Selasa, 06 Desember 2016

Tugas Temu 9

Halaman 153-154

Latihan 1

Lakukan prediksi CHOL dengan variabel independen TRIG, UM, dan UM kuadrat.
a.       Hitung SS for Regression ;
b.      Hitung SS for Residual;
c.       Hitung Means SS for Regression ;
d.      Hitung Means SS for Residual;
e.       Hitung nilai F parsial;
f.       Hitung nilai r2;
g.      Buktikan bahwa penambahan X3 berperan dalam memprediksi Y. 

UM
CHOL
TRIG
40
218
194
46
165
188
69
197
134
44
188
155
41
217
191
56
240
207
48
222
155
49
244
235
41
190
167
38
209
186
36
208
179
39
214
129
59
238
220
56
219
155
44
241
201
37
212
140
40
244
132
32
217
140
56
227
279
49
218
101
50
241
213
46
234
168
52
231
242
51
297
142
46
230
240
60
258
173
47
243
175
58
236
199
66
193
201
52
193
193
55
319
191
58
212
216
41
209
154
60
224
198
50
284
129
48
222
115
49
229
148
39
204
164
40
211
104
47
230
218
67
230
239
57
222
183
50
213
190
43
238
259
55
234
156
UM      = Umur
CHOL = Cholesterol
TRIG   = Trigliserida
Model 1 : CHOL = β0 + β1 TRIG
Coefficientsa
Model
Unstandardized Coefficients
Standardized Coefficients
t
Sig.
B
Std. Error
Beta
1
(Constant)
203.123
17.156

11.840
.000
Trigliserida
.127
.093
.203
1.360
.181
a. Dependent Variable: Cholesterol
Estimated model 1: CHOL = 203.123 + 0.127 TRIG
ANOVAa
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
1181.676
1
1181.676
1.850
.181b
Residual
27464.768
43
638.716


Total
28646.444
44



a. Dependent Variable: Cholesterol
b. Predictors: (Constant), Trigliserida
Model 2 : CHOL = β0 + β1 UM
Coefficientsa
Model
Unstandardized Coefficients
Standardized Coefficients
t
Sig.
B
Std. Error
Beta
1
(Constant)
204.048
22.093

9.236
.000
Umur
.445
.444
.151
1.004
.321
a. Dependent Variable: Cholesterol
Estimated model 2: CHOL = 204.048 + 0.445 UM
ANOVAa
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
655.625
1
655.625
1.007
.321b
Residual
27990.819
43
650.949


Total
28646.444
44



a. Dependent Variable: Cholesterol
b. Predictors: (Constant), Umur
Model 3 : CHOL = β0 + β1 UMSQ
Coefficientsa
Model
Unstandardized Coefficients
Standardized Coefficients
t
Sig.
B
Std. Error
Beta
1
(Constant)
217.420
11.555

18.816
.000
Umur Kuadrat
.003
.004
.118
.777
.442
a. Dependent Variable: Cholesterol
Estimated model : CHOL = 217.420 + 0.003 UMSQ
ANOVAa
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
396.227
1
396.227
.603
.442b
Residual
28250.217
43
656.982


Total
28646.444
44



a. Dependent Variable: Cholesterol
b. Predictors: (Constant), Umur Kuadrat
Model 4 : CHOL = β0 + β1 TRIG + β2 UM
Coefficientsa
Model
Unstandardized Coefficients
Standardized Coefficients
t
Sig.
B
Std. Error
Beta
1
(Constant)
192.155
24.554

7.826
.000
Trigliserida
.108
.098
.173
1.099
.278
Umur
.292
.464
.099
.629
.533
a. Dependent Variable: Cholesterol
Estimated model : CHOL = 192.155 + 0.108 TRIG + 0.292 UM
ANOVAa
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
1437.719
2
718.860
1.110
.339b
Residual
27208.725
42
647.827


