How to export rStudio analysis detail - r

Question

working on a document classification model using the code provided by Tim DAuria (https://www.youtube.com/watch?v=j1V2McKbkLo), but can not figure out how to actually see the detail analysis of the classification on the 'Test' data.
I am using the model to help classify contracts by type and want to see the specific classification assigned to the different contracts. For example, the model assigns 15 contracts as "x" type of contract. How can I view those 15 file names? The code below works great for the classification piece. Just posting it for reference.
Please help! Really new at this and I'm sure I'm missing something obvious but could not find anything on the web.
Classification Code below:
#int
libs <- c("tm", "plyr","class")
lapply(libs, require, character.only = TRUE)
#set options
options(stringsAsFactors = FALSE)
#set parameters
contract <- c("build construction", "other")
pathname <- ..Desktop/ML/ContractReview"
#clean text
cleanCorpus <- function(corpus) {
corpus.tmp <- tm_map(corpus, removePunctuation)
corpus.tmp <- tm_map(corpus.tmp, stripWhitespace)
corpus.tmp <- tm_map(corpus.tmp, content_transformer(tolower))
corpus.tmp <- tm_map(corpus.tmp, removeWords, stopwords("english"))
corpus.tmp <- tm_map(corpus.tmp, stemDocument)
return(corpus.tmp)
}
#build TDM
generateTDM <- function(contract, path) {
c.dir <- sprintf ("%s/%s", path, contract)
c.cor <- VCorpus(DirSource(directory = c.dir), readerControl = list(reader=readPlain))
c.cor.cl <- cleanCorpus(c.cor)
c.tdm <- TermDocumentMatrix(c.cor.cl)
c.tdm <- removeSparseTerms(c.tdm, .07)
result <- list(name = contract, tdm = c.tdm)
}
tdm <- lapply(contract, generateTDM, path = pathname)
# attach name
bindcontractToTDM <- function(tdm) {
c.mat <-t(data.matrix(tdm[["tdm"]]))
c.df <- as.data.frame(c.mat, stringsAsFactors = FALSE)
c.df <- cbind(c.df, rep(tdm[["name"]], nrow(c.df)))
colnames(c.df) [ncol(c.df)] <- "targetcontract"
return(c.df)
}
contractTDM <- lapply(tdm, bindcontractToTDM)
#stack if you have more than one dataframe
tdm.stack <- do.call(rbind.fill, contractTDM)
tdm.stack[is.na(tdm.stack)] <-0
#hold-out
train.idx <- sample(nrow(tdm.stack), ceiling(nrow(tdm.stack)* 0.7))
test.idx <- (1:nrow(tdm.stack))[- train.idx]
#model - knn
tdm.contract <-tdm.stack[, "targetcontract"]
tdm.stack.nl <- tdm.stack[, !colnames(tdm.stack) %in% "targetcontract"]
knn.pred <- knn(tdm.stack.nl[train.idx, ], tdm.stack.nl[test.idx, ], tdm.contract[train.idx])
#accuracy
conf.mat <- table("predictions"= knn.pred, Actual = tdm.contract[test.idx])
(accuracy <- sum(diag(conf.mat)) / length(test.idx)*100)

Answer 1

The answer to your question is hidden in the knn.pred object, which stores predicted labels for the test cases.
Since there is no input data provided (I don't think it was provided by the author of the video?), I am not sure about the details of the contract classes from your example. However, the output of the knn function from the class package you are using is a factor with labels that the algorithm predicted for the test documents
(you will notice that
length(knn.pred)
is equal to 0.3 * nrow(tdm.stack) ).
If you would like to view/store the predicted label and the actual label for each entry, then you can create a suitable data frame:
label_df <- cbind(label_pred = knn.pred, label_actual = tdm.contract[test.idx])
Alternatively, you can also choose to include the remaining columns of the tdm.contract (if you would like to re-examine the tdm.stack information in that context)
label_df <- cbind(label_pred = knn.pred, label_actual = tdm.contract[test.idx], tdm.stack.nl[test.idx,])
You can then filter either of these data frames to see how your entries of interest have been labelled.
Alternatively, you can choose to run k nearest neighbors algorithm using a function from a different package, in which case the output might differ.