Title: | Mixture and Flexible Discriminant Analysis |
---|---|
Description: | Mixture and flexible discriminant analysis, multivariate adaptive regression splines (MARS), BRUTO, and vector-response smoothing splines. Hastie, Tibshirani and Friedman (2009) "Elements of Statistical Learning (second edition, chap 12)" Springer, New York. |
Authors: | Trevor Hastie [aut, cre] (Original co-author of the S package `mda`), Robert Tibshirani [aut] (Original co-author of the S package `mda`), Balasubramanian Narasimhan [ctb] (Contributed to the upgrading of code), Friedrich Leisch [ctb] (Original R port from the S package), Kurt Hornik [ctb] (Original R port from the S package), Brian Ripley [ctb] (Original R port from the S package) |
Maintainer: | Trevor Hastie <[email protected]> |
License: | GPL-2 |
Version: | 0.5-5 |
Built: | 2024-12-08 07:19:26 UTC |
Source: | CRAN |
Fit an additive spline model by adaptive backfitting.
bruto(x, y, w, wp, dfmax, cost, maxit.select, maxit.backfit, thresh = 0.0001, trace.bruto = FALSE, start.linear = TRUE, fit.object, ...)
bruto(x, y, w, wp, dfmax, cost, maxit.select, maxit.backfit, thresh = 0.0001, trace.bruto = FALSE, start.linear = TRUE, fit.object, ...)
x |
a matrix of numeric predictors (does not include the column of 1s). |
y |
a vector or matrix of responses. |
w |
optional observation weight vector. |
wp |
optional weight vector for each column of |
dfmax |
a vector of maximum df (degrees of freedom) for each term. |
cost |
cost per degree of freedom; default is 2. |
maxit.select |
maximum number of iterations during the selection stage. |
maxit.backfit |
maximum number of iterations for the final backfit stage (with fixed lambda). |
thresh |
convergence threshold (default is 0.0001); iterations cease when the relative change in GCV is below this threshold. |
trace.bruto |
logical flag. If |
start.linear |
logical flag. If |
fit.object |
This the object returned by |
... |
further arguments to be passed to or from methods. |
A multiresponse additive model fit object of class "bruto"
is
returned. The model is fit by adaptive backfitting using smoothing
splines. If there are np
columns in y
, then np
additive models are fit, but the same amount of smoothing (df) is
used for each term. The procedure chooses between df = 0
(term omitted), df = 1
(term linear) or df > 0
(term
fitted by smoothing spline). The model selection is based on an
approximation to the GCV criterion, which is used at each step of the
backfitting procedure. Once the selection process stops, the model is
backfit using the chosen amount of smoothing.
A bruto object has the following components of interest:
lambda |
a vector of chosen smoothing parameters, one for each
column of |
df |
the df chosen for each column of |
type |
a factor with levels |
gcv.select gcv.backfit df.select |
The sequence of gcv values and df selected during the execution of the function. |
nit |
the number of iterations used. |
fitted.values |
a matrix of fitted values. |
residuals |
a matrix of residuals. |
call |
the call that produced this object. |
Trevor Hastie and Rob Tibshirani, Generalized Additive Models, Chapman and Hall, 1990 (page 262).
Trevor Hastie, Rob Tibshirani and Andreas Buja “Flexible Discriminant Analysis by Optimal Scoring” JASA 1994, 89, 1255-1270.
predict.bruto
data(trees) fit1 <- bruto(trees[,-3], trees[3]) fit1$type fit1$df ## examine the fitted functions par(mfrow=c(1,2), pty="s") Xp <- matrix(sapply(trees[1:2], mean), nrow(trees), 2, byrow=TRUE) for(i in 1:2) { xr <- sapply(trees, range) Xp1 <- Xp; Xp1[,i] <- seq(xr[1,i], xr[2,i], len=nrow(trees)) Xf <- predict(fit1, Xp1) plot(Xp1[ ,i], Xf, xlab=names(trees)[i], ylab="", type="l") }
data(trees) fit1 <- bruto(trees[,-3], trees[3]) fit1$type fit1$df ## examine the fitted functions par(mfrow=c(1,2), pty="s") Xp <- matrix(sapply(trees[1:2], mean), nrow(trees), 2, byrow=TRUE) for(i in 1:2) { xr <- sapply(trees, range) Xp1 <- Xp; Xp1[,i] <- seq(xr[1,i], xr[2,i], len=nrow(trees)) Xf <- predict(fit1, Xp1) plot(Xp1[ ,i], Xf, xlab=names(trees)[i], ylab="", type="l") }
a method for coef for extracting the canonical coefficients from an fda or mda object
## S3 method for class 'fda' coef(object, ...)
## S3 method for class 'fda' coef(object, ...)
object |
an |
... |
not relevant |
See the references for details.
