I want an implementation of a basic sentiment classifier on a set of text files.
The attached files represent news, speeches, whatever. A whole bunch of text.
For example, on Sept. 13, there is one speech transcript.
The output would be something like the following:
13 Sept 12: 65%
Where the percentage is some characterization of the positive or negative communication inherent in the document. This would probably be done by a binary read of the sentences.
The weather looks good.
The weather outlook is bad.
13 Sept 12: 50% ... would be the result from those two sentences.
I want to do that for each sentence within each document.
I want a summary by document in a spreadsheet.
The implemented algorithm should be either Sentaword net )[login to view URL]), OpinionFinder ([login to view URL]) or both.
The task of processing the texts against the SentiWordNet word base looks clear. What environment should the script be run in? Please see the private message for the summary.