Briefly: a couple of news items:
- What Counts at the Box Office is The Buzz in which the NY Times reports on a project out of the University of Arizona. The research finds that movie success is best predicted simply by online buzz - no other feature, such as polarity/sentiment, really helps. Compare this with work by Gilad Mishne and Natalie Glance which found that sentiment did provide some information: Predicting Movies Sales from Blogger Sentiment.
- Text Mining the New York Times [via SlashDot] in which a method is presented which apparantly heralds the coming of age of text mining in industrial applications. Great - maybe I can go and get a job in that field now!
Comments