A coefficient matrix
Trevor Hastie and Robert Tibshirani
“Flexible Disriminant Analysis by Optimal Scoring” by Hastie, Tibshirani and Buja, 1994, JASA, 1255-1270.
“Penalized Discriminant Analysis” by Hastie, Buja and Tibshirani, 1995, Annals of Statistics, 73-102.
“Elements of Statisical Learning - Data Mining, Inference and Prediction” (2nd edition, Chapter 12) by Hastie, Tibshirani and Friedman, 2009, Springer
predict.fda
,
plot.fda
,
mars
,
bruto
,
polyreg
,
softmax
,
confusion
,
data(iris) irisfit <- fda(Species ~ ., data = iris) coef(irisfit) mfit=mda(Species~.,data=iris,subclass=2) coef(mfit)
data(iris) irisfit <- fda(Species ~ ., data = iris) coef(irisfit) mfit=mda(Species~.,data=iris,subclass=2) coef(mfit)
Compute the confusion matrix between two factors, or for an fda or mda object.
## Default S3 method: confusion(object, true, ...) ## S3 method for class 'fda' confusion(object, data, ...)
## Default S3 method: confusion(object, true, ...) ## S3 method for class 'fda' confusion(object, data, ...)
object |
the predicted factor, or an fda or mda model object. |
true |
the true factor. |
data |
a data frame (list) containing the test data. |
... |
further arguments to be passed to or from methods. |
This is a generic function.
For the default method essentially table(object, true)
, but
with some useful attribute(s).
data(iris) irisfit <- fda(Species ~ ., data = iris) confusion(predict(irisfit, iris), iris$Species) ## Setosa Versicolor Virginica ## Setosa 50 0 0 ## Versicolor 0 48 1 ## Virginica 0 2 49 ## attr(, "error"): ## [1] 0.02
data(iris) irisfit <- fda(Species ~ ., data = iris) confusion(predict(irisfit, iris), iris$Species) ## Setosa Versicolor Virginica ## Setosa 50 0 0 ## Versicolor 0 48 1 ## Virginica 0 2 49 ## attr(, "error"): ## [1] 0.02
A list with training data and other details for the mixture example
data(ESL.mixture)
data(ESL.mixture)
This list contains the following elements:
a 200x2 matrix of predictors.
a 200 vector of y values taking values 0 or 1.
a 6831x2 matrix of prediction points, on a 69x99 grid.
a vector of 6831 probabilities - the true probabilities
of a 1 at each point in xnew
.
the marginal distribution of the predictors t each
point in xnew
.
grid values for first coordinate in xnew
.
grid values for second coordinate in xnew
.
a 20 x 2 matrix of means used in the generation of these data.
"Elements of Statistical Learning (second edition)", Hastie, T., Tibshirani, R. and Friedman, J. (2009), Springer, New York. https://hastie.su.domains/ElemStatLearn/
Flexible discriminant analysis.
fda(formula, data, weights, theta, dimension, eps, method, keep.fitted, ...)
fda(formula, data, weights, theta, dimension, eps, method, keep.fitted, ...)
formula |
of the form |
data |
data frame containing the variables in the formula (optional). |
weights |
an optional vector of observation weights. |
theta |
an optional matrix of class scores, typically with less
than |
dimension |
The dimension of the solution, no greater than
|
eps |
a threshold for small singular values for excluding
discriminant variables; default is |
method |
regression method used in optimal scaling. Default is
linear regression via the function |
keep.fitted |
a logical variable, which determines whether the
(sometimes large) component |
... |
additional arguments to |
an object of class "fda"
. Use predict
to extract
discriminant variables, posterior probabilities or predicted class
memberships. Other extractor functions are coef
,
confusion
and plot
.
