Re: NLP question tonight

[Hector Castro <hectcastro@gmail.com>]

Hi Joe,

Thanks a lot for following up on this.  The Tech Talk video looks very interesting.

A professor at Penn actually pointed me to General Inquirer a few years back and asked me to process a ton of NYT articles for positive and negative sentiment.  I remember access to it being complex and having to exchange several e-mails with its administrators to resolve errors.  Unfortunately, I don't remember how well the results turned out.

If I recall correctly, we ended up using LIWC to process the articles:  http://www.liwc.net/

Thanks again,

--
Hector

On Jan 11, 2012, at 10:21 PM, Joe Snikeris wrote:

> Hi all,
> 
> Someone had a question about inferring positive or negative sentiment
> from a given text. His example was processing SEC filings to determine
> if their content reflected negatively on a particular company.
> 
> There is a system called the General Inquirer[1] that can provide word
> scores indicating how positive or negative a given word is. For
> example, words like 'dire', 'shortage', and 'layoff' would have scores
> that indicate a negative connotation. For an example of how this can
> be used to make trading decisions see the Google Tech Talk by David
> Leinweber[2]. His slides[3] mention the General Inquirer on page 10.
> 
> Regards,
> Joe Snikeris
> 
> [1] http://www.wjh.harvard.edu/~inquirer/
> [2] http://www.youtube.com/watch?v=HJqtqNl5G4E
> [3] http://tinyurl.com/74nh8zn (PDF)