Total
28646.444
44



a. Dependent Variable: Cholesterol
b. Predictors: (Constant), Umur, Trigliserida
Model 5 : CHOL = β0 + β1 TRIG + β3 UMSQ
Coefficientsa
Model
Unstandardized Coefficients
Standardized Coefficients
t
Sig.
B
Std. Error
Beta
1
(Constant)
200.525
18.433

10.879
.000
Trigliserida
.115
.098
.185
1.173
.247
Umur Kuadrat
.002
.005
.065
.413
.682
a. Dependent Variable: Cholesterol
Estimated model : CHOL = 200.525 + 0.115 TRIG + 0.002 UMSQ
ANOVAa
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
1292.618
2
646.309
.992
.379b
Residual
27353.826
42
651.282


Total
28646.444
44



a. Dependent Variable: Cholesterol
b. Predictors: (Constant), Umur Kuadrat, Trigliserida
Model 6 : CHOL = β0 + β1 TRIG + β2 UMSQ + β3 UMSQ
Coefficientsa
Model
Unstandardized Coefficients
Standardized Coefficients
t
Sig.
B
Std. Error
Beta
1
(Constant)
-21.969
104.532

-.210
.835
Trigliserida
.079
.095
.126
.825
.414
Umur
9.220
4.269
3.132
2.160
.037
Umur Kuadrat
-.088
.042
-3.035
-2.103
.042
a. Dependent Variable: Cholesterol
Estimated model : CHOL = -21.969 + 0.079 TRIG + 9.220 UM - 0.088 UMSQ
ANOVAa
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
4086.344
3
1362.115
2.274
.094b
Residual
24560.100
41
599.027


Total
28646.444
44



a. Dependent Variable: Cholesterol
b. Predictors: (Constant), Umur Kuadrat, Trigliserida, Umur
Kita lakukan uji parsial F seperti berikut (bedasarkan hasil-hasil yang sudah kita lakukan diatas)

ANOVA Tabel untuk CHOL dengan TRIG, UM, UMSQ

Sumber

Df

SS

MS

F

r2
Regresi
X1
X2|X1
X3|X1, X2

1
1
1

1181.676
256.043
2648.625

1181.676
256.043
2648.625

1.972*
0.427
4.421

0.143
Residual
41
24560.100
599.027


Total
44
28646.444



*p<0.05

No.

Model Estimasi

F

r2
1.
Y = 203.123 + 0.127 TRIG
                         (0.093)*
1.850
0.041
2.
Y = 204.048 + 0.445 UM
                         (0.044)*
1.007
0.023
3
Y = 217.420 + 0.003 UMSQ
                         (0.004)*
0.603
0.014
4.
Y = 192.155 + 0.108 TRIG + 0.292 UM
                         (0.098)*         (0.464)*
1.110
0.050
5.
Y = 200.525 + 0.115 TRIG + 0.002 UMSQ
                         (0.098)*          (0.005)*
0.992
0.045
6.
Y = -21.969 + 0.079 TRIG + 9.220 UM - 0.088 UMSQ
                            (0.095)*          (4.269)*         (0.042)
2.274
0.143
Angka dalam tanda kurung adalah Standar Error dari parameter
*bermakna (p<0.05)

            Dari ke enam model estimasi terlihat bahwa variable TRIG secara konsisten sangat berpengaruh terhadap CHOL (p<0.05). Pada model estimasi 1 tampak nilai r2 sebesar 0.041 dan bila dibanding dengan model estimasi 4,5 dan 6 penambahan nilai r2 relatif kecil masing-masing 0.050, 0.045, dan 0.143 atau hanya bertambah sekitar 0.009, 0.004, dan 0.102 ini sangat tidak berarti.
Dengan demikian kita bisa berkesimpulan bahwa variable TRIG sangat bermakna pengeruhnya terhadap CHOL. Sebaliknya penambahan variable  UM dan UMSQ tidak berperan dalam menjelaskan variasi CHOL dan tidak perlu menambahkan kedua variable tersebut ke dalam model. Model akhir yaitu : Y = 203.123 + 0.127 TRIG
 