The object has the following components:
percent.explained |
the percent between-group variance explained by each dimension (relative to the total explained.) |
values |
optimal scaling regression sum-of-squares for each
dimension (see reference). The usual discriminant analysis
eigenvalues are given by |
means |
class means in the discriminant space. These are also
scaled versions of the final theta's or class scores, and can be
used in a subsequent call to |
theta.mod |
(internal) a class scoring matrix which allows
|
dimension |
dimension of discriminant space. |
prior |
class proportions for the training data. |
fit |
fit object returned by |
call |
the call that created this object (allowing it to be
|
confusion |
confusion matrix when classifying the training data. |
The method
functions are required to take arguments x
and y
where both can be matrices, and should produce a matrix
of fitted.values
the same size as y
. They can take
additional arguments weights
and should all have a ...
for safety sake. Any arguments to method
can be passed on via
the ...
argument of fda
. The default method
polyreg
has a degree
argument which allows
polynomial regression of the required total degree. See the
documentation for predict.fda
for further requirements
of method
. The package earth
is suggested for this
package as well; earth
is a more detailed implementation of
the mars model, and works as a method
argument.
Trevor Hastie and Robert Tibshirani
“Flexible Disriminant Analysis by Optimal Scoring” by Hastie, Tibshirani and Buja, 1994, JASA, 1255-1270.
“Penalized Discriminant Analysis” by Hastie, Buja and Tibshirani, 1995, Annals of Statistics, 73-102.
“Elements of Statisical Learning - Data Mining, Inference and Prediction” (2nd edition, Chapter 12) by Hastie, Tibshirani and Friedman, 2009, Springer
predict.fda
,
plot.fda
,
mars
,
bruto
,
polyreg
,
softmax
,
confusion
,
data(iris) irisfit <- fda(Species ~ ., data = iris) irisfit ## fda(formula = Species ~ ., data = iris) ## ## Dimension: 2 ## ## Percent Between-Group Variance Explained: ## v1 v2 ## 99.12 100.00 ## ## Degrees of Freedom (per dimension): 5 ## ## Training Misclassification Error: 0.02 ( N = 150 ) confusion(irisfit, iris) ## Setosa Versicolor Virginica ## Setosa 50 0 0 ## Versicolor 0 48 1 ## Virginica 0 2 49 ## attr(, "error"): ## [1] 0.02 plot(irisfit) coef(irisfit) ## [,1] [,2] ## [1,] -2.126479 -6.72910343 ## [2,] -0.837798 0.02434685 ## [3,] -1.550052 2.18649663 ## [4,] 2.223560 -0.94138258 ## [5,] 2.838994 2.86801283 marsfit <- fda(Species ~ ., data = iris, method = mars) marsfit2 <- update(marsfit, degree = 2) marsfit3 <- update(marsfit, theta = marsfit$means[, 1:2]) ## this refits the model, using the fitted means (scaled theta's) ## from marsfit to start the iterations
data(iris) irisfit <- fda(Species ~ ., data = iris) irisfit ## fda(formula = Species ~ ., data = iris) ## ## Dimension: 2 ## ## Percent Between-Group Variance Explained: ## v1 v2 ## 99.12 100.00 ## ## Degrees of Freedom (per dimension): 5 ## ## Training Misclassification Error: 0.02 ( N = 150 ) confusion(irisfit, iris) ## Setosa Versicolor Virginica ## Setosa 50 0 0 ## Versicolor 0 48 1 ## Virginica 0 2 49 ## attr(, "error"): ## [1] 0.02 plot(irisfit) coef(irisfit) ## [,1] [,2] ## [1,] -2.126479 -6.72910343 ## [2,] -0.837798 0.02434685 ## [3,] -1.550052 2.18649663 ## [4,] 2.223560 -0.94138258 ## [5,] 2.838994 2.86801283 marsfit <- fda(Species ~ ., data = iris, method = mars) marsfit2 <- update(marsfit, degree = 2) marsfit3 <- update(marsfit, theta = marsfit$means[, 1:2]) ## this refits the model, using the fitted means (scaled theta's) ## from marsfit to start the iterations
Perform a penalized regression, as used in penalized discriminant analysis.
gen.ridge(x, y, weights, lambda=1, omega, df, ...)
gen.ridge(x, y, weights, lambda=1, omega, df, ...)
x , y , weights
|
the x and y matrix and possibly a weight vector. |
lambda |
the shrinkage penalty coefficient. |
omega |
a penalty object; omega is the eigendecomposition of the penalty matrix, and need not have full rank. By default, standard ridge is used. |
df |
an alternative way to prescribe lambda, using the notion of equivalent degrees of freedom. |
... |
currently not used. |
A generalized ridge regression, where the coefficients are penalized
according to omega. See the function definition for further details.
No functions are provided for producing one dimensional penalty
objects (omega).
laplacian()
creates a two-dimensional penalty
object, suitable for (small) images.
