Sentiment and Contextual Analysis

In Chapter 18 of the book "Share This: The Social Media Handbook for PR Professionals," Andrew Smith writes about social media monitoring and two ways to analysis such. More so, he brings up two forms of analysis that one can use: sentiment analysis and contextual analysis. Being that I am soon to enter the marketing world, knowing different analysis styles is always a good thing especially in the terms of social media monitoring. So when I read Smith's chapter about these two styles, I found it to be particularly interesting. 

Smith concludes that sentiment analysis is basically an "attempt to deduce how somebody feels about a particular person, topic, issue or organization based on what they say." So in other words, the technique assigns certain words to have either a positive or a negative relation. The social media is then analyzed, creating a total count of positive verse negatives. If the negatives count is higher, than the subject is deemed negative and vice versa.  

I do like this technique but it does propose one major issue, accuracy. Smith touches on this topic of lacking accuracy by explaining how the technique does not consider the slang words. For example, "sick" would traditionally be considered a negative word, but in today's slang, it could be meant in a positive way. 

From reading Smith's description of contextual analysis, I concluded that it is sentiment analysis's "bigger brother." I call it that because it uses "sophisticated computer algorithms and semantic analysis" making it more accurate than sentiment analysis, or you could say it is smarter if you are still thinking about it terms of a bigger brother. The cost for this technique is much higher of course, but the results have such a higher degree of accuracy. 

Until Next Time,

Matt