BuzzMetric‘s Matthew Hurst’s excellent blog, “Data Mining: Text Mining, Visualization and Social Media,” has run some great examples of why sentiment analysis—figuring out the attitudes expressed about a topic—is so damn hard. Here‘s someone posting about Steve Rubel and Wal-Mart: “Certainly, if you don’t post about it, I’m personally not going to not respect you for not doing so, however others may.” Let’s see a computer make sense of that!

Of course, that’s just overly-complex writing that even humans have trouble parsing. But how about this example, drawn from a review of the Su Hong restaurant: “When we moved from the area, my wife’s most frequent complaint was that no one made orange peel beef like Su Hong.” As Matthew points out, it’s a positive review in the form of a complaint.

Sentiment analysis remains real hard. For judging overall market trends, it doesn’t have to be perfect, but I’d love to see some analysis of how good it actually is right now. [Tags: sentiment marketing matthew_hurst everything_is_miscellaneous]

Categories: Uncategorized dw