The glass
data frame has 214 observations and 10 variables,
representing glass fragments.
data(glass)
data(glass)
This data frame contains the following columns:
refractive index
weight percent in corresponding oxide
weight percent in corresponding oxide
weight percent in corresponding oxide
weight percent in corresponding oxide
weight percent in corresponding oxide
weight percent in corresponding oxide
weight percent in corresponding oxide
weight percent in corresponding oxide
Type of glass:
building_windows_float_processed,
building_windows_non_float_processed,
vehicle_windows_float_processed,
vehicle_windows_non_float_processed (none in this database),
containers,
tableware,
headlamps
P. M. Murphy and D. W. Aha (1999), UCI Repository of Machine Learning Databases, http://archive.ics.uci.edu/ml/datasets/glass+identification
Creates a penalty matrix for use by gen.ridge
for
two-dimensional smoothing.
laplacian(size, compose) laplacian(size = 16, compose = FALSE)
laplacian(size, compose) laplacian(size = 16, compose = FALSE)
size |
dimension of the image is |
compose |
default is |
Formulas are used to construct a laplacian for smoothing a square image.
If compose=FALSE
, an eigen-decomposition object is
returned. The vectors
component is a size^2 x size^2
orthogonal matrix, and the $values
component is a size^2
vector of non-negative eigen-values. If compose=TRUE
, these are
multiplied together to form a single matrix.
Trevor Hastie <[email protected]
Here we follow very closely the material on page 635 in JASA 1991 of O'Sullivan's article on discretized Laplacian Smoothing
Multivariate adaptive regression splines.
mars(x, y, w, wp, degree, nk, penalty, thresh, prune, trace.mars, forward.step, prevfit, ...)
mars(x, y, w, wp, degree, nk, penalty, thresh, prune, trace.mars, forward.step, prevfit, ...)
x |
a matrix containing the independent variables. |
y |
a vector containing the response variable, or in the case of multiple responses, a matrix whose columns are the response values for each variable. |
w |
an optional vector of observation weights (currently ignored). |
wp |
an optional vector of response weights. |
degree |
an optional integer specifying maximum interaction degree (default is 1). |
nk |
an optional integer specifying the maximum number of model terms. |
penalty |
an optional value specifying the cost per degree of freedom charge (default is 2). |
thresh |
an optional value specifying forward stepwise stopping threshold (default is 0.001). |
prune |
an optional logical value specifying whether the model
should be pruned in a backward stepwise fashion (default is
|
trace.mars |
an optional logical value specifying whether info
should be printed along the way (default is |
forward.step |
an optional logical value specifying whether
forward stepwise process should be carried out (default is
|
prevfit |
optional data structure from previous fit. To see the
effect of changing the penalty parameter, one can use prevfit with
|
... |
further arguments to be passed to or from methods. |
An object of class "mars"
, which is a list with the following
components:
call |
call used to |
all.terms |
term numbers in full model. |
selected.terms |
term numbers in selected model. |
penalty |
the input penalty value. |
degree |
the input degree value. |
thresh |
the input threshold value. |
gcv |
gcv of chosen model. |
factor |
matrix with |
cuts |
matrix with |
residuals |
residuals from fit. |
fitted |
fitted values from fit. |
lenb |
length of full model. |
coefficients |
least squares coefficients for final model. |
x |
a matrix of basis functions obtained from the input x matrix. |
This function was coded from scratch, and did not use any of Friedman's mars code. It gives quite similar results to Friedman's program in our tests, but not exactly the same results. We have not implemented Friedman's anova decomposition nor are categorical predictors handled properly yet. Our version does handle multiple response variables, however.
Trevor Hastie and Robert Tibshirani
J. Friedman, “Multivariate Adaptive Regression Splines” (with discussion) (1991). Annals of Statistics, 19/1, 1–141.
predict.mars
,
model.matrix.mars
.
