Hierarchical clustering using cosine distance in R

问题

I want to do hierarchical clustering by using cosine similarity with the R programming language for corpus of documents, but I got the following error:

Error in if (is.na(n) || n > 65536L) stop("size cannot be NA nor exceed 65536") : missing value where TRUE/FALSE needed

What should I do?

To reproduce it, here's an example:

library(tm)
doc <- c( "The sky is blue.", "The sun is bright today.", "The sun in the sky is bright.", "We can see the shining sun, the bright sun." )
doc_corpus <- Corpus( VectorSource(doc) )
control_list <- list(removePunctuation = TRUE, stopwords = TRUE, tolower = TRUE)
tdm <- TermDocumentMatrix(doc_corpus, control = control_list)



tf <- as.matrix(tdm)
( idf <- log( ncol(tf) / ( 1 + rowSums(tf != 0) ) ) )
( idf <- diag(idf) )
tf_idf <- crossprod(tf, idf)
colnames(tf_idf) <- rownames(tf)

tf_idf

cosine_dist = 1-crossprod(tf_idf) /(sqrt(colSums(tf_idf^2)%*%t(colSums(tf_idf^2))))
cluster1 <- hclust(cosine_dist, method = "ward.D")

Then I get the error:

Error in if (is.na(n) || n > 65536L) stop("size cannot be NA nor exceed 65536") : missing value where TRUE/FALSE needed

回答1:

There are 2 issues:

1: cosine_dist = 1-crossprod(tf_idf) /(sqrt(colSums(tf_idf^2)%*%t(colSums(tf_idf^2)))) creates NaN's because you divide by 0.

2: hclust needs a dist object, not just a matrix. See ?hclust for more details

Both can be solved with the following code:

.....
cosine_dist = 1-crossprod(tf_idf) /(sqrt(colSums(tf_idf^2)%*%t(colSums(tf_idf^2))))

# remove NaN's by 0
cosine_dist[is.na(cosine_dist)] <- 0

# create dist object
cosine_dist <- as.dist(cosine_dist)

cluster1 <- hclust(cosine_dist, method = "ward.D")

plot(cluster1)

来源：https://stackoverflow.com/questions/52391558/hierarchical-clustering-using-cosine-distance-in-r

标签

hierarchical-clustering