Regression Adequacy
Regression Adequacy
Regression Adequacy
05
Adequacy of Models for Regression
06.05.1
06.05.2 Chapter 06.05
Figure 1 Plot of coefficient of thermal expansion vs. temperature data points and regression
line.
Table 2 shows the residuals of the data to calculate the sum of the square of residuals as
S r ( 0.28571) 2 (0.068571) 2 (0.23286) 2 (0.21714) 2
(0.021429) 2 ( 0.25429) 2
0.25283
The standard error of estimate
Sr
s / T
n2
0.25283
62
0.25141
The units of s / T are same as the units of . How is the value of the standard error of
estimate interpreted? We may say that on average the difference between the observed and
predicted values is 0.25141 μin/in/ F . Also, we can look at the value as follows. About
95% of the observed values are between 2 s / T of the predicted value (see Figure 2).
This would lead us to believe that the value of in the example is expected to be accurate
within 2 s / T = 2 0.25141 = 0.50282 μin/in/ F .
Figure 2 Plotting the linear regression line and showing the regression standard error.
One can also look at this criterion as finding if 95% of the scaled residuals for the model are
in the domain [-2,2], that is
06.05.4 Chapter 06.05
i a0 a1Ti
Scaled residual
s / T
For the example,
s / T 0.25141
Table 4 Residuals and scaled residuals for data.
Ti i i a 0 a1Ti Scaled Residuals
-340 2.45 -0.28571 -1.1364
-260 3.58 0.068571 0.27275
-180 4.52 0.23286 0.92622
-100 5.28 0.21714 0.86369
-20 5.86 0.021429 0.085235
60 6.36 -0.25429 -1.0115
and the scaled residuals are calculated in Table 4. All the scaled residuals are in the [-2,2]
domain.
i 1
n
(1)
i a 0 a1Ti
2
i 1
n
S t i
2
(2)
i 1
where
n
i
i 1
n
For the example data
6
i
i 1
6
2.45 3.58 4.52 5.28 5.86 6.36
6
4.6750 μin/in/ F
n
S t i
2
i 1
Adequacy of Regression Model 06.05.5
Going back to the definition of the coefficient of determination, one can see that S t is the
variation without any relationship of y vs. x , while S r is the variation with the straight-
line relationship.
The limits of the values of r 2 are between 0 and 1. What do these limiting values of
r 2 mean? If r 2 0 , then S t S r , which means that regressing the data to a straight line
does nothing to explain the data any further. If r 2 1 , then S r 0 , which means that the
straight line is passing through all the data points and is a perfect fit.
Instantaneous
Temperature
Thermal Expansion
F μin/in/F
80 6.47
60 6.36
40 6.24
20 6.12
0 6.00
-20 5.86
-40 5.72
-60 5.58
-80 5.43
-100 5.28
-120 5.09
-140 4.91
-160 4.72
-180 4.52
-200 4.30
-220 4.08
-240 3.83
-260 3.58
-280 3.33
-300 3.07
-320 2.76
-340 2.45
06.05.8 Chapter 06.05
Figure 3 Plot of thermal expansion coefficient vs. temperature data points and regression line
for more data points.
Regressing the data from Table 2 to the straight line regression line
(T ) a 0 a1T
and following the procedure for conducting linear regression as given in Chapter 06.03, we
get (Figure 3)
6.0248 0.0093868T
Adequacy of Regression Model 06.05.9
References