Question

我正在尝试将POS标记为从Yelp数据集质询（第11回合）下载的评论文本的文本。下面列出了我编写的功能：

r <- unlist(lapply(rec_texts, function(x) { str_split(x, "\n") }))

R_tagPOS <-  function(x, ...) {
r <- as.String(x)
word_token_annotator <- Maxent_Word_Token_Annotator()
a2 <- Annotation(1L, "sentence", 1L, nchar(r))
a2 <- annotate(r, word_token_annotator, a2)
a3 <- annotate(r, Maxent_POS_Tag_Annotator(), a2)
a3w <- a3[a3$type == "word"]
POStags <- unlist(lapply(a3w$features, `[[`, "POS"))
POStagged <- paste(sprintf("%s/%s", s[a3w], POStags), collapse = " ")
list(POStagged = POStagged, POStags = POStags)
}

POS_R <- lapply(r, R_tagPOS)

但是它返回如下结果：

Error in e(s, a) : no word token annotations found 
8.stop("no word token annotations found") 
7. e(s, a) 
6.paste(y$id, y$type, y$start, y$end, sep = "\r") 
5.merge.Annotation(a, e(s, a)) 
4.merge(a, e(s, a)) 
3.annotate(r, Maxent_POS_Tag_Annotator(), a2) 
2.FUN(X[[i]], ...) 
1.lapply(r, R_tagPOS)

rec_texts是Yelp数据集质询中的评论文本。

Classes 'tbl_df', 'tbl' and 'data.frame':   2000 obs. of  1 variable:
$ text: chr  "Mix is such a beautiful restaurant and you have a gorgeous 
view of the strip while you're enjoying your dinner!"| 
__truncated__"WooooooHooooo 100th review!!!  I eat here all the time, 
but I've gotta take the chance to be the 100th review.\"| __truncated__ 
"My bf and I came for a small snack and some brews. We got the pretzel 
and cheese appetizer. The pretzel was ver"| __truncated__ "The food was 
amazing! This is a must for me now when I come visit Las Vegas! Everyone 
was very friendly. The ow"| __truncated__ ...

请问是否有人遇到过这种情况，或者有人知道如何解决此问题？我认为这与数据有关。因为此代码对第三方提供的另一个审阅文本数据集工作正常。

Answer 1

我收到了相同的错误消息，在我的情况下，我的字符串向量中有一个空元素。

一旦我从向量中删除了所有空元素，它就可以正常工作。

R-openNLP找不到单词标记注释

1 个答案: