Question

我创建了一个NaiveBayes模型，我认为该模型不存在（无法创建），因此不能用于预测。我想知道它会如何以及在预测中该如何处理（如何识别这种模型？）

这是我运行NaiveBayes之后得到的模型：

> NB_TRAIN_model[[24,1]]

Naive Bayes Classifier for Discrete Predictors

Call:
naiveBayes.default(x = X, y = Y, laplace = laplace)

A-priori probabilities:
Y
F I M 
1 0 0 

Conditional probabilities:
   V2
Y    [,1] [,2]
  F 0.545   NA
  I    NA   NA
  M    NA   NA

   V3
Y    [,1] [,2]
  F 0.445   NA
  I    NA   NA
  M    NA   NA

   V4
Y   [,1] [,2]
  F 0.15   NA
  I   NA   NA
  M   NA   NA

   V5
Y   [,1] [,2]
  F  0.8   NA
  I   NA   NA
  M   NA   NA

   V6
Y     [,1] [,2]
  F 0.3535   NA
  I     NA   NA
  M     NA   NA

   V7
Y    [,1] [,2]
  F 0.163   NA
  I    NA   NA
  M    NA   NA

   V8
Y    [,1] [,2]
  F 0.207   NA
  I    NA   NA
  M    NA   NA

   V9
Y   [,1] [,2]
  F    9   NA
  I   NA   NA
  M   NA   NA

当我尝试根据新数据进行预测时，我得到以下信息：

Error in a[i, j] <- predict(NB_TRAIN_model[[transnames[[i, j]], j]], test[i,  : 
  replacement has length zero

如何在预测模型是否存在之前对其进行检查？

这是我得到模型的方式：

require(party)
require (data.table)
require (e1071)
require (randomForest)
dat1 <- fread('https://archive.ics.uci.edu/ml/machine-learning-databases/abalone/abalone.data',stringsAsFactors=T)

## split data to train and test
set.seed(123)
dat1 <- subset(dat1, !is.na(V1))
smp_size<-100
train_ind <- sample(seq_len(nrow(dat1)), size = smp_size)
train <- dat1[train_ind, ]
test <- dat1[-train_ind, ]

#dat1$V1<-as.factor(dat1$V1)
rf <- randomForest(V1 ~ ., data = train, ntree = 10, keep.inbag = TRUE)

rf_train<-predict(rf,train[,V2:V9], nodes=TRUE)
train_nodes<-attr(rf_train,"nodes")

rf_test<-predict(rf,test[,V2:V9], nodes=TRUE)
test_nodes<-attr(rf_test,"nodes")

### each b[i] holds the rows for each terminal node for tree no. i
b<-list()

max_no_terminalnodes<- length (split(train, train_nodes[,1]) )
for (i in 1:(rf$ntree))
{
  b[[i]]<-split(train, train_nodes[,i])
  if ( max_no_terminalnodes < length (split(train, train_nodes[,i]) ) ) {
    max_no_terminalnodes <- length (split(train, train_nodes[,i]) ) }
}
### Holds the naiveBayes models per terminal node per tree 
#NB_TRAIN_model<-list()
NB_TRAIN_model<-matrix(list(), nrow=max_no_terminalnodes, ncol=rf$ntree)
NAMETRANS<-matrix(list(), nrow=max_no_terminalnodes, ncol=rf$ntree)
NAMETRANS[] <- 0L


k=0
for (i in 1:(rf$ntree))
{ 
  for (j in as.numeric (names( b[[i]]))){
    k=k+1
  NB_TRAIN_model[[k,i]]<- naiveBayes(as.factor(V1) ~ ., data =  b[[i]][[k]] )
  NAMETRANS[k,i]<-j

   }
  k=0
}
### Find for each row and tree what NaiveBayes model to run (based on the terminal node NB model per tree)
transnames<-matrix(list(), nrow=max_no_terminalnodes, ncol=rf$ntree)

for(j in 1:ncol(transnames)){
  for(i in 1:nrow(transnames)){

    tryCatch(transnames[i,j] <- which(NAMETRANS[,j]==test_nodes[i,j]),
             error = function(e) return(NA)) }
  }


a<-matrix(list(), nrow=nrow(test), ncol=rf$ntree)
for(j in 1:ncol(transnames)){
  for(i in 1:nrow(transnames)){
   a[i,j]<-  predict (NB_TRAIN_model[[transnames[[i, j]],j]],test[i,V2:V9],type = "class") 
  }
}

由于模型存在问题，R中的NaiveBayes无法预测

0 个答案: