aicc.am2
function
5
x1
missing
0
x2
missing
0
y
missing
0
start
call
3

name
1
c

call
3

name
1
/

name
1
n

numeric
1
2

call
3

name
1
/

name
1
n

numeric
1
2

{
4

comment.expression
2

comment
1

character
13
#
# Function to determine the smoothing parameters based on AICC for
# an additive model for y based on two variables x1 and x2, using 
# cubic smoothing splines for the smooths. For discussion see
# Simonoff and Tsai (1999, Journal of Computational & Graphical 
# Statistics).
#
# The function can be generalized to more predictors by adding the
# appropriate terms to those for x1 and x2.
#
# The function should be run using several different starting values
# to assure that the true minimum of AICC has been found.
#

<-
2

name
1
n

call
2

name
1
length

name
1
y

<-
2

name
1
df

call
8

name
1
nlminb
start
name
1
start
obj
name
1
crit.am2
lower
call
3

name
1
c

numeric
1
1

numeric
1
1
upper
call
3

name
1
c

call
3

name
1
-

name
1
n

numeric
1
2

call
3

name
1
-

name
1
n

numeric
1
2
x1
name
1
x1
x2
name
1
x2
y
name
1
y

<-
2

name
1
am

call
2

name
1
gam

call
3

name
1
~

name
1
y

call
3

name
1
+

call
3

name
1
s

name
1
x1
df
call
3

name
1
[

call
3

name
1
$

name
1
df

character
1
parameters

numeric
1
1

call
3

name
1
s

name
1
x2
df
call
3

name
1
[

call
3

name
1
$

name
1
df

character
1
parameters

numeric
1
2

call
4

name
1
list
am
name
1
am
aicc
call
3

name
1
$

name
1
df

character
1
objective
df
call
3

name
1
$

name
1
df

character
1
parameters
aicc.am3
function
6
x1
missing
0
x2
missing
0
x3
missing
0
y
missing
0
start
call
4

name
1
c

call
3

name
1
/

name
1
n

numeric
1
2

call
3

name
1
/

name
1
n

numeric
1
2

call
3

name
1
/

name
1
n

numeric
1
2

{
4

comment.expression
2

comment
1

character
13
#
# Function to determine the smoothing parameters based on AICC for
# an additive model for y based on three variables x1, x2 and x3, using 
# cubic smoothing splines for the smooths. For discussion see
# Simonoff and Tsai (1999, Journal of Computational & Graphical 
# Statistics).
#
# The function can be generalized to more predictors by adding the
# appropriate terms to those for x1, x2 and x3.
#
# The function should be run using several different starting values
# to assure that the true minimum of AICC has been found.
#

<-
2

name
1
n

call
2

name
1
length

name
1
y

<-
2

name
1
df

call
9

name
1
nlminb
start
name
1
start
obj
name
1
crit.am3
lower
call
4

name
1
c

numeric
1
1

numeric
1
1

numeric
1
1
upper
call
4

name
1
c

call
3

name
1
-

name
1
n

numeric
1
2

call
3

name
1
-

name
1
n

numeric
1
2

call
3

name
1
-

name
1
n

numeric
1
2
x1
name
1
x1
x2
name
1
x2
x3
name
1
x3
y
name
1
y

<-
2

name
1
am

call
2

name
1
gam

call
3

name
1
~

name
1
y

call
3

name
1
+

call
3

name
1
+

call
3

name
1
s

name
1
x1
df
call
3

name
1
[

call
3

name
1
$

name
1
df

character
1
parameters

numeric
1
1

call
3

name
1
s

name
1
x2
df
call
3

name
1
[

call
3

name
1
$

name
1
df

character
1
parameters

numeric
1
2

call
3

name
1
s

name
1
x3
df
call
3

name
1
[

call
3

name
1
$

name
1
df

character
1
parameters

numeric
1
3

call
4

name
1
list
am
name
1
am
aicc
call
3

name
1
$

name
1
df

character
1
objective
df
call
3

name
1
$

name
1
df

character
1
parameters
aicc.spline
function
4
x
missing
0
y
missing
0
start
call
3

name
1
/

name
1
n

numeric
1
2

{
4

comment.expression
2

comment
1

character
8
#
# Function to determine the AICC-based smoothing parameter for a cubic
# smoothing spline fit of y on x. See Hurvich, Simonoff, and Tsai (1998,
# JRSSB) for discussion.
#
# The function should be run using several different starting values
# to assure that the true minimum of AICC has been found.
#

<-
2

name
1
n

call
2

name
1
length

name
1
y

<-
2

name
1
df

call
7

name
1
nlminb
start
name
1
start
obj
name
1
crit.spline
lower
numeric
1
2
upper
call
3

name
1
-

name
1
n

numeric
1
2
x
name
1
x
y
name
1
y

<-
2

name
1
s

call
4

name
1
smooth.spline

name
1
x

name
1
y
df
call
3

name
1
$

name
1
df

character
1
parameters

call
3

name
1
list
spline
name
1
s
aicc
call
3

name
1
$

name
1
df

character
1
objective
crit.am2
function
5
df
missing
0
x1
missing
0
x2
missing
0
y
missing
0

