Sentiment analysis is an analysis made on a computer whose goal is to discover what writers feel about a certain topic or the feelings they are expressing while they made their text. The sentiment being analyzed is usually categorized into two binary distinctions, which may be positive or negative.
However, it may also determine much more detailed or fine-grained feelings such as, if the writer feels anger, joy, disgust, sadness or fear. Sentiment Analysis may also be known as opinion mining, which derives from the attitude or the opinion given by the writer. This technology is used nowadays to analyze what people would feel about a certain topic and is especially applied in social media. It may also be used in evaluating survey responses and evaluating whether product reviews are positive or negative.
HOW DOES IT WORK?
There is a various amount of ways to do a sentiment analysis. Generally, the approaches to making a sentiment analysis use the same idea. In conducting this analysis, you must create a list of words that strongly associate with strong feelings which portray positive or negative sentiments.
However, this may be the most time- consuming stage since you have to modify or add to the list depending on the topic you have. The second step is to count all the positive or negative words in the text. And then analyze the mix of positive and negative words. If there’s a lot of positive words than negative, then it means that the sentiment is positive. If there’s more negative than positive, then overall it is negative.
Sentiment analysis performs its analysis in textual contents in R. This means it will utilize various existing dictionaries such as Harvard IV, QDAP or Loughran-McDonald. This may also create new and customized dictionaries and uses LASSO regulation as a statistical approach in selecting relevant terms based on an exogenous response variable.