Package earth also provides multivariate adaptive regression
spline models based on the Hastie/Tibshirani mars code in package
mda, adding some extra features. It can be used in the
method
argument of fda
or mda
.
data(trees) fit1 <- mars(trees[,-3], trees[3]) showcuts <- function(obj) { tmp <- obj$cuts[obj$sel, ] dimnames(tmp) <- list(NULL, names(trees)[-3]) tmp } showcuts(fit1) ## examine the fitted functions par(mfrow=c(1,2), pty="s") Xp <- matrix(sapply(trees[1:2], mean), nrow(trees), 2, byrow=TRUE) for(i in 1:2) { xr <- sapply(trees, range) Xp1 <- Xp; Xp1[,i] <- seq(xr[1,i], xr[2,i], len=nrow(trees)) Xf <- predict(fit1, Xp1) plot(Xp1[ ,i], Xf, xlab=names(trees)[i], ylab="", type="l") }
data(trees) fit1 <- mars(trees[,-3], trees[3]) showcuts <- function(obj) { tmp <- obj$cuts[obj$sel, ] dimnames(tmp) <- list(NULL, names(trees)[-3]) tmp } showcuts(fit1) ## examine the fitted functions par(mfrow=c(1,2), pty="s") Xp <- matrix(sapply(trees[1:2], mean), nrow(trees), 2, byrow=TRUE) for(i in 1:2) { xr <- sapply(trees, range) Xp1 <- Xp; Xp1[,i] <- seq(xr[1,i], xr[2,i], len=nrow(trees)) Xf <- predict(fit1, Xp1) plot(Xp1[ ,i], Xf, xlab=names(trees)[i], ylab="", type="l") }
Mixture discriminant analysis.
mda(formula, data, subclasses, sub.df, tot.df, dimension, eps, iter, weights, method, keep.fitted, trace, ...)
mda(formula, data, subclasses, sub.df, tot.df, dimension, eps, iter, weights, method, keep.fitted, trace, ...)
formula |
of the form |
data |
data frame containing the variables in the formula (optional). |
subclasses |
Number of subclasses per class, default is 3. Can be a vector with a number for each class. |
sub.df |
If subclass centroid shrinking is performed, what is the effective degrees of freedom of the centroids per class. Can be a scalar, in which case the same number is used for each class, else a vector. |
tot.df |
The total df for all the centroids can be specified rather than separately per class. |
dimension |
The dimension of the reduced model. If we know our final model will be confined to a discriminant subspace (of the subclass centroids), we can specify this in advance and have the EM algorithm operate in this subspace. |
eps |
A numerical threshold for automatically truncating the dimension. |
iter |
A limit on the total number of iterations, default is 5. |
weights |
NOT observation weights! This is a special
weight structure, which for each class assigns a weight (prior
probability) to each of the observations in that class of belonging
to one of the subclasses. The default is provided by a call to
|
method |
regression method used in optimal scaling. Default is
linear regression via the function |
keep.fitted |
a logical variable, which determines whether the
(sometimes large) component |
trace |
if |
... |
additional arguments to |
An object of class c("mda", "fda")
. The most useful extractor
is predict
, which can make many types of predictions from this
object. It can also be plotted, and any functions useful for fda
objects will work here too, such as confusion
and coef
.
The object has the following components:
percent.explained |
the percent between-group variance explained by each dimension (relative to the total explained.) |
values |
optimal scaling regression sum-of-squares for each dimension (see reference). |
means |
subclass means in the discriminant space. These are also
scaled versions of the final theta's or class scores, and can be
used in a subsequent call to |
theta.mod |
(internal) a class scoring matrix which allows
|
dimension |
dimension of discriminant space. |
sub.prior |
subclass membership priors, computed in the fit. No effort is currently spent in trying to keep these above a threshold. |
prior |
class proportions for the training data. |
fit |
fit object returned by |
call |
the call that created this object (allowing it to be
|
confusion |
confusion matrix when classifying the training data. |
weights |
These are the subclass membership probabilities for each member of the training set; see the weights argument. |
assign.theta |
a pointer list which identifies which elements of certain lists belong to individual classes. |
deviance |
The multinomial log-likelihood of the fit. Even though
the full log-likelihood drives the iterations, we cannot in general
compute it because of the flexibility of the |
The method
functions are required to take arguments x
and y
where both can be matrices, and should produce a matrix
of fitted.values
the same size as y
. They can take
additional arguments weights
and should all have a ...
for safety sake. Any arguments to method() can be passed on via the
...
argument of mda
. The default method
polyreg
has a degree
argument which allows polynomial
regression of the required total degree. See the documentation for
predict.fda
for further requirements of method
.
The package earth
is suggested for this package as well;
earth
is a more detailed implementation of the mars model, and
works as a method
argument.
The function mda.start
creates the starting weights; it takes
additional arguments which can be passed in via the ...
argument to mda
. See the documentation for mda.start
.
Trevor Hastie and Robert Tibshirani
“Flexible Disriminant Analysis by Optimal Scoring” by Hastie, Tibshirani and Buja, 1994, JASA, 1255-1270.
