R/build_forest_predict.R

In MultivariateRandomForest: Models Multivariate Cases Using Random Forests

build_forest_predict <- function(trainX, trainY, n_tree, m_feature, min_leaf, testX){
  if (class(n_tree)=="character" || n_tree%%1!=0 || n_tree<1) stop('Number of trees in the forest can not be fractional or negative integer or string')
  if (class(m_feature)=="character" || m_feature%%1!=0 || m_feature<1) stop('Number of randomly selected features considered for a split can not be fractional or negative integer or string')
  if (class(min_leaf)=="character" || min_leaf%%1!=0 || min_leaf<1 || min_leaf>nrow(trainX)) stop('Minimum leaf number can not be fractional or negative integer or string or greater than number of samples')
  if (ncol(trainX)!=ncol(testX) || nrow(trainX)!=nrow(trainY)) stop('Data size is inconsistant')

  theta <- function(trainX){trainX}
  results <- bootstrap::bootstrap(1:nrow(trainX),n_tree,theta)
  b=results$thetastar

  Variable_number=ncol(trainY)
  if (Variable_number>1){
    Command=2
  }else if(Variable_number==1){
    Command=1
  }

  Y_HAT=matrix(  0*(1:Variable_number*nrow(testX)),  ncol=Variable_number,   nrow=nrow(testX)  )
  Y_pred=NULL

  for (i in 1:n_tree){
    Single_Model=NULL
    X=trainX[ b[ ,i],  ]
    Y=matrix(trainY[ b[ ,i],  ],ncol=Variable_number)
    Inv_Cov_Y = solve(cov(Y)) # calculate the V inverse
    if (Command==1){
      Inv_Cov_Y=matrix(rep(0,4),ncol=2)
    }
    Single_Model=build_single_tree(X, Y, m_feature, min_leaf,Inv_Cov_Y,Command)
    Y_pred=single_tree_prediction(Single_Model,testX,Variable_number)
    for (j in 1:Variable_number){
      Y_HAT[,j]=Y_HAT[,j]+Y_pred[,j]
    }
  }
  Y_HAT=Y_HAT/n_tree
  return(Y_HAT)
}

# build_forest_predict <- function(trainX, trainY, n_tree, m_feature, min_leaf, testX){
#   if (class(n_tree)=="character" || n_tree%%1!=0 || n_tree<1) stop('Number of trees in the forest can not be fractional or negative integer or string')
#   if (class(m_feature)=="character" || m_feature%%1!=0 || m_feature<1) stop('Number of randomly selected features considered for a split can not be fractional or negative integer or string')
#   if (class(min_leaf)=="character" || min_leaf%%1!=0 || min_leaf<1 || min_leaf>nrow(trainX)) stop('Minimum leaf number can not be fractional or negative integer or string or greater than number of samples')
#   if (ncol(trainX)!=ncol(testX) || nrow(trainX)!=nrow(trainY)) stop('Data size is inconsistant')
#
#   theta <- function(trainX){trainX}
#   results <- bootstrap::bootstrap(1:nrow(trainX),n_tree,theta)
#   b=results$thetastar
#
#   Variable_number=ncol(trainY)
#   if (Variable_number>1){
#     Command=2
#   }else if(Variable_number==1){
#     Command=1
#   }
#
#   Y_HAT=matrix(  0*(1:Variable_number*nrow(testX)),  ncol=Variable_number,   nrow=nrow(testX)  )
#   Y_pred=rep(NULL,n_tree)
#   ##############################
#     for (i in 1:n_tree){
#       Single_Model=NULL
#       X=trainX[ b[ ,i],  ]
#       Y=matrix(trainY[ b[ ,i],  ],ncol=Variable_number)
#       Inv_Cov_Y = solve(cov(Y)) # calculate the V inverse
#       if (Command==1){
#         Inv_Cov_Y=matrix(rep(0,4),ncol=2)
#       }
#       Single_Model=build_single_tree(X, Y, m_feature, min_leaf,Inv_Cov_Y,Command)
#       Y_pred=single_tree_prediction(Single_Model,testX,Variable_number)
#       for (j in 1:Variable_number){
#         Y_HAT[,j]=Y_HAT[,j]+Y_pred[,j]
#       }
#     }
#     Y_HAT=Y_HAT/n_tree
#
#   ######################
#
#   no_cores <- parallel::detectCores() - 1
#   cl<-parallel::makeCluster(no_cores)
#   # doParallel::registerDoParallel(cl)
#
#   # Y_pred <- foreach::foreach(i=1:n_tree, .combine = list, .multicombine = TRUE) %dopar% {
#   #   Single_Model=NULL
#   #   X=trainX[ b[ ,i],  ]
#   #   Y=matrix(trainY[ b[ ,i],  ],ncol=Variable_number)
#   #   if (Command==1){
#   #     Inv_Cov_Y=matrix(rep(0,4),ncol=2) # for RF, covariance is not needed, it is 2x2 zeros matrix
#   #   }else if (Command==2){
#   #     Inv_Cov_Y = solve(cov(Y)) # calculate covariance among the output features for MRF
#   #   }
#   #   Single_Model=build_single_tree(X, Y, m_feature, min_leaf,Inv_Cov_Y,Command)
#   #   Y_pred3=single_tree_prediction(Single_Model,testX,Variable_number)
#   #   Y_pred3
#   # }
#   # Y_HAT=plyr::aaply(plyr::laply(Y_pred, as.matrix), c(2, 3), mean)
#
#   parallel::stopCluster(cl)
#   return(Y_HAT)
# }