September 28, 2010 – Consumer-generated content continues to grow at a rapid pace. Much of this is caused by online purchases and the consumer reviews/ratings spurred by that growth. So it should be no surprise that marketers are spending a great deal of their time trying to monitor and manage the online consumer-generated information, and many are using natural language processing (NLP).
This interesting information was found on a blog post, “Behind machine based sentiment analysis.” Many of the larger marketing/PR firms have moved towards automated platforms that use NLP to discover, measure and report on various topics. This is commonly called online reputation management (ORM). The platform uses a training database of words and phrases that indicate negative and positive tone. When a document is being processed, the dominant words/phrases are extracted and then compared with the language taxonomy. From this, a whole number is then assigned to that document based on an n-point scale, e.g., very good, good, neutral, bad, very bad.
Interesting post and although automated sentiment analysis will never be as accurate as human analysis (because, well, for one thing, it doesn’t have a sense of humor), the technology is good for monitoring online buzz and measuring crisis management.
Melody K. Smith
Sponsored by Data Harmony, a unit of Access Innovations, the world leader in indexing and making content findable.
[…] Original post: Consumer Products Getting a Machine Assigned Reputation : TaxoDiary […]