{
3

comment.expression
2

comment
1

character
3
#
# AICC criterion for additive model with two predictors
#

<-
2

name
1
n

call
2

name
1
length

name
1
y

<-
2

name
1
ggg

call
2

name
1
gam

call
3

name
1
~

name
1
y

call
3

name
1
+

call
3

name
1
s

name
1
x1
df
call
3

name
1
[

name
1
df

numeric
1
1

call
3

name
1
s

name
1
x2
df
call
3

name
1
[

name
1
df

numeric
1
2

call
3

name
1
+

call
2

name
1
log

call
3

name
1
/

call
3

name
1
$

name
1
ggg

character
1
deviance

name
1
n

(
2

name
1
(

call
3

name
1
/

(
2

name
1
(

call
3

name
1
-

call
3

name
1
*

numeric
1
2

name
1
n

call
3

name
1
$

name
1
ggg

character
1
df

(
2

name
1
(

call
3

name
1
-

call
3

name
1
$

name
1
ggg

character
1
df

numeric
1
2
crit.am3
function
6
df
missing
0
x1
missing
0
x2
missing
0
x3
missing
0
y
missing
0

{
3

comment.expression
2

comment
1

character
3
#
# AICC criterion for additive model with three predictors
#

<-
2

name
1
n

call
2

name
1
length

name
1
y

<-
2

name
1
ggg

call
2

name
1
gam

call
3

name
1
~

name
1
y

call
3

name
1
+

call
3

name
1
+

call
3

name
1
s

name
1
x1
df
call
3

name
1
[

name
1
df

numeric
1
1

call
3

name
1
s

name
1
x2
df
call
3

name
1
[

name
1
df

numeric
1
2

call
3

name
1
s

name
1
x3
df
call
3

name
1
[

name
1
df

numeric
1
3

call
3

name
1
+

call
2

name
1
log

call
3

name
1
/

call
3

name
1
$

name
1
ggg

character
1
deviance

name
1
n

(
2

name
1
(

call
3

name
1
/

(
2

name
1
(

call
3

name
1
-

call
3

name
1
*

numeric
1
2

name
1
n

call
3

name
1
$

name
1
ggg

character
1
df

(
2

name
1
(

call
3

name
1
-

call
3

name
1
$

name
1
ggg

character
1
df

numeric
1
2
crit.expo1
function
3
x
missing
0
y
missing
0

{
7

comment.expression
2

comment
1

character
8
#
# Function to calculate the AICC value for an exponetial model with one predictor.
# The target variable is y and the predictor is x. For this function, p must be
# the number of parameters (2), since X-tilde is of full rank; the formula is
# given to illustrate the general formulation. For discussion see
# Simonoff and Tsai (1999, Journal of Computational & Graphical 
# Statistics).
#

<-
2

name
1
n

call
2

name
1
length

name
1
y

<-
2

name
1
logy

call
2

name
1
log

name
1
y

<-
2

name
1
startb

call
2

name
1
coefficients

call
2

name
1
lm

call
3

name
1
~

name
1
logy

name
1
x

<-
2

name
1
nlsmodel

call
4

name
1
nls

call
3

name
1
~

name
1
y

call
2

name
1
exp

call
3

name
1
+

name
1
beta0

call
3

name
1
*

name
1
beta1

name
1
x
data
call
3

name
1
data.frame

name
1
x

name
1
y
start
call
3

name
1
list
beta0
call
3

name
1
[

name
1
startb

numeric
1
1
beta1
call
3

name
1
[

name
1
startb

numeric
1
2

<-
2

name
1
xtilde

call
3

name
1
cbind

call
2

name
1
fitted

name
1
nlsmodel

call
3

name
1
*

call
2

name
1
fitted

name
1
nlsmodel

name
1
x

<-
2

name
1
p

call
2

name
1
mattrace

call
3

name
1
%*%

call
3

name
1
%*%

name
1
xtilde

call
2

name
1
solve

call
3

name
1
%*%

call
2

name
1
t

name
1
xtilde

name
1
xtilde

call
2

name
1
t

name
1
xtilde

call
3

name
1
+

call
2

name
1
log

call
3

name
1
/

call
2

name
1
sum

call
3

name
1
^

call
3

name
1
$

name
1
nlsmodel

character
1
residual

numeric
1
2

name
1
n

call
3

name
1
/

(
2

name
1
(

call
3

name
1
+

numeric
1
1

call
3

name
1
/

name
1
p

name
1
n

(
2

name
1
(

call
3

name
1
-

call
3

name
1
-

numeric
1
1

call
3

name
1
/

name
1
p

name
1
n

call
3

name
1
/

numeric
1
2

name
1
n
crit.expo2
function
4
x1
missing
0
x2
missing
0
y
missing
0

{
7

comment.expression
2

comment
1

character
8
#
# Function to calculate the AICC value for an exponetial model with two predictors.
# The target variable is y and the predictors are x1 and x2.  For this function, p 
# must be the number of parameters (3), since X-tilde is of full rank; the formula is
# given to illustrate the general formulation. For discussion see
# Simonoff and Tsai (1999, Journal of Computational & Graphical 
# Statistics).
#