“Penalized Discriminant Analysis” by Hastie, Buja and Tibshirani, 1995, Annals of Statistics, 73-102
“Discriminant Analysis by Gaussian Mixtures” by Hastie and Tibshirani, 1996, JRSS-B, 155-176.
“Elements of Statisical Learning - Data Mining, Inference and Prediction” (2nd edition, Chapter 12) by Hastie, Tibshirani and Friedman, 2009, Springer
predict.mda
,
mars
,
bruto
,
polyreg
,
gen.ridge
,
softmax
,
confusion
data(iris) irisfit <- mda(Species ~ ., data = iris) irisfit ## Call: ## mda(formula = Species ~ ., data = iris) ## ## Dimension: 4 ## ## Percent Between-Group Variance Explained: ## v1 v2 v3 v4 ## 96.02 98.55 99.90 100.00 ## ## Degrees of Freedom (per dimension): 5 ## ## Training Misclassification Error: 0.02 ( N = 150 ) ## ## Deviance: 15.102 data(glass) # random sample of size 100 samp <- c(1, 3, 4, 11, 12, 13, 14, 16, 17, 18, 19, 20, 27, 28, 31, 38, 42, 46, 47, 48, 49, 52, 53, 54, 55, 57, 62, 63, 64, 65, 67, 68, 69, 70, 72, 73, 78, 79, 83, 84, 85, 87, 91, 92, 94, 99, 100, 106, 107, 108, 111, 112, 113, 115, 118, 121, 123, 124, 125, 126, 129, 131, 133, 136, 139, 142, 143, 145, 147, 152, 153, 156, 159, 160, 161, 164, 165, 166, 168, 169, 171, 172, 173, 174, 175, 177, 178, 181, 182, 185, 188, 189, 192, 195, 197, 203, 205, 211, 212, 214) glass.train <- glass[samp,] glass.test <- glass[-samp,] glass.mda <- mda(Type ~ ., data = glass.train) predict(glass.mda, glass.test, type="post") # abbreviations are allowed confusion(glass.mda,glass.test)
data(iris) irisfit <- mda(Species ~ ., data = iris) irisfit ## Call: ## mda(formula = Species ~ ., data = iris) ## ## Dimension: 4 ## ## Percent Between-Group Variance Explained: ## v1 v2 v3 v4 ## 96.02 98.55 99.90 100.00 ## ## Degrees of Freedom (per dimension): 5 ## ## Training Misclassification Error: 0.02 ( N = 150 ) ## ## Deviance: 15.102 data(glass) # random sample of size 100 samp <- c(1, 3, 4, 11, 12, 13, 14, 16, 17, 18, 19, 20, 27, 28, 31, 38, 42, 46, 47, 48, 49, 52, 53, 54, 55, 57, 62, 63, 64, 65, 67, 68, 69, 70, 72, 73, 78, 79, 83, 84, 85, 87, 91, 92, 94, 99, 100, 106, 107, 108, 111, 112, 113, 115, 118, 121, 123, 124, 125, 126, 129, 131, 133, 136, 139, 142, 143, 145, 147, 152, 153, 156, 159, 160, 161, 164, 165, 166, 168, 169, 171, 172, 173, 174, 175, 177, 178, 181, 182, 185, 188, 189, 192, 195, 197, 203, 205, 211, 212, 214) glass.train <- glass[samp,] glass.test <- glass[-samp,] glass.mda <- mda(Type ~ ., data = glass.train) predict(glass.mda, glass.test, type="post") # abbreviations are allowed confusion(glass.mda,glass.test)
Provide starting weights for the mda
function which
performs discriminant analysis by gaussian mixtures.
mda.start(x, g, subclasses = 3, trace.mda.start = FALSE, start.method = c("kmeans", "lvq"), tries = 5, criterion = c("misclassification", "deviance"), ...)
mda.start(x, g, subclasses = 3, trace.mda.start = FALSE, start.method = c("kmeans", "lvq"), tries = 5, criterion = c("misclassification", "deviance"), ...)
x |
The x data, or an mda object. |
g |
The response vector g. |
subclasses |
number of subclasses per class, as in |
trace.mda.start |
Show results of each iteration. |
start.method |
Either |
tries |
Number of random starts. |
criterion |
By default, classification errors on the training data. Posterior deviance is also an option. |
... |
arguments to be passed to the mda fitter when using posterior deviance. |
A list of weight matrices, one for each class.
Produce a design matrix from a ‘mars’ object.
