model.frame.default(Terms, newdata, na.action = na.action, xlev = object$xlevels) 中的错误：对象不是矩阵

如何解决model.frame.default(Terms, newdata, na.action = na.action, xlev = object$xlevels) 中的错误：对象不是矩阵

我正在研究定性数据的分类模型 (SVM)，数据已下载到 UCI 存储库中。这是所用数据集的 link，它已被拆分为训练集和测试集，我的问题是我已经拟合了一个模型，但我无法在测试数据集上预测我的模型 (review_ted_model)。任何人都可以帮助我解决错误，因为我受够了，好吗？

这就是我遇到的问题。

```
predict_svm_rbf <- predict(review_ted_model,newdata = test)
Error in model.frame.default(Terms,newdata,na.action = na.action,xlev = object$xlevels) : 
  object is not a matrix
In addition: Warning message:
'newdata' had 53766 rows but variable found had 1 row 
 ```

代码的输出太多了，我无法复制到这里。 这些是已经尝试过的代码，除了上面已经说明的最后一行代码外，它们运行得很好。

library(caret)
library(mlbench)
library(tidyverse)
library(tidytext)
library(data.table)
train=as.data.frame(fread("drugsComTrain_raw.tsv"))
test=as.data.frame(fread("drugsComTest_raw.tsv"))
train %>% head()
summary(train)
str(train)
nrow(train)
train[duplicated(train)|duplicated(train,fromLast = T),]
sapply(train,function(x) sum(is.na(x)))

#Finding the total count of missing values.
sum(!complete.cases(train))
head(train)

train  <- train %>% mutate(document = row_number())
train$condition <- as.character(train$condition)
train$review <- as.character(train$review)
train$drugName <- as.character(train$drugName)
summary(train)
str(train)


trainText <- train %>% 
  unnest_tokens(word,review) %>% 
  anti_join(stop_words)


length(unique(trainText$drugName))


unique(trainText$drugName)
abc = trainText[which(trainText$drugName %in% c("Methadose","Didrex","Forteo","Kava")),]
abc

abc %>% 
  count(word,drugName,sort = TRUE) %>% 
  mutate(drugName = factor(drugName),word = as.factor(word),word = fct_reorder(word,n)
  ) %>% 
  group_by(drugName) %>% 
  top_n(10)  %>% 
  ungroup() %>% 
  ggplot(aes(word,n)) +
  geom_col(aes(fill = drugName)) +
  facet_wrap(~drugName,scales = "free") +
  coord_flip()

View(abc)
abc 
library(caret)
library(LiblineaR)
review_ted_model <- train(condition ~ word,data = abc,method = "svmLinear3")

plot(review_ted_model)

View(test)
predict_svm_rbf <- predict(review_ted_model,newdata = test)