Can anyone help explain to me why the two codes below have different result? I
thought I can use log(time)~. to replace log(time)~dist+climb+timef.I am using
CVlm from DAAG package. I think nihills is preloaded with the package. Thanks in
advance.> CVlm(df=nihills,
form.lm=formula(log(time)~.),plotit="Observed",m=2)Analysis of
Variance Table
Response: log(time) Df Sum Sq Mean Sq F value Pr(>F) dist
1 6.34 6.34 384.31 4.6e-14 ***climb 1 0.12 0.12 7.24 0.0145
* timef 1 0.19 0.19 11.29 0.0033 ** Residuals 19 0.31 0.02
---Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
fold 1 Observations in test set: 11 Slieve Gullion McVeigh Classic
Tollymore Mountain Moughanmore Hen & Cock Annalong Horseshoe Rocky Meelbeg
Meelmore Slieve Donard Seven Sevens Slieve GallionPredicted -0.680
-0.538 -0.610 -0.786 -0.849 7.87e-01 -0.682
-6.20e-01 -2.98e-01 1.41e+00 -5.87e-01cvpred -0.762
-0.614 -0.727 -0.769 -0.799 6.67e-01 -0.648
-7.89e-01 -5.30e-02 1.36e+00 -4.52e-01log(time) -0.762
-0.614 -0.727 -0.769 -0.799 6.67e-01 -0.648
-7.89e-01 -5.30e-02 1.36e+00 -4.52e-01CV residual 0.000
0.000 0.000 0.000 0.000 1.11e-16 0.000
1.11e-16 -4.86e-17 4.44e-16 -5.55e-17
Sum of squares = 0 Mean square = 0 n = 11
fold 2 Observations in test set: 12 Binevenagh Glenariff Mountain
Donard & Commedagh Slieve Martin Monument Race Loughshannagh Horseshoe
Donard Forest Flagstaff to Carling Slieve Bearnagh Lurig Challenge Scrabo Hill
Race BARF Turkey TrotPredicted -1.78e-01 -4.89e-01
2.78e-02 -0.577 -7.27e-01 -0.623 -0.571
3.03e-01 -0.401 -7.20e-01 -0.889 -4.88e-01cvpred
-1.53e-01 -3.52e-01 3.79e-02 -0.597 -7.51e-01
-0.435 -0.657 3.76e-01 -0.374 -8.33e-01
-1.125 -3.38e-01log(time) -1.53e-01 -3.52e-01
3.79e-02 -0.597 -7.51e-01 -0.435 -0.657
3.76e-01 -0.374 -8.33e-01 -1.125 -3.38e-01CV
residual -2.78e-17 -5.55e-17 -6.94e-18 0.000
1.11e-16 0.000 0.000 -1.11e-16
0.000 1.11e-16 0.000 -5.55e-17
Sum of squares = 0 Mean square = 0 n = 12
Overall (Sum over all 12 folds) ms 1.18e-32 Warning messages:1: In
predict.lm(subs.lm, newdata = df[rows.out, ]) : prediction from a
rank-deficient fit may be misleading2: In predict.lm(subs.lm, newdata =
df[rows.out, ]) : prediction from a rank-deficient fit may be misleading3: In
CVlm(df = nihills, form.lm = formula(log(time) ~ .), plotit =
"Observed", :
As there is >1 explanatory variable, cross-validation predicted values for a
fold are not a linear function of corresponding overall predicted values. Lines
that are shown for the different folds are approximate> CVlm(df=nihills,
form.lm=formula(log(time)~dist+climb+timef),plotit="Observed",m=2)Analysis
of Variance Table
Response: log(time) Df Sum Sq Mean Sq F value Pr(>F) dist
1 6.34 6.34 384.31 4.6e-14 ***climb 1 0.12 0.12 7.24 0.0145
* timef 1 0.19 0.19 11.29 0.0033 ** Residuals 19 0.31 0.02
---Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
fold 1 Observations in test set: 11 Slieve Gullion McVeigh Classic
Tollymore Mountain Moughanmore Hen & Cock Annalong Horseshoe Rocky Meelbeg
Meelmore Slieve Donard Seven Sevens Slieve GallionPredicted -0.6801
-0.53759 -0.610 -0.7857 -0.84929 0.787 -0.6824
-0.620 -0.298 1.41 -0.587cvpred -0.7068
-0.60517 -0.680 -0.7501 -0.80507 1.053 -0.6856
-0.642 -0.185 2.81 -0.588log(time) -0.7621
-0.61413 -0.727 -0.7687 -0.79913 0.667 -0.6481
-0.789 -0.053 1.36 -0.452CV residual -0.0554
-0.00896 -0.047 -0.0186 0.00595 -0.386 0.0376
-0.147 0.132 -1.45 0.136
Sum of squares = 2.31 Mean square = 0.21 n = 11
fold 2 Observations in test set: 12 Binevenagh Glenariff Mountain
Donard & Commedagh Slieve Martin Monument Race Loughshannagh Horseshoe
Donard Forest Flagstaff to Carling Slieve Bearnagh Lurig Challenge Scrabo Hill
Race BARF Turkey TrotPredicted -0.178 -0.489
0.0278 -0.5771 -0.72713 -0.623 -0.571
0.303 -0.4008 -0.720 -0.889 -0.488cvpred
-0.308 -0.583 0.0268 -0.5822 -0.75753
-0.614 -0.600 0.125 -0.3245 -0.751
-0.891 -0.558log(time) -0.153 -0.352
0.0379 -0.5968 -0.75148 -0.435 -0.657
0.376 -0.3743 -0.833 -1.125 -0.338CV
residual 0.156 0.231 0.0111 -0.0147
0.00604 0.178 -0.057 0.251
-0.0498 -0.082 -0.234 0.219
Sum of squares = 0.29 Mean square = 0.02 n = 12
Overall (Sum over all 12 folds) ms 0.113 Warning message:In CVlm(df =
nihills, form.lm = formula(log(time) ~ dist + climb + :
As there is >1 explanatory variable, cross-validation predicted values for a
fold are not a linear function of corresponding overall predicted values. Lines
that are shown for the different folds are approximate> head(nihills)
dist climb time timefBinevenagh 7.5 1740 0.858
1.064Slieve Gullion 4.2 1110 0.467 0.623Glenariff Mountain 5.9 1210
0.703 0.887Donard & Commedagh 6.8 3300 1.039 1.214McVeigh Classic 5.0
1200 0.541 0.637Tollymore Mountain 4.8 950 0.483 0.589
[[alternative HTML version deleted]]