## S3 method for class 'mars' model.matrix(object, x, which, full = FALSE, ...)
## S3 method for class 'mars' model.matrix(object, x, which, full = FALSE, ...)
object |
a mars object. |
x |
optional argument; if supplied, the mars basis functions are evaluated at these new observations. |
which |
which columns should be used. The default is to use the
columns described by the component |
full |
if |
... |
further arguments to be passed from or to methods. |
A model matrix corresponding to the selected columns.
Fit a smoothing spline to a matrix of responses, single x.
mspline(x, y, w, df = 5, lambda, thresh = 1e-04, ...)
mspline(x, y, w, df = 5, lambda, thresh = 1e-04, ...)
x |
x variable (numeric vector). |
y |
response matrix. |
w |
optional weight vector, defaults to a vector of ones. |
df |
requested degrees of freedom, as in |
lambda |
can provide penalty instead of df. |
thresh |
convergence threshold for df inversion (to lambda). |
... |
holdall for other arguments. |
This function is based on the ingredients of smooth.spline
,
and allows for simultaneous smoothing of multiple responses
A list is returned, with a number of components, only some of which are of interest. These are
lambda |
The value of lambda used (in case df was supplied) |
df |
The df used (in case lambda was supplied) |
s |
A matrix like |
lev |
Self influences (diagonal of smoother matrix) |
Trevor Hastie
x=rnorm(100) y=matrix(rnorm(100*10),100,10) fit=mspline(x,y,df=5)
x=rnorm(100) y=matrix(rnorm(100*10),100,10) fit=mspline(x,y,df=5)
Plot in discriminant (canonical) coordinates a fda
or (by inheritance) a mda
object.
## S3 method for class 'fda' plot(x, data, coords, group, colors, pch, mcolors, mpch, pcex, mcex, ...)
## S3 method for class 'fda' plot(x, data, coords, group, colors, pch, mcolors, mpch, pcex, mcex, ...)
x |
an object of class |
data |
the data to plot in the discriminant coordinates. If
|
coords |
vector of coordinates to plot, with default
|
group |
if |
colors |
a vector of colors to be used in the plotting. |
pch |
a vector of plotting characters. |
mcolors |
a vector of colors for the class centroids; default is |
mpch |
a vector of plotting characters for the centroids. |
pcex |
character expansion factor for the points; defualt is |
mcex |
character expansion factor for the centroids; defualt is |
... |
further arguments to be passed to or from methods. |
data(iris) irisfit <- fda(Species ~ ., data = iris) plot(irisfit) data(ESL.mixture) ## Not a data frame mixture.train=ESL.mixture[c("x","y")] mixfit=mda(y~x, data=mixture.train) plot(mixfit, mixture.train) plot(mixfit, data=ESL.mixture$xnew, group="pred")
data(iris) irisfit <- fda(Species ~ ., data = iris) plot(irisfit) data(ESL.mixture) ## Not a data frame mixture.train=ESL.mixture[c("x","y")] mixfit=mda(y~x, data=mixture.train) plot(mixfit, mixture.train) plot(mixfit, data=ESL.mixture$xnew, group="pred")
Simple minded polynomial regression.
polyreg(x, y, w, degree = 1, monomial = FALSE, ...)
polyreg(x, y, w, degree = 1, monomial = FALSE, ...)
x |
predictor matrix. |
y |
response matrix. |
w |
optional (positive) weights. |
degree |
total degree of polynomial basis (default is 1). |
monomial |
If |
... |
currently not used. |
A polynomial regression fit, containing the essential ingredients for
its predict
method.
Predicted values based on ‘bruto’ additive spline models which are fit by adaptive backfitting.
## S3 method for class 'bruto' predict(object, newdata, type=c("fitted", "terms"), ...)
## S3 method for class 'bruto' predict(object, newdata, type=c("fitted", "terms"), ...)
object |
a fitted bruto object |
newdata |
values at which predictions are to be made. |
type |
if type is |
... |
further arguments to be passed to or from methods. |
Either a fit matrix or a list of fitted terms.
data(trees) fit1 <- bruto(trees[,-3], trees[3]) fitted.terms <- predict(fit1, as.matrix(trees[,-3]), type = "terms") par(mfrow=c(1,2), pty="s") for(tt in fitted.terms) plot(tt, type="l")
data(trees) fit1 <- bruto(trees[,-3], trees[3]) fitted.terms <- predict(fit1, as.matrix(trees[,-3]), type = "terms") par(mfrow=c(1,2), pty="s") for(tt in fitted.terms) plot(tt, type="l")
Classify observations in conjunction with fda
.