<-
2

name
1
n

call
2

name
1
length

name
1
y

<-
2

name
1
logy

call
2

name
1
log

name
1
y

<-
2

name
1
startb

call
2

name
1
coefficients

call
2

name
1
lm

call
3

name
1
~

name
1
logy

call
3

name
1
+

name
1
x1

name
1
x2

<-
2

name
1
nlsmodel

call
4

name
1
nls

call
3

name
1
~

name
1
y

call
2

name
1
exp

call
3

name
1
+

call
3

name
1
+

name
1
beta0

call
3

name
1
*

name
1
beta1

name
1
x1

call
3

name
1
*

name
1
beta2

name
1
x2
data
call
4

name
1
data.frame

name
1
x1

name
1
x2

name
1
y
start
call
4

name
1
list
beta0
call
3

name
1
[

name
1
startb

numeric
1
1
beta1
call
3

name
1
[

name
1
startb

numeric
1
2
beta2
call
3

name
1
[

name
1
startb

numeric
1
3

<-
2

name
1
xtilde

call
4

name
1
cbind

call
2

name
1
fitted

name
1
nlsmodel

call
3

name
1
*

call
2

name
1
fitted

name
1
nlsmodel

name
1
x1

call
3

name
1
*

call
2

name
1
fitted

name
1
nlsmodel

name
1
x2

<-
2

name
1
p

call
2

name
1
mattrace

call
3

name
1
%*%

call
3

name
1
%*%

name
1
xtilde

call
2

name
1
solve

call
3

name
1
%*%

call
2

name
1
t

name
1
xtilde

name
1
xtilde

call
2

name
1
t

name
1
xtilde

call
3

name
1
+

call
2

name
1
log

call
3

name
1
/

call
2

name
1
sum

call
3

name
1
^

call
3

name
1
$

name
1
nlsmodel

character
1
residual

numeric
1
2

name
1
n

call
3

name
1
/

(
2

name
1
(

call
3

name
1
+

numeric
1
1

call
3

name
1
/

name
1
p

name
1
n

(
2

name
1
(

call
3

name
1
-

call
3

name
1
-

numeric
1
1

call
3

name
1
/

name
1
p

name
1
n

call
3

name
1
/

numeric
1
2

name
1
n
crit.expo3
function
5
x1
missing
0
x2
missing
0
x3
missing
0
y
missing
0

{
7

comment.expression
2

comment
1

character
8
#
# Function to calculate the AICC value for an exponetial model with three predictors.
# The target variable is y and the predictors are x1, x2, and x3.  For this function, p 
# must be the number of parameters (4), since X-tilde is of full rank; the formula is
# given to illustrate the general formulation. For discussion see
# Simonoff and Tsai (1999, Journal of Computational & Graphical 
# Statistics).
#

<-
2

name
1
n

call
2

name
1
length

name
1
y

<-
2

name
1
logy

call
2

name
1
log

name
1
y

<-
2

name
1
startb

call
2

name
1
coefficients

call
2

name
1
lm

call
3

name
1
~

name
1
logy

call
3

name
1
+

call
3

name
1
+

name
1
x1

name
1
x2

name
1
x3

<-
2

name
1
nlsmodel

call
4

name
1
nls

call
3

name
1
~

name
1
y

call
2

name
1
exp

call
3

name
1
+

call
3

name
1
+

call
3

name
1
+

name
1
beta0

call
3

name
1
*

name
1
beta1

name
1
x1

call
3

name
1
*

name
1
beta2

name
1
x2

call
3

name
1
*

name
1
beta3

name
1
x3
data
call
5

name
1
data.frame

name
1
x1

name
1
x2

name
1
x3

name
1
y
start
call
5

name
1
list
beta0
call
3

name
1
[

name
1
startb

numeric
1
1
beta1
call
3

name
1
[

name
1
startb

numeric
1
2
beta2
call
3

name
1
[

name
1
startb

numeric
1
3
beta3
call
3

name
1
[

name
1
startb

numeric
1
4

<-
2

name
1
xtilde

call
5

name
1
cbind

call
2

name
1
fitted

name
1
nlsmodel

call
3

name
1
*

call
2

name
1
fitted

name
1
nlsmodel

name
1
x1

call
3

name
1
*

call
2

name
1
fitted

name
1
nlsmodel

name
1
x2

call
3

name
1
*

call
2

name
1
fitted

name
1
nlsmodel

name
1
x3

<-
2

name
1
p

call
2

name
1
mattrace

call
3

name
1
%*%

call
3

name
1
%*%

name
1
xtilde

call
2

name
1
solve

call
3

name
1
%*%

call
2

name
1
t

name
1
xtilde

name
1
xtilde

call
2

name
1
t

name
1
xtilde

call
3

name
1
+

call
2

name
1
log

call
3

name
1
/

call
2

name
1
sum

call
3

name
1
^

call
3

name
1
$

name
1
nlsmodel

character
1
residual

numeric
1
2

name
1
n

call
3

name
1
/

(
2

name
1
(

call
3

name
1
+

numeric
1
1

call
3

name
1
/

name
1
p

name
1
n

(
2

name
1
(

call
3

name
1
-

call
3

name
1
-

numeric
1
1

call
3

name
1
/

name
1
p

name
1
n

call
3

name
1
/

numeric
1
2

name
1
n
crit.linear
function
3
x
missing
0
y
missing
0

{
4

comment.expression
2

comment
1

character
5
#
# Function to calculate the AICC value for a linear model. The target
# variable is y and the predictors are the columns of x. For discussion see
# Hurvich and Tsai (1989, Biometrika).
#

