Tree Regression Dialog Functions

DESCRIPTION:

These functions are used by the Tree Regression dialog. menuTree calls tabSummary.tree, tabPlot.tree, tabPrune.tree, tabShrink.tree , and tabPredict.tree if summary, plotting, pruning, shrinking, or prediction results are requested.

USAGE:

menuTree(formula, data, weights, subset, na.omit.p=T, mincut=5, 
         minsize=10, mindev=0.01, print.summary.p=T, print.tree.p=F, 
         save.name=NULL, saveRsdMisclass=F, saveRsdPearson=F,  
         saveRsdDeviance=F, plot.it=T, plotUniform=F, plot.addText=T,  
         plot.addText.what="yval", prune.p=F, prune.k=NULL,  
         prune.best=NULL, reduced.newdata=NULL, prune.method="deviance",  
         shrink.p=F, shrink.k=F, reduced.saved.tree=NULL, plot.seq=F,  
         predict.newdata=NULL, predict.type="vector", predict.save.name= 
         NULL) 
tabSummary.tree(treeobj, print.summary.p=T, print.tree.p=F, save.name= 
         NULL, saveRsdMisclass=F, saveRsdPearson=F, saveRsdDeviance=F) 
tabPlot.tree(treeobj, plot.it=T, plotUniform=F, plot.addText=T,  
         plot.addText.what="yval") 
tabPrune.tree(treeobj, prune.k=NULL, prune.best=NULL, reduced.newdata= 
         NULL, prune.method="deviance", plot.seq=F, reduced.saved.tree, 
         prune.p=T) 
tabShrink.tree(treeobj, shrink.k, reduced.newdata=NULL, plot.seq,  
         reduced.saved.tree, shrink.p=T) 
tabPredict.tree(object, newdata=NULL, predict.type="vector", save.name) 

REQUIRED ARGUMENTS:

formula
a formula object, with the response on the left of a `~' operator, and the terms, separated by + operators, on the right.
treeobj
an object that inherits from class tree.

OPTIONAL ARGUMENTS:

data
a data.frame in which to interpret the variables named in the formula, or in the subset and the weights argument. If this is missing, then the variables in the formula should be on the search list.
weights
vector of observation weights. The length of weights must be the same as the number of observations. The weights must be nonnegative and it is strongly recommended that they be strictly positive, since zero weights are ambiguous, compared to use of the subset argument.
subset
expression saying which subset of the rows of the data should be used in the fit. This can be a logical vector (which is replicated to have length equal to the number of observations), or a numeric vector indicating which observation numbers are to be included, or a character vector of the row names to be included. All observations are included by default.
na.omit.p
if TRUE, then any observation with missing values are removed from the analysis. If FALSE and there are missing values then the function will exit with a message that missing values are not allowed. If na.omit.p is TRUE then na.action is set to na.omit in the call to tree. If na.omit.p is FALSE then na.action is set to na.fail in the call to tree.
mincut
minimum number of observations to include before the first cut on a variable. See the help file for tree.control for more details on this and other fitting options.
minsize
node size at which the last split is performed, i.e. minsize == 5 means that growing continues if there are at least 5 observations in a node. See the help file for tree.control for more details on this and other fitting options.
mindev
minimum node deviance before growing stops. See the help file for tree.control for more details on this and other fitting options.
print.summary.p
if TRUE, a short summary of the tree model fitted is printed. This output is from the function summary.tree.
print.tree.p
if TRUE, the full fitted tree is printed. This output is from the function print.tree.
save.name
a character string for the name of the data frame to save the fit and residuals in. If data frame with this name already exists in database 1 and it has the appropriate number of rows then the saved values will be appended to the data frame. If the object already exist in database 1 and it is not a data frame or it does not have the appropriate number of rows then a new name is created by appending a number to save.name and the results are saved in the data frame with the new name.
saveRsdMisclass
if TRUE, the misclassification errors are returned in save.name..
saveRsdPearson
if TRUE, the Pearson residuals from the regression are saved in the data frame save.name.
saveRsdDeviance
if TRUE, residuals of type "deviance" are returned in save.name.
plot.it
if TRUE, a plot of the fitted tree will be produced.
plotUniform
if TRUE, the plot of the tree will have branches uniformily sized. The default is to plot the branches with sizes proportional to node deviance.
plot.addText
if TRUE, text labels will be added at each node on the dendrogram.
plot.addText.what
select what the labels at each node should be. Choices are "Response Value" (default), "Node Size", or "Node Deviance".
prune.p
if TRUE, the resulting tree will be pruned. The result will then be saved in an object named with the string given in reduced.saved.tree.
prune.k
cost-complexity parameter defining either a specific subtree of tree ( prune.k a scalar) or the (optional) sequence of subtrees minimizing the cost-complexity measure ( prune.k a vector). If missing, prune.k is determined algorithmically.
prune.best
integer specifying the size (number of terminal nodes) of a specific subtree in the cost-complexity sequence to be returned. This is an alternative way to select a subtree than by supplying a scalar cost- complexity parameter prune.k.
reduced.newdata
the name of a data frame containing values are which predictions are required. The sequence of optimally reduced (shrunken or pruned) trees is evaluated on newdata. If missing, the data used to grow the tree are used.
prune.method
select one of two possibilities, "deviance" or "misclass". It denotes the measure of node heterogeneity used to guide cost-complexity pruning.
shrink.p
if TRUE, the resulting tree will be shrunken. The result will then be saved in an object named with the string given in reduced.saved.tree.
shrink.k
the value of the shrinkage parameter of each tree in the sequence.
reduced.saved.tree
a character string. Results from pruning or shrinking a tree will be saved in an object with that name, or a derivative of the string, if an object with that name already exists. It defaults to "last.pruned" for pruned trees, or "last.shrunken" for shrunken ones.
plot.seq
if TRUE, a plot of the optimal tree sequence that is obtained for the various values given as shrinkage scalars or pruning tree sizes. It displays the value of the deviance for each reduced tree.
predict.newdata
a data frame to use for computing predictions. It must contain the same names as the terms in the right side of the formula for the model. If missing, the predictions for the original data are computed.
predict.type
selection that denotes whether the predictions are to be returned as a vector (default) or as a tree object.
predict.save.name
a character string for the name of the object to save the predictions in. If an object with this name already exists in database 1 another name will be generated by appending an integer to the string given.

VALUE:

an object of class tree. See the tree.object help file for details.

SEE ALSO:

, , , , , , , .