Sentiment Analysis is essentially a process of understanding the
context of any non-technical writing. It combines aspects from
natural language processing and machine learning algorithms to
determine sentiment of the writing. They these techniques to
categorize writing, weighing each of the words and ultimately
providing a "score" in sentiment. This analysis is very useful when
trying to gauge public opinion, provide better understanding of
context and much more.
How it works, is that the analysis will break down the writing to
components, whether it is sentences, or even just words, it will
identify which words have sentiment by categorizing them as either
context or stop words. Then it will categorize it by sentiment,
positive or negative. This can be further explored where varying
degrees of positive to negative sentiment can be established.
In a world where huge amounts of unstructured data in the form of
written information exists, there comes a need for data scientists to
leverage them for various use cases. Mainly dealing with market
research and customer experience, these kinds of analysis brings in
a new dimension of information that was very hard to process
before. Data mining can be useful if the sentiment analysis can
identify proper sentiment.
Although these are heavily used in areas of customer interactions, I
believe this could be used in healthcare settings to differentiate
levels of pain, and identify the appropriate symptoms from a
patients' description of their ailments.