<-
2

name
1
n

call
2

name
1
length

name
1
y

<-
2

name
1
p

call
3

name
1
+

call
2

name
1
ncol

call
2

name
1
as.matrix

name
1
x

numeric
1
1

<-
2

name
1
lll

call
2

name
1
lm

call
3

name
1
~

name
1
y

name
1
x

call
3

name
1
+

call
2

name
1
log

call
3

name
1
/

call
2

name
1
deviance

name
1
lll

name
1
n

call
3

name
1
/

(
2

name
1
(

call
3

name
1
+

numeric
1
1

call
3

name
1
/

name
1
p

name
1
n

(
2

name
1
(

call
3

name
1
-

call
3

name
1
-

numeric
1
1

call
3

name
1
/

name
1
p

name
1
n

call
3

name
1
/

numeric
1
2

name
1
n
crit.semi1
function
5
df
missing
0
x
missing
0
z
missing
0
y
missing
0

{
3

comment.expression
2

comment
1

character
14
#
# Function to calculate the AICC value for a semiparametric model. The target
# variable is y, the parametric predictors are the columns of x, and the one
# nonparametric (smooth) predictor is z. The nonparametric portion of the 
# model is fit using a cubic smoothing spline. For discussion see
# Simonoff and Tsai (1999, Journal of Computational & Graphical 
# Statistics).
#
# Choosing the semiparametric model based on AICC is a two-step procedure. 
# First, for a given set of parametric predictors, use a grid
# search to find the df for the smooth with minimum AICC. Then, cycle through
# all possible sets of parametric predictors. The model with overall smallest
# value of AICC is the model of choice.
#

<-
2

name
1
n

call
2

name
1
length

name
1
y

<-
2

name
1
ggg

call
2

name
1
gam

call
3

name
1
~

name
1
y

call
3

name
1
+

name
1
x

call
3

name
1
s

name
1
z
df
name
1
df

call
3

name
1
+

call
2

name
1
log

call
3

name
1
/

call
3

name
1
$

name
1
ggg

character
1
deviance

name
1
n

(
2

name
1
(

call
3

name
1
/

(
2

name
1
(

call
3

name
1
-

call
3

name
1
*

numeric
1
2

name
1
n

call
3

name
1
$

name
1
ggg

character
1
df

(
2

name
1
(

call
3

name
1
-

call
3

name
1
$

name
1
ggg

character
1
df

numeric
1
2
crit.semi2
function
6
df
missing
0
x
missing
0
z1
missing
0
z2
missing
0
y
missing
0

{
3

comment.expression
2

comment
1

character
14
#
# Function to calculate the AICC value for a semiparametric model. The target
# variable is y, the parametric predictors are the columns of x, and the two 
# nonparametric (smooth) predictors are z1 and z2. The nonparametric portion 
# of the model is fit using cubic smoothing splines. For discussion see
# Simonoff and Tsai (1999, Journal of Computational & Graphical 
# Statistics).
#
# Choosing the semiparametric model based on AICC is a two-step procedure. 
# First, for a given set of parametric predictors, use a grid
# search to find the df for the smooth with minimum AICC. Then, cycle through
# all possible sets of parametric predictors. The model with overall smallest
# value of AICC is the model of choice.
#

<-
2

name
1
n

call
2

name
1
length

name
1
y

<-
2

name
1
ggg

call
2

name
1
gam

call
3

name
1
~

name
1
y

call
3

name
1
+

call
3

name
1
+

name
1
x

call
3

name
1
s

name
1
z1
df
call
3

name
1
[

name
1
df

numeric
1
1

call
3

name
1
s

name
1
z2
df
call
3

name
1
[

name
1
df

numeric
1
2

call
3

name
1
+

call
2

name
1
log

call
3

name
1
/

call
3

name
1
$

name
1
ggg

character
1
deviance

name
1
n

(
2

name
1
(

call
3

name
1
/

(
2

name
1
(

call
3

name
1
-

call
3

name
1
*

numeric
1
2

name
1
n

call
3

name
1
$

name
1
ggg

character
1
df

(
2

name
1
(

call
3

name
1
-

call
3

name
1
$

name
1
ggg

character
1
df

numeric
1
2
crit.semi3
function
7
df
missing
0
x
missing
0
z1
missing
0
z2
missing
0
z3
missing
0
y
missing
0

{
3

comment.expression
2

comment
1

character
14
#
# Function to calculate the AICC value for a semiparametric model. The target
# variable is y, the parametric predictors are the columns of x, and the three 
# nonparametric (smooth) predictors are z1, z2 and z3. The nonparametric portion 
# of the model is fit using cubic smoothing splines. For discussion see
# Simonoff and Tsai (1999, Journal of Computational & Graphical 
# Statistics).
#
# Choosing the semiparametric model based on AICC is a two-step procedure. 
# First, for a given set of parametric predictors, use a grid
# search to find the df for the smooth with minimum AICC. Then, cycle through
# all possible sets of parametric predictors. The model with overall smallest
# value of AICC is the model of choice.
#

