Chapter 6. Sentiment Analysis

"Are you happy or not; that's the question!"

Sentiment Analysis (SA) or opinion mining is a technique used to figure out the polarity (positivity or negativity) of a sentence automatically.

This can be quite difficult given that natural language is difficult for computers to decipher. There is another related concept called Emotion Detection (ED). While the task of SA is to determine whether a given sentence or a phrase represents a positive or negative sentiment, ED tries to do something more challenging. It tries to find the actual emotion being expressed in a text.

So, the output of ED algorithms is categorical (joy, sadness, anger, violence, feel-good) while that of SA algorithms is mostly Boolean (the sentence being examined by the algorithm has either a positive or a negative polarity).

Sometimes, it makes sense to return the polarity percentage from SA algorithms as that can be used as a degree of positivity. SA is important because...