## S3 method for class 'fda' predict(object, newdata, type, prior, dimension, ...)
## S3 method for class 'fda' predict(object, newdata, type, prior, dimension, ...)
object |
an object of class |
newdata |
new data at which to make predictions. If missing, the training data is used. |
type |
kind of predictions: |
prior |
the prior probability vector for each class; the default is the training sample proportions. |
dimension |
the dimension of the space to be used, no larger
than the dimension component of |
... |
further arguments to be passed to or from methods. |
An appropriate object depending on type
. object
has a
component fit
which is regression fit produced by the
method
argument to fda
. There should be a
predict
method for this object which is invoked. This method
should itself take as input object
and optionally newdata
.
fda
,
mars
,
bruto
,
polyreg
,
softmax
,
confusion
data(iris) irisfit <- fda(Species ~ ., data = iris) irisfit ## Call: ## fda(x = iris$x, g = iris$g) ## ## Dimension: 2 ## ## Percent Between-Group Variance Explained: ## v1 v2 ## 99.12 100 confusion(predict(irisfit, iris), iris$Species) ## Setosa Versicolor Virginica ## Setosa 50 0 0 ## Versicolor 0 48 1 ## Virginica 0 2 49 ## attr(, "error"): ## [1] 0.02
data(iris) irisfit <- fda(Species ~ ., data = iris) irisfit ## Call: ## fda(x = iris$x, g = iris$g) ## ## Dimension: 2 ## ## Percent Between-Group Variance Explained: ## v1 v2 ## 99.12 100 confusion(predict(irisfit, iris), iris$Species) ## Setosa Versicolor Virginica ## Setosa 50 0 0 ## Versicolor 0 48 1 ## Virginica 0 2 49 ## attr(, "error"): ## [1] 0.02
Predicted values based on ‘mars’ multivariate adaptive regression spline models.
## S3 method for class 'mars' predict(object, newdata, ...)
## S3 method for class 'mars' predict(object, newdata, ...)
object |
an object of class |
newdata |
values at which predictions are to be made. |
... |
further arguments to be passed to or from methods. |
the fitted values.
mars
,
predict
,
model.matrix.mars
Classify observations in conjunction with mda
.
## S3 method for class 'mda' predict(object, newdata, type, prior, dimension, g, ...)
## S3 method for class 'mda' predict(object, newdata, type, prior, dimension, g, ...)
object |
a fitted mda object. |
newdata |
new data at which to make predictions. If missing, the training data is used. |
type |
kind of predictions: |
prior |
the prior probability vector for each class; the default is the training sample proportions. |
dimension |
the dimension of the space to be used, no larger
than the dimension component of |
g |
??? |
... |
further arguments to be passed to or from methods. |
An appropriate object depending on type
. object
has a
component fit
which is regression fit produced by the
method
argument to mda
. There should be a
predict
method for this object which is invoked. This method
should itself take as input object
and optionally newdata
.
mda
,
fda
,
mars
,
bruto
,
polyreg
,
softmax
,
confusion
data(glass) samp <- sample(1:nrow(glass), 100) glass.train <- glass[samp,] glass.test <- glass[-samp,] glass.mda <- mda(Type ~ ., data = glass.train) predict(glass.mda, glass.test, type = "post") # abbreviations are allowed confusion(glass.mda, glass.test)
data(glass) samp <- sample(1:nrow(glass), 100) glass.train <- glass[samp,] glass.test <- glass[-samp,] glass.mda <- mda(Type ~ ., data = glass.train) predict(glass.mda, glass.test, type = "post") # abbreviations are allowed confusion(glass.mda, glass.test)
Find the maximum in each row of a matrix.
softmax(x, gap = FALSE)
softmax(x, gap = FALSE)
x |
a numeric matrix. |
gap |
if |
A factor with levels the column labels of x
and values the
columns corresponding to the maximum column. If gap = TRUE
a
list is returned, the second component of which is the difference
between the largest and next largest column of x
.
predict.fda
,
confusion
,
fda
mda
data(iris) irisfit <- fda(Species ~ ., data = iris) posteriors <- predict(irisfit, type = "post") confusion(softmax(posteriors), iris[, "Species"])
data(iris) irisfit <- fda(Species ~ ., data = iris) posteriors <- predict(irisfit, type = "post") confusion(softmax(posteriors), iris[, "Species"])