<-
2

name
1
n

call
2

name
1
length

name
1
y

<-
2

name
1
ggg

call
2

name
1
gam

call
3

name
1
~

name
1
y

call
3

name
1
+

call
3

name
1
+

call
3

name
1
+

name
1
x

call
3

name
1
s

name
1
z1
df
call
3

name
1
[

name
1
df

numeric
1
1

call
3

name
1
s

name
1
z2
df
call
3

name
1
[

name
1
df

numeric
1
2

call
3

name
1
s

name
1
z3
df
call
3

name
1
[

name
1
df

numeric
1
3

call
3

name
1
+

call
2

name
1
log

call
3

name
1
/

call
3

name
1
$

name
1
ggg

character
1
deviance

name
1
n

(
2

name
1
(

call
3

name
1
/

(
2

name
1
(

call
3

name
1
-

call
3

name
1
*

numeric
1
2

name
1
n

call
3

name
1
$

name
1
ggg

character
1
df

(
2

name
1
(

call
3

name
1
-

call
3

name
1
$

name
1
ggg

character
1
df

numeric
1
2
crit.spline
function
4
df
missing
0
x
missing
0
y
missing
0

{
2

comment.expression
2

comment
1

character
3
#
# AICC criterion for cubic smoothing spline estimator
#

<-
2

name
1
n

call
2

name
1
length

name
1
y

call
3

name
1
+

call
2

name
1
log

call
3

name
1
/

call
2

name
1
sum

call
3

name
1
^

(
2

name
1
(

call
3

name
1
-

call
3

name
1
$

call
3

name
1
predict

call
4

name
1
smooth.spline

name
1
x

name
1
y
df
name
1
df

name
1
x

character
1
y

name
1
y

numeric
1
2

name
1
n

call
3

name
1
/

(
2

name
1
(

call
3

name
1
+

numeric
1
1

call
3

name
1
/

name
1
df

name
1
n

(
2

name
1
(

call
3

name
1
-

call
3

name
1
-

numeric
1
1

call
3

name
1
/

name
1
df

name
1
n

call
3

name
1
/

numeric
1
2

name
1
n
goftest1
function
5
x
missing
0
y
missing
0
tail
name
1
T
nsim
numeric
1
100

{
2

comment.expression
2

comment
1

character
7
#
# Function to calculate the AICC-based goodness-of-fit test of linearity. The
# target variable is y, and the predicting variable is x. By default, a Monte
# Carlo-based tail probability is given based on 100 simulation replications.
# For discussion see Simonoff and Tsai (1999, Journal of Computational & Graphical 
# Statistics).
#

<-
2

name
1
test

call
3

name
1
-

call
3

name
1
crit.linear

name
1
x

name
1
y

call
3

name
1
$

call
3

name
1
aicc.spline

name
1
x

name
1
y

character
1
aicc

if
3

name
1
tail

{
5

<-
2

name
1
p

numeric
1
0

<-
2

name
1
lll

call
3

name
1
$

call
3

name
1
lsfit

name
1
x

name
1
y

character
1
residuals

<-
2

name
1
yfit

call
3

name
1
-

name
1
y

name
1
lll

for
3
i
NULL
0

call
3

name
1
:

numeric
1
1

name
1
nsim

{
3

<-
2

name
1
yy

call
3

name
1
+

name
1
yfit

call
3

name
1
sample

name
1
lll
replace
name
1
T

<-
2

name
1
testt

call
3

name
1
-

call
3

name
1
crit.linear

name
1
x

name
1
yy

call
3

name
1
$

call
3

name
1
aicc.spline

name
1
x

name
1
yy

character
1
aicc

if
2

call
3

name
1
<=

name
1
test

name
1
testt

<-
2

name
1
p

call
3

name
1
+

name
1
p

numeric
1
1

call
3

name
1
list
test
name
1
test
p
call
3

name
1
/

name
1
p

name
1
nsim

call
3

name
1
list
test
name
1
test
p
name
1
NA
goftest2
function
7
x1
missing
0
x2
missing
0
y
missing
0
tail
name
1
T
nsim
numeric
1
100
start
call
3

name
1
c

call
3

name
1
/

name
1
n

numeric
1
2

call
3

name
1
/

name
1
n

numeric
1
2

{
2

comment.expression
2

comment
1

character
8
#
# Function to calculate the AICC-based goodness-of-fit test of linearity based 
# on two predictors. The target variable is y, and the predicting variables
# are x1 and x2. By default, a Monte Carlo-based tail probability is given 
# based on 100 simulation replications. For discussion see
# Simonoff and Tsai (1999, Journal of Computational & Graphical 
# Statistics).
#

