power_ranger: Cascade Forest's implementation in R
In ablanda/deepForest:

Description Usage Arguments Details Author(s) Examples

This function attempts to replicate Cascade Forest in original way of paper by package ranger.

power_ranger(x = NULL, y, training_frame, validation_frame = NULL,
  num.trees = 200, pmtry = NULL, n_forest = 4, random_forest = 2,
  eta = FALSE, num.threads = 30, work.dir = getwd(), early.stop = 10,
  continue = 0, write.forest = TRUE, save.memory = FALSE, id = NULL,
  k = NULL)

`x`	A vector containing the names or indices of the predictor variables to use in building the model. If x is missing,then all columns except y are used.
`y`	The name of the response variable in the model.If the data does not contain a header, this is the column index number starting at 0, and increasing from left to right. (The response must be either an integer or a categorical variable).
`training_frame`	Training data of class `data.frame` or `matrix`.
`validation_frame`	Validation data.
`num.trees`	Number of trees.
`pmtry`	Percentage of variables to possibly split at in each node. Default is the (rounded down) square root of the number variables divided by total numbers.
`n_forest`	Total forest number for every layer.
`random_forest`	Number of Random forest.
`num.threads`	Number of threads.
`work.dir`	Type: character. When `out.of.memory == TRUE`, the four models of each layers will be saved to disk in file `df1:early.stop.RData`. If you don't provide a working directory, the models will be saved inside that directory.
`early.stop`	Number of layers.
`continue`	It's used for prediction if `n_forest==continue`. Else if `n_forest>continue` for add layers to Cascade Forest.
`write.forest`	Save `ranger.forest` object, required for prediction. Set to `FALSE` to reduce memory usage if no prediction intended.
`save.memory`	Use memory saving (but slower) splitting mode. Warning: This option slows down the tree growing, use only if you encounter memory problems.
`id`	prefix of saved files
`k`	number of k-folds, if is NULL (default) use OOB

For implementation details of Cascade Forest: https://arxiv.org/pdf/1702.08835.pdf.

Blanda Alessandro

## Not run: 
rm(list=ls())
# Load libraries
library(devtools)
install_github( 'ablanda/deepForest ')
library(deepForest)
# Download MNIST data here: \url{https://pjreddie.com/projects/mnist-in-csv/}
dati<-read.csv('mnist_train.csv',header=F)
dativ<-read.csv('mnist_test.csv',header=F)

dati[,1]<-as.factor(dati[,1])
dativ[,1]<-as.factor(dativ[,1])
m<-power_ranger(y=1,training_frame = dati[1:100,],validation_frame = dativ[1:100,],n_forest=8,random_forest = 4,early.stop=4,k=3)
pred<-matrix(0,nrow(dativ),nlevels(dati[,1])-1)
for(h in 1:(nlevels(dati[,1])-1)){
pred_level<-sapply(1:8,function(j) m$pred_val[[1]][[j]][,h])
 pred[,h] <-rowMeans(pred_level)}

## End(Not run)