Hey, Rober Zinkov (from Los Angeles) dropped by the Metalab yesterday and told me about the talk he is going to give at the WU. It's about sentiment analysis in text data, and sounds really interesting. So if you want to know how emotions are extracted from your twitter posts for fun and profit: Abstract: Over the last decade, researchers have discovered that text can be analyzed for its emotion content. Opinions can be readily extracted from documents about various topics and products. This presentation will cover the various smaller tasks involved in performing sentiment analysis. This will include extraction patterns, both unsupervised and feature-based learning methods. I will cover how sentiment dictionaries are built and sentiment is seeded. I will also discuss how sentiment analysis interacts with other existing natural language processing systems such as entity extractors and summarization algorithms. This talk will assume basic familiarity with machine learning and natural language processing algorithms. It will be in the Computer Lab on the 5th floor of the Statistics Building at the Business School. It will be on 26/5/2011 at 16:00. Everyone is welcome to go. Institut für Statistik und Mathematik WU (Wirtschaftsuniversität Wien) Augasse 2-6 1090 Wien Austria