<-
2

name
1
test

call
3

name
1
-

call
3

name
1
crit.linear

call
3

name
1
cbind

name
1
x1

name
1
x2

name
1
y

call
3

name
1
$

call
5

name
1
aicc.am2

name
1
x1

name
1
x2

name
1
y
start
name
1
start

character
1
aicc

if
3

name
1
tail

{
5

<-
2

name
1
p

numeric
1
0

<-
2

name
1
lll

call
3

name
1
$

call
3

name
1
lsfit

call
3

name
1
cbind

name
1
x1

name
1
x2

name
1
y

character
1
residuals

<-
2

name
1
yfit

call
3

name
1
-

name
1
y

name
1
lll

for
3
i
NULL
0

call
3

name
1
:

numeric
1
1

name
1
nsim

{
3

<-
2

name
1
yy

call
3

name
1
+

name
1
yfit

call
3

name
1
sample

name
1
lll
replace
name
1
T

<-
2

name
1
testt

call
3

name
1
-

call
3

name
1
crit.linear

call
3

name
1
cbind

name
1
x1

name
1
x2

name
1
yy

call
3

name
1
$

call
5

name
1
aicc.am2

name
1
x1

name
1
x2

name
1
yy
start
name
1
start

character
1
aicc

if
2

call
3

name
1
<=

name
1
test

name
1
testt

<-
2

name
1
p

call
3

name
1
+

name
1
p

numeric
1
1

call
3

name
1
list
test
name
1
test
p
call
3

name
1
/

name
1
p

name
1
nsim

call
3

name
1
list
test
name
1
test
p
name
1
NA
goftest3
function
8
x1
missing
0
x2
missing
0
x3
missing
0
y
missing
0
tail
name
1
T
nsim
numeric
1
100
start
call
4

name
1
c

call
3

name
1
/

name
1
n

numeric
1
2

call
3

name
1
/

name
1
n

numeric
1
2

call
3

name
1
/

name
1
n

numeric
1
2

{
2

comment.expression
2

comment
1

character
8
#
# Function to calculate the AICC-based goodness-of-fit test of linearity based 
# on three predictors. The target variable is y, and the predicting variables
# are x1, x2, and x3. By default, a Monte Carlo-based tail probability is given 
# based on 100 simulation replications. For discussion see
# Simonoff and Tsai (1999, Journal of Computational & Graphical 
# Statistics).
#

<-
2

name
1
test

call
3

name
1
-

call
3

name
1
crit.linear

call
4

name
1
cbind

name
1
x1

name
1
x2

name
1
x3

name
1
y

call
3

name
1
$

call
6

name
1
aicc.am3

name
1
x1

name
1
x2

name
1
x3

name
1
y
start
name
1
start

character
1
aicc

if
3

name
1
tail

{
5

<-
2

name
1
p

numeric
1
0

<-
2

name
1
lll

call
3

name
1
$

call
3

name
1
lsfit

call
4

name
1
cbind

name
1
x1

name
1
x2

name
1
x3

name
1
y

character
1
residuals

<-
2

name
1
yfit

call
3

name
1
-

name
1
y

name
1
lll

for
3
i
NULL
0

call
3

name
1
:

