Open SamMckaylin opened 4 years ago
you should not have to mess with that. Is the code broke?
The Foote novelty estimation is outputting an error saying the object width.vector can’t be found or something to that effect. I’m at a meeting about posters at the moment but I can check the exact output when I’m back
It all seems to run again, not entirely sure what happened
Just to check, should it write "thresholds" as a dataframe to the "Global Environment" tab?
Sam --- are you using the version on GitHub? It should just run in Studio. Do you want me to check?
I was trying to replicate the analysis for different topic disciplines and I think I'd accidentally missed changing a few of the topic names. While trying to run everything line by line to look for my errors I noticed the elements of the FN calculator weren't working or writing to the Global Environment which I thought may have been my fault. Now I've corrected all of my mistakes I've realised that the FNC needed to be run as a cluster and it all seems to work. The different topics disciplines are all producing different results. Although notably, while there is a short "revolution" in the 1930s it isn't as significant or long lived as several others. I'll save everything which seems notable and try to work out what's happening. Apologies for being a bit of a div earlier.
Looking through the code I've realised on line 122 in the original code (Within the section on Foote Novelty estimation) the object "width.vector" is called but it hasn't been formally specified in the code prior to this
thresholds <- data.frame(matrix(nrow = length(width.vector), ncol = 3))
It's mentioned a couple of lines prior:
novelty.analysis <- function(dm, width.vector = c(1,2,3), ft.save.drcty = "foote.results.csv", th.save.drcty = 'foote.thresh.csv', n.boot = 1000)
What should I specify it as in "thresholds"?