ ⎛ 0.8857142857 - 0.0257142857 0.0001587302 ⎞ ⎛ 23.1 ⎞ ⎜ b ⎟ = ⎜ - 0.0257142857 0.0013809524 - 0.0000105820 ⎟ ⎜ 1749 ⎟ ⎜⎜ 1 ⎟⎟ ⎜ ⎟⎜ ⎟ ⎝ b 2 ⎠ ⎝ 0.0001587302 - 0.0000105820 0.0000000882 ⎠ ⎝166410 ⎠ 1.9 ⎛ ⎞ = ⎜ 0.0603333 ⎟ ⎜ - 0.000167 ⎟ ⎝ ⎠ e. Persamaan Regresinya:
Yˆ = b0 + b1 X + b2 X 2 Yˆ = 1.9 + 0.0603333 X − 0.000167 X 2
4