numeric
1
1

name
1
nsim

{
3

<-
2

name
1
yy

call
3

name
1
+

name
1
yfit

call
3

name
1
sample

name
1
lll
replace
name
1
T

<-
2

name
1
testt

call
3

name
1
-

call
3

name
1
crit.linear

call
4

name
1
cbind

name
1
x1

name
1
x2

name
1
x3

name
1
yy

call
3

name
1
$

call
6

name
1
aicc.am3

name
1
x1

name
1
x2

name
1
x3

name
1
yy
start
name
1
start

character
1
aicc

if
2

call
3

name
1
<=

name
1
test

name
1
testt

<-
2

name
1
p

call
3

name
1
+

name
1
p

numeric
1
1

call
3

name
1
list
test
name
1
test
p
call
3

name
1
/

name
1
p

name
1
nsim

call
3

name
1
list
test
name
1
test
p
name
1
NA
mattrace
function
2
x
missing
0

{
1

comment.expression
2

comment
1

character
3
#
# Function to calculate the trace of a matrix x
#

call
2

name
1
sum

call
3

name
1
[

name
1
x

call
4

name
1
seq

numeric
1
1

call
3

name
1
^

call
2

name
1
ncol

name
1
x

numeric
1
2

call
3

name
1
+

call
2

name
1
ncol

name
1
x

numeric
1
1
cars93.dat
structure
3
.Data
list
8
Auto
structure
3
.Data
integer
93
1
2
4
3
5
6
7
9
8
10
11
15
16
13
18
19
12
14
17
20
22
21
24
26
27
23
25
28
29
30
34
33
38
35
36
31
37
32
39
40
43
42
41
45
44
46
47
48
49
50
51
52
53
56
54
55
57
58
59
60
61
63
62
67
64
66
65
68
69
71
70
72
76
77
74
75
73
78
79
80
82
81
83
87
85
84
86
90
89
91
88
92
93
.Label
character
93
Acura_Integra
Acura_Legend
Audi_100
Audi_90
BMW_535i
Buick_Century
Buick_LeSabre
Buick_Riviera
Buick_Roadmaster
Cadillac_DeVille
Cadillac_Seville
Chevrolet_Astro
Chevrolet_Camaro
Chevrolet_Caprice
Chevrolet_Cavalier
Chevrolet_Corsica
Chevrolet_Corvette
Chevrolet_Lumina
Chevrolet_Lumina_APV
Chrylser_Concorde
Chrysler_Imperial
Chrysler_LeBaron
Dodge_Cara6
Dodge_Colt
Dodge_Dynasty
Dodge_Shadow
Dodge_Spirit
Dodge_Stealth
Eagle_Summit
Eagle_Vision
Ford_Aerostar
Ford_Crown_Victoria
Ford_Escort
Ford_Festiva
Ford_Mustang
Ford_Probe
Ford_Taurus
Ford_Tempo
Geo_Metro
Geo_Storm
Honda_Accord
Honda_Civic
Honda_Prelude
Hyundai_Elantra
Hyundai_Excel
Hyundai_Scoupe
Hyundai_Sonata
Infiniti_Q45
Lexus_ES300
Lexus_SC300
Lincoln_Continental
Lincoln_Town_Car
Mazda_323
Mazda_626
Mazda_MPV
Mazda_Protege
Mazda_RX-7
Mercedes-Benz_190E
Mercedes-Benz_300E
Mercury_Capri
Mercury_Cougar
Mitsubishi_Diamante
Mitsubishi_Mirage
Nissan_Altima
Nissan_Maxima
Nissan_Quest
Nissan_Sentra
Oldsmobile_Achieva
Oldsmobile_Cutlass_Ci
Oldsmobile_Eighty-Eig
Oldsmobile_Silhouette
Plymouth_Laser
Pontiac_Bonneville
Pontiac_Firebird
Pontiac_Grand_Prix
Pontiac_LeMans
Pontiac_Sunbird
Saab_900
Saturn_SL
Subaru_Justy
Subaru_Legacy
Subaru_Loyale
Suzuki_Swift
Toyota_Camry
Toyota_Celica
Toyota_Previa
Toyota_Tercel
Volkswagen_Corrado
Volkswagen_Euro6
Volkswagen_Fox
Volkswagen_Passat
Volvo_240
Volvo_850
class
character
2
AsIs
factor
Hiway.MPG
structure
2
.Data
numeric
93
31
25
26
26
30
31
28
25
27
25
25
36
34
28
29
23
20
26
25
28
28
26
33
29
27
21
27
24
33
28
33
30
27
29
30
20
30
26
50
36
31
46
31
33
29
34
27
22
24
23
26
26
37
36
34
24
25
29
25
26
26
33
24
33
30
23
26
31
31
23
28
30
41
31
28
27
28
26
38
37
30
30
43
37
32
29
22
33
21
30
25
28
28
.guiColInfo
character
5
TXPROP_ColWidth<-15
TXPROP_ColJustification<-Right
TXPROP_ObjectNote<-
TXPROP_ColFloatFormat<-Decimal
TXPROP_ColPrecision<-2
Engine.size
structure
2
.Data
numeric
93
1.8
3.2
2.8
2.8
3.5
2.2
3.8
5.7
3.8
4.9
4.6
2.2
2.2
3.4
2.2
3.8
4.3
5
5.7
3.3
3
3.3
1.5
2.2
2.5
3
2.5
3
1.5
3.5
1.3
1.8
2.3
2.3
2
3
3
4.6
1
1.6
2.3
1.5
2.2
1.5
1.8
1.5
2
4.5
3
3
3.8
4.6
1.6
1.8
2.5
3
1.3
2.3
3.2
1.6
3.8
1.5
3
1.6
2.4
3
3
2.3
2.2
3.8
3.8
1.8
1.6
2
3.4
3.4
3.8
2.1
1.9
1.2
1.8
2.2
1.3
1.5
2.2
2.2
2.4
1.8
2.5
2
2.8
2.3
2.4
.guiColInfo
character
5
TXPROP_ColWidth<-16
TXPROP_ColJustification<-Right
TXPROP_ObjectNote<-
TXPROP_ColFloatFormat<-Decimal
TXPROP_ColPrecision<-2
Horsepower
structure
2
.Data
numeric
93
140
200
172
172
208
110
170
180
170
200
295
110
110
160
110
170
165
170
300
153
141
147
92
93
100
142
100
300
92
214
63
127
96
105
115
145
140
190
55
90
160
102
140
81
124
92
128
278
185
225
160
210
82
103
164
155
255
130
217
100
140
92
202
110
150
151
160
155
110
170
170
92
74
110
160
200
170
140
85
73
90
130
70
82
135
130
138
81
109
134
178
114
168
.guiColInfo
character
5
TXPROP_ColWidth<-13
TXPROP_ColJustification<-Right
TXPROP_ObjectNote<-
TXPROP_ColFloatFormat<-Decimal
TXPROP_ColPrecision<-2
Weight
numeric
93
2705
3560
3375
3405
3640
2880
3470
4105
3495
3620
3935
2490
2785
3240
3195
3715
4025
3910
3380
3515
3085
3570
2270
2670
2970
3705
3080
3805
2295
3490
1845
2530
2690
2850
2710