LLatihan 2
Lakukan prediksi BB dengan variabel independen TB, BTL, dan AK.
a.       Hitung SS for Regression ;
b.      Hitung SS for Residual;
c.       Hitung Means SS for Regression ;
d.      Hitung Means SS for Residual;
e.       Hitung nilai F parsial;
f.       Hitung nilai r2;
g.      Buktikan bahwa penambahan X3 berperan dalam memprediksi Y.
BB
TB
BTL
AK
79.2
149.0
54.1
2670.0
64.0
152.0
44.3
820.0
67.0
155.7
47.8
1210.0
78.4
159.0
53.9
2678.0
66.0
163.3
47.5
1205.0
63.0
166.0
43.0
815.0
65.9
169.0
47.1
1200.0
63.1
172.0
44.0
1180.0
73.2
174.5
44.1
1850.0
66.5
176.1
48.3
1260.0
61.9
176.5
43.5
1170.0
72.5
179.0
43.3
1852.0
101.1
182.0
66.4
1790.0
66.2
170.4
47.5
1250.0
99.9
184.9
66.0
1889.0
63.0
169.0
44.0
915.0
 



















BB       = Berat Badan
TB       = Tinggi Badan
BTL     = Berat Badan Tanpa Lemak
AK      = Asupan Kalori

Model 1 : BB = β0 + β1 TB
Coefficientsa
Model
Unstandardized Coefficients
Standardized Coefficients
T
Sig.
B
Std. Error
Beta
1
(Constant)
-2.492
48.880

-.051
.960
Tinggi Badan
.441
.289
.378
1.525
.149
a. Dependent Variable: Berat Badan

Estimated model 1 : BB = -2.492 + 0.441 TB
ANOVAa
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
326.204
1
326.204
2.327
.149b
Residual
1962.751
14
140.196


Total
2288.954
15



a. Dependent Variable: Berat Badan
b. Predictors: (Constant), Tinggi Badan

Model 2 : BB = β0 + β2 BTL
Coefficientsa
Model
Unstandardized Coefficients
Standardized Coefficients
t
Sig.
B
Std. Error
Beta
1
(Constant)
-4.303
7.112

-.605
.555
Berat Badan Tanpa Lemak
1.554
.143
.945
10.836
.000
a. Dependent Variable: Berat Badan

Estimated model 2 : BB = -4.303 + 1.554 BTL
ANOVAa
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
2045.099
1
2045.099
117.411
.000b
Residual
243.855
14
17.418


Total
2288.954
15



a. Dependent Variable: Berat Badan
b. Predictors: (Constant), Berat Badan Tanpa Lemak
 
Model 3 : BB = β0 + β3 AK
Coefficientsa
Model
Unstandardized Coefficients
Standardized Coefficients
t
Sig.
B
Std. Error
Beta
1
(Constant)
52.517
7.074

7.423
.000
Asupan Kalori
.013
.004
.617
2.936
.011
a.     Dependent Variable: Berat Badan
 
Estimasi model 3 : BB = 52.517 + 0.013 AK
ANOVAa
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
872.301
1
872.301
8.620
.011b
Residual
1416.653
14
101.190


Total
2288.954
15



a. Dependent Variable: Berat Badan
b. Predictors: (Constant), Asupan Kalori
 
Model 4 : BB = β0 + β1 TB + β2 BTL
Coefficientsa
Model
Unstandardized Coefficients
Standardized Coefficients
t
Sig.
B
Std. Error
Beta
1
(Constant)
-27.527
16.631

-1.655
.122
Tinggi Badan
.155
.101
.132
1.530
.150
Berat Badan Tanpa Lemak
1.496
.142
.910
10.511
.000
a. Dependent Variable: Berat Badan
 
Estimasi model 4 : BB = -27.527 + 0.155 TB + 1.496 BTL
ANOVAa
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
2082.309
2
1041.154
65.499
.000b
Residual
206.645
13
15.896


Total
2288.954
15



a. Dependent Variable: Berat Badan
b. Predictors: (Constant), Berat Badan Tanpa Lemak, Tinggi Badan
 
