A small addition
Folk have requested the possibility to run their dictionaries over all project documents at once to generate a ‘unified dictionary report’ mirroring the unified word frequency report. This function is now attached to a menu item in the latest preview release. It drops the resulting report straight into a CSV file, for easy import into whatever you like to do your data analysis in.
Dear Will & other developers,
I am currently writing my master’s thesis and want to analyze the level of negativity of company specific news articles and how it affects inevstors. I will download the articles from Factiva or LexisNexis. I will perform pretty basic analysis on the articles, e.g. proportion of negative words relative to the total amount of words in article, and so on.
So, on to my question: Is Yoshikoder suitable for this kind of research? Does it have the capacity to analyze say 100,000 news articles (word length will be less than 1,000 per article) in reasonable time? Initially I was considering using the General Inquirer (since it features the Harvard IV-4 dictionary’s negative word tag, which is very important for my research), but was unable to find it on the internet, nor was I successful at contacting its developers. However, I believe the same negative word-tag might be included in Yoshikoder, and if not, it could be added there relatively easily?
Thanks for your help,
Best regards,
Mika
Mika
November 13, 2008 at 8:01 am
Just fyi, I answered this one response offline. I’ll post the reply in a mo’.
Will
December 10, 2008 at 3:32 pm
About the General Inquirer:
As far as I know it’s only available publicly via a web interface at http://www.webuse.umd.edu:9090/ For some reason UMD haven’t answered my intermittent email about accessing other forms of the GI. (Actually they don’t answer my email at all). So it’s all a bit mysterious.
The relevant GI dictionaries are still available for download though, and can be massaged quite easily into, say, VBPro format. Bear in mind that the GI does somewhat sophisticated word sense disambiguation which most programs no longer attempt, Yoshikoder included. Hence your results won’t be quite the same as the GI’s. But I’d expect them to be very similar in large samples.
Will
December 19, 2008 at 9:25 am