3735
3325
3950
1695
2475
2865
2350
3040
2345
2620
2285
2885
4000
3510
3515
3695
4055
2325
2440
2970
3735
2895
2920
3525
2450
3610
2295
3730
2545
3050
4100
3200
2910
2890
3715
3470
2640
2350
2575
3240
3450
3495
2775
2495
2045
2490
3085
1965
2055
2950
3030
3785
2240
3960
2985
2810
2985
3245
Manual
numeric
93
1
1
1
1
1
0
0
0
0
0
0
1
1
1
0
0
0
0
1
0
0
0
1
1
1
0
0
1
1
0
1
1
1
1
1
1
0
0
1
1
1
1
1
1
1
1
1
0
1
1
0
0
1
1
1
0
1
1
0
1
0
1
0
1
1
0
0
0
0
0
0
1
1
1
1
1
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Domestic
numeric
93
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Van
numeric
93
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
1
0
0
0
0
class
character
1
data.frame
row.names
character
93
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
collrank.dat
structure
3
.Data
list
7
University
structure
3
.Data
integer
48
47
22
16
11
19
24
10
3
4
21
8
38
41
9
17
23
40
46
12
44
43
26
14
39
28
45
35
5
2
42
29
18
7
30
20
27
37
1
6
36
34
33
25
48
13
31
15
32
.Label
character
48
Boston College
Brandeis University
Brown University
California Institute of Technol
Carnegie Mellon University
Case Western Reserve Univ.
College of William and Mary
Columbia University
Cornell University
Dartmouth College
Duke University
Emory University
George Washington University
Georgetown University
Georgia Institute of Technology
Harvard University
Johns Hopkins University
Lehigh University
Massachusetts Inst. of Technolo
New York University
Northwestern University
Princeton University
Rice University
Stanford University
Syracuse University
Tufts University
Tulane University
U. of North Carolina--Chapel Hi
Univ. of California--Los Angele
Univ. of California--San Diego
Univ. of California--Santa Barb
Univ. of Illinois--Urbana-Champ
Univ. of Southern California
Univ. of Wisconsin--Madison
University of California--Berke
University of California--Davis
University of California--Irvin
University of Chicago
University of Michigan--Ann Arb
University of Notre Dame
University of Pennsylvania
University of Rochester
University of Virginia
Vanderbilt University
Wake Forest University
Washington University
Yale University
Yeshiva University
class
character
2
AsIs
factor
Size.gt.50
numeric
48
9
13
21
7
12
14
13
13
5
10
9
5
16
11
11
9
13
7
7
5
15
7
14
14
13
5
18
12
8
13
29
8
8
26
9
8
16
10
19
22
20
16
8
1
10
19
7
17
Retention
numeric
48
98
97
96
96
97
97
97
96
94
96
95
91
95
95
94
95
97
94
91
91
97
97
95
94
94
92
94
88
90
93
94
90
93
93
86
86
93
94
91
91
91
89
89
88
89
86
86
91
Graduation
numeric
48
95
95
97
95
89
93
94
93
85
90
88
86
87
89
87
89
94
85
90
83
93
88
90
85
84
87
79
70
82
77
77
86
91
74
70
73
68
87
71
74
72
65
71
72
69
70
69
79
Reputation
numeric
48
3
1
3
11
1
3
11
11
7
11
11
3
11
7
7
18
30
24
30
24
18
41
28
11
21
65
7
21
41
55
21
65
30
41
41
55
49
55
49
41
18
37
55
103
65
65
30
24
U.S.News.rank
numeric
48
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
25
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
43
44
45
46
47
48
50
Expenditure
numeric
48
4.658078206
4.510772821
4.632477539
4.500401157
4.572592821
4.563991023
4.498186452
4.37819797
4.869038005
4.473632927
4.515052144
4.609444995
4.505448453
4.348012638
4.790313318
4.412930709
4.200686376
4.73255458
4.479474534
4.406369835
4.150510833
4.305910003
4.320644711
4.212001058
4.264770615
4.690647866
4.214154786
4.411468181
4.247727833
4.474711485
4.33243846
4.190415747
4.031691169
4.319626484
4.382647303
4.268297087
4.224766074
4.089940418
4.299725154
4.23230984
4.079145053
4.26904571
4.188028041
4.362463625
4.226728757
4.023992805
4.12762305
4.04743064
class
character
1
data.frame
row.names
character
48
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
salinity.dat
structure
3
.Data
list
4
Salinity
numeric
28
7.6
7.7
4.3
5.9
5
6.5
8.300000000000001
8.199999999999999
13.2
12.6
10.4
10.8
13.1
12.3
10.4
10.5
7.7
9.5
12
12.6
13.6
14.1
13.5
11.5
12
13
14.1
15.1
Lagged.salinity
numeric
28
8.199999999999999
7.6
4.6
4.3
5.9
5
6.5
8.300000000000001
10.1
13.2
12.6
10.4
10.8
13.1
13.3
10.4
10.5
7.7
10
12
12.1
13.6
15
13.5
11.5
12
13
14.1
Trend
numeric
28
4
5
0
1
2
3
4
5
0
1
2
3
4
5
0
1
2
3
0
1
4
5
0
1
2
3
4
5
Water.flow
numeric
28
23.005
23.873
26.417
24.868
29.895
24.2
23.215
21.862
22.274
23.83
25.144
22.43
21.785
22.38
23.927
23.443
24.859
22.686
21.789
22.041
21.033
21.005
25.865
26.29
22.932
21.313
20.769
21.393
row.names
character
28
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
class
character
1
data.frame