Prediction on Test Data for Super Greedy Forests

Obtain predicted values on test data using a trained super greedy forest.

predict(object, newdata,  get.tree = NULL,
     block.size = 10, seed = NULL, do.trace = FALSE,...)

Arguments

object

rfsgt object obtained from previous training call using rfsgt.

newdata

Test data. If not provided the training data is used and the original training forest is restored.

get.tree

Vector of integer(s) identifying trees over which the ensemble is calculated over. By default, uses all trees in the forest.

block.size

Determines how cumulative error rate is calculated. To obtain the cumulative error rate on every nth tree, set the value to an integer between 1 and ntree.

seed

Negative integer specifying seed for the random number generator.

do.trace

Number of seconds between updates to the user on approximate time to completion.

...

Additional options.

Details

Returns the predicted values for a super greedy forest.

Author

Hemant Ishwaran and Udaya B. Kogalur

References

Ishwaran H. (2023). Super greedy regression trees with coordinate descent. Technical Report.

See also

rfsgt

Examples

# \donttest{
## ------------------------------------------------------------
##
## train/test using friedman 3
##
## ------------------------------------------------------------

## train sgf on friedman 3
d.trn <- data.frame(mlbench:::mlbench.friedman3(500))
o <- rfsgt(y~.,d.trn, hcut=1)
print(o)

## test sgf
d.tst <- data.frame(mlbench:::mlbench.friedman3(1000))
y.tst <- d.tst$y
x.tst <- d.tst[, colnames(d.tst)!= "y"]
yhat <- predict(o, x.tst)$predicted
cat("test set mse:", mean((yhat - y.tst)^2), "\n")

## ------------------------------------------------------------
##
## restore a trained super greedy forest using boston
##
## ------------------------------------------------------------

## run sgf on boston
data(BostonHousing, package = "mlbench")
o <- rfsgt(medv~., BostonHousing)
print(o)

## restore the forest
print(predict(o))

## ------------------------------------------------------------
##
## coherence check using boston housing with factors
##
## ------------------------------------------------------------

## boston housing data: make factors
data(BostonHousing, package = "mlbench")
Boston <- BostonHousing[1:40,]
Boston$zn <- factor(Boston$zn)
Boston$chas <- factor(Boston$chas)
Boston$lstat <- factor(round(0.2 * Boston$lstat))
Boston$nox <- factor(round(20 * Boston$nox))
Boston$rm <- factor(round(Boston$rm))
     
## grow a single tree - save inbag information
o <- rfsgt(medv~., Boston, hcut=2, filter=FALSE, ntree=1, membership=TRUE, nodesize=3)

## coherence matrix
pred <- data.frame(
      inbag=o$inbag,
      pred.inb=o$predicted,
      pred.oob=o$predicted.oob,
      pred.inb.restore=predict(o)$predicted,
      pred.oob.restore=predict(o)$predicted.oob,
      pred.test=predict(o,Boston)$predicted)
print(pred)

## coherence check
cat("coherence for inbag data:", sum(pred$pred.inb-pred$pred.test,na.rm=TRUE)==0, "\n")
cat("  coherence for oob data:", sum(pred$pred.oob-pred$pred.test,na.rm=TRUE)==0, "\n")


## canonical example of train/test with prediction
trn <- sample(1:nrow(Boston), nrow(Boston)/2, replace=FALSE)
o.trn <- rfsgt(medv~.,Boston[trn,],hcut=2)
predict(o.trn,Boston[-trn,])


## ------------------------------------------------------------
## prediction using tuning hcut and pre-filtering with tune.hcut 
## ------------------------------------------------------------

## fit the forest to the tuned hcut
dta <- data.frame(mlbench:::mlbench.friedman3(500))
f <- tune.hcut(y~., dta, hcut=5, verbose=TRUE)
o <- rfsgt(y~., dta, filter=f)
print(o)

## test the tuned forest on new data
print(predict(o, data.frame(mlbench:::mlbench.friedman3(25000))))

## over-ride the optimized hcut
o2 <- rfsgt(y~., dta, filter=use.tune.hcut(f, hcut=2))
print(o2)
print(predict(o2, data.frame(mlbench:::mlbench.friedman3(25000))))


# }