(函数(classes,fdef,mtable)中的错误

时间:2014-10-04 18:51:26

标签: topic-modeling

我在两个月前运行了这个主题建模脚本SUCCESSFULLY,但它突然给了我一条错误信息(在最后三行)。

post <- posterior(TM1, newdata = dtm[-c(1:20),]) #this script gives me an error message.
perplex <- perplexity(TM1, newdata = dtm[-c(1:20),]) #this script does not give me an error message. 

有人可以帮我解决这里发生的事吗?请~~

=====================

library("tm")

library("slam")

library("topicmodels")

library("SnowballC")

corpus <- Corpus(DirSource(directory="/Users/loni/Documents/TextMining/test", encoding="UTF-8"))

dtm <- DocumentTermMatrix(corpus, control=list(stemming=TRUE, stopwords=TRUE, removePunctuation=FALSE))
term_tfidf <- tapply(dtm$v/row_sums(dtm)[dtm$i], dtm$j, mean) * log2(nDocs(dtm)/col_sums(dtm>0))

dim(dtm)

[1] 26 919

dtm <- dtm[, term_tfidf >= .06] # petition corpus

dtm <- dtm[row_sums(dtm) > 0,]

dim(dtm)

[1] 26 499

k<-5

SEED <- 2

TM <- list(VEM = LDA(dtm, k = k, control = list(seed = SEED))) 

TM1 <- list(VEM = LDA(dtm[c(1:20),], k = k, control = list(seed = SEED))) #validation
Topic <- topics(TM[["VEM"]],1)
Terms <- terms(TM[["VEM"]], 8)
Terms[, 1:5]
post <- posterior(TM1, newdata = dtm[-c(1:20),])

(函数(classes,fdef,mtable)中的错误:   无法为签名'“list”,“DocumentTermMatrix”'

找到函数'后验'的继承方法

1 个答案:

答案 0 :(得分:0)

可能是因为列表索引错误。在TM1上尝试[[]]或[]