Table 4.8
Correlation matrix of the numerical independent variables on the “new” data.
Variable
LOAN MORTDUE VALUE YOJ
DEROG DELINQ CLAGE NINQ CLNO DEBTINC
LOAN
1.00
0.23
0.33
0.08
-0.02
-0.04
0.10
0.08
0.08
0.07
MORTDUE
0.23
1.00
0.85
0.01
-0.01
-0.02
0.11
0.05
0.31
0.14
VALUE
0.33
0.85
1.00
0.06
-0.04
0.04
0.13
0.04
0.33
0.12
YOJ
0.08
0.01
0.06
1.00
-0.05
0.01
0.16
-0.06
-0.01
-0.09
DEROG
-0.02
-0.01
-0.04
-0.05
1.00
0.14
-0.02
0.07
0.00
0.07
DELINQ
-0.04
-0.02
0.04
0.01
0.14
1.00
0.06
0.02
0.07
0.06
CLAGE
0.10
0.11
0.13
0.16
-0.02
0.06
1.00
-0.1
0.21
-0.07
NINQ
0.08
0.05
0.04
-0.06
0.07
0.02
-0.1
1.00
0.07
0.20
CLNO
0.08
0.31
0.33
-0.01
0.00
0.07
0.21
0.07
1.00
0.17
DEBTINC
0.07
0.14
0.12
-0.09
0.07
0.06
-0.07
0.20
0.17
1.00
80
From this correlation matrix, we see that there are no large pair-wise correlations. The
largest correlation is 0.85 between VALUE and MORTDUE. The other pair-wise
correlations are all very small and insignificant. Worrying correlations will occur when the
correlation between two variables is greater than 0.9. The variance inflation factors for
each numerical variable are given in Table 4.9.
Table 4.9
Variance inflation factors (VIF) of numerical independent variables on the
“new” data.
Variable
VIF
LOAN
1.156804
MORTDUE 3.84004
VALUE
4.112173
YOJ
1.047812
DEROG
1.03521
DELINQ
1.046096
CLAGE
1.108899
NINQ
1.063397
CLNO
1.202153
DEBTINC
1.100282
From Table 4.9, there are no large variance inflation factors. Therefore, there is no serious
problem with collinearity in the “new” data.
Outliers and influential observations in the model are now considered. A half-normal plot
of the residuals is given in Figure 4.17.
81
Fig. 4.17
Half-normal plot of residuals for the “new” data.
Figure 4.17 shows no indication of outliers. Secondly, a half-normal plot of the leverages
is given in Figure 4.18.
Fig. 4.18
Half-normal plot of leverages for the “new” data.
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0
1
2
3
4
Half-normal quantiles
So
rte
d
D
at
a
240
245
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0.
0
0.
1
0.
2
0.
3
0.
4
0.
5
Half-normal quantiles
So
rte
d
Da
ta
49
86
82
Figure 4.18 shows that there may be some indication of leverage from observations
numbered 49 and 86. Finally a half-normal plot of the Cook’s distance statistics is given in
Figure 4.19.
Fig. 4.19
Half-normal plot of the Cook’s distance statistics for the “new” data.
From Figure 4.19, there may be some leverage from observations numbered 220 and 245.
Thus, observations numbered 49, 86, 220 and 245 may be influential observations. In order
to see whether these observations are influential we delete them from the “new” data and
re-fit the model. The estimated parameters of this model are then compared to the
parameters of the model with no observations deleted. Table 4.10 shows the estimated
coefficients.
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0.
00
0.
02
0.
04
0.
06
0.
08
0.
10
Half-normal quantiles
So
rte
d
Da
ta
220
245
83
Do'stlaringiz bilan baham: |