Sentiment analysis using Python

[Henrik Lied]

Hi there,

An experiment I'm currently involved with requires some form of texual
analysis to figure out the "mood" of sentences (ergo, if the sentence
in any way seem negative or positive).

Do any of you have experience with something equivalent, and perhaps
some pointers on where I should go forward?

I've had some experience with Naive Bayes in the past, but I assume
the training would be colossal (I've got texts in the thousands to
examine).


Any tips appreciated!
-- Henrik L.

 

[Paul Rubin]

Do any of you have experience with something equivalent, and perhaps
some pointers on where I should go forward?

en.wikipedia.org/wiki/Latent_semantic_analysis perhaps.

 

[alex23]

An experiment I'm currently involved with requires some form of texual
analysis to figure out the "mood" of sentences (ergo, if the sentence
in any way seem negative or positive).

I vaguely recall seeing a presentation on exactly this at an open
source con a few years back:

http://osdc2006.cgpublisher.com/proposals/73/index_html

The paper is available here:

http://maurice.vodien.com/unrefereed/OSDC06-MontyLingua.pdf

The "emotional estimation" aspect revolved mostly around Open Mind
Common Sense:

http://openmind.media.mit.edu/

Unfortunately, there are no code samples in the presentation, but you
may be able to approach the author, Maurice Ling.

Hope this helps.