Sentiment analysis

NYT has a very enlightening article on Mining the Web for Feelings, not Facts.

This article very correctly points out:

An emerging field known as sentiment analysis is taking shape around one of the computer world’s unexplored frontiers: translating the vagaries of human emotion into hard data.

This is more than just an interesting programming exercise. For many businesses, online opinion has turned into a kind of virtual currency that can make or break a product in the marketplace.

There are many companies working in this area now. Obviously, marketing could use sentiment analysis extensively and this is the first application. But sentiment analysis will go much farther than just marketing. I liked:

As sentiment analysis algorithms grow more sophisticated, they should begin to yield more accurate results that may eventually point the way to more sophisticated filtering mechanisms. They could become a part of everyday Web use.

Understanding sentiments and using them is relevant to all aspects of humans. Nice that finally the Web is getting some feelings.

2 thoughts on “Sentiment analysis

  1. Mark

    Nice article Ramesh! The original piece looks at analysing sentiment – positive or negative opinions or reviews, whether the opinion is based on logic or feelings. Gathering an overall view of whether people are for or against something is one thing. Analysing why is not a job for any machine I’ve yet seen.

Leave a Reply