Query: Can you please elaborate this sentence: “Sentences with subjective expressions are retained and that which conveys objective expressions are discarded.” Why would be of interest to analyse sentences with subjective expressions, aren’t those conveying objective expressions of more value? To elaborate my question.It is clear that sentences with subjective expressions hold writer's stance that he is probably very passionate about, but if we consider product reviews where people often give negative review to a product on account of some shortcoming - that could be considered as objective comment and wouldn't it have a value to a business owner to analyse those reviews as well? Therefore my question, why do we eliminate these sentences and should it be always done like that?



Answer: 

In sentiment analysis there are two kinds of analysis- Polarity and Emotional Analytics.


Polarity Analytics - it will be positive, negative or neutral.

Emotion Analytics - it comes under sad, happy, joy, anger, surprise, unknown.


According to the above two types of analytics, we don’t ignore anything. 100% we concentrate more on negative comments where user gives neutral or negative comments.


What actual portion of the text which says negative  is known as vibes.


Vibes will actually give what is the actual form of text or a sentences which gives you negative or positive or neutral impression about the product.

E.g: “I don’t like the product” – Negative Vibe


Rather than whether sentence negative or positive, we actually find the information what makes the sentence negative and what makes the sentence positive.

“Product is really good” - Positive Vibe.


To be short nothing is omitted while you are doing the analysis. The one thing is that how you are going to clean this data and how you are going to apply this polarity, emotional analytics on your data which gives the better accuracy results of what form you want.


If you go for a normal text mining or any sentiment analysis package, you will not get to the expectation what you need for the business. Definitely we need to go lot of customization to make sure that the entire thing is in line with your needs.


None of the sentiment analysis packages come in and which suits your business need. Only with the social media data if you go ahead directly and apply which is of generic in terms you can get those sentiment analysis package working but with little customization is required.

When you want some in specific in terms of objective or subjective in the commands definitely you have to undergo lot of customization on text mining and sentiment analysis package. 


You may refer to the below recording for reference in which it has been explained in the Doubt Clearing session.

https://edureka.wistia.com/medias/o648iah5qa