Model 5 : BB = β0 + β1 TB + β3 AK
Coefficientsa
Model
Unstandardized Coefficients
Standardized Coefficients
T
Sig.
B
Std. Error
Beta
1
(Constant)
-31.333
37.369

-.838
.417
Tinggi Badan
.492
.216
.421
2.275
.040
Asupan Kalori
.014
.004
.646
3.491
.004
a. Dependent Variable: Berat Badan
 
Estimasi model 5 : BB = -31.333 + 0.492 TB + 0.014 AK
ANOVAa
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
1275.821
2
637.911
8.185
.005b
Residual
1013.133
13
77.933


Total
2288.954
15



a. Dependent Variable: Berat Badan
b. Predictors: (Constant), Asupan Kalori, Tinggi Badan
 
Model 6 : BB = β0 + β1 TB + β2 BTL + β3 AK
Coefficientsa
Model
Unstandardized Coefficients
Standardized Coefficients
T
Sig.
B
Std. Error
Beta
1
(Constant)
-33.412
14.489

-2.306
.040
Tinggi Badan
.210
.090
.180
2.339
.037
Berat Badan Tanpa Lemak
1.291
.150
.785
8.631
.000
Asupan Kalori
.004
.002
.209
2.375
.035
a. Dependent Variable: Berat Badan
 
Estimasi model 6 : BB = -33.412 + 0.210 TB + 1.291 BTL + 0.004 AK
ANOVAa
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
2148.400
3
716.133
61.141
.000b
Residual
140.554
12
11.713


Total
2288.954
15



a. Dependent Variable: Berat Badan
b. Predictors: (Constant), Asupan Kalori, Tinggi Badan, Berat Badan Tanpa Lemak
 
Kita lakukan uji parsial F seperti berikut (bedasarkan hasil-hasil yang sudah kita lakukan diatas)

ANOVA Tabel untuk CHOL dengan TRIG, UM, UMSQ

Sumber

Df

SS

MS

F

r2
Regresi
X1
X2|X1
X3|X1, X2

1
1
1

326.204
1756.105
66.091

326.204
1756.105
66.091

27.850
149.928
5.643

0.939
Residual
12
140.554
11.713


Total
15
2288.954



*p<0.05

No.

Model Estimasi

F

r2
1.
Y = -2.492 + 0.441 TB
                         (0.289)*
2.327
0.143
2.
Y = -4.303 + 1.554 BTL
                         (0.143)*
117.411
0.893
3
Y = 52.517 + 0.013 AK
                         (0.004)*
8.620
0.381
4.
Y = -27.527 + 0.155 TB + 1.496 BTL
                         (0.101)*         (0.142)*
65.499
0.910
5.
Y = -31.333 + 0.492 TB + 0.014 AK
                         (0.216)*          (0.004)*
8.185
0.557
6.
Y = -33.412 + 0.210 TB + 1.291 BTL + 0.004 AK                           
                       (0.090)*          (0.150)*         (0.002)*
61.141
0.939
Angka dalam tanda kurung adalah Standar Error dari parameter
*bermakna (p<0.05)

            Dari ke enam model estimasi terlihat bahwa variable TB secara konsisten sangat berpengaruh terhadap BB (p<0.05). Pada model estimasi 1tampak nilai r2 sebesar 0.143 dan bila dibanding dengan model estimasi 4,5 dan 6 penambahan nilai r2 relatif kecil dengan nilai 0.910, 0.557, 0.939 atau hanya bertambah sekitar 0.767, 0.414 dan 0.792 nilai ini tidak berarti.

            Dengan demikian bisa diambil kesimpulan bahwa variable TB sangat bermakna pengaruhnya terhadap BB. Sebaliknya penambahan variable BBTL dan AK tidak berperan dalam menjelaskan variasi BB dan tidak perlu menambahkan kedua variable tersebut ke dalam model. Model akhir yaitu : Y= -2.492 + 0.441 TB
   
 
 
 

Tidak ada komentar:

Posting Komentar