Sentimental analysis
Sentiment analysis is the process of determining the opinion or feeling of the piece of customer reviews. We humans are pretty good at sorting out people sentiments based on the texts which are positive, negative or neutral. Companies across the world are developing machine learning to do it automatically. It is done on social media conversation and forums for targeted brands. This is really important when you deal with consumer brands. Sentiment analysis has many other names like Opinion extraction, Opinion mining, Sentiment mining, Subjectivity analysis. These opinions are stated differently.
Ø Organisation internal data: Customer feedback from emails, telephone.
Ø News and reports: Opinions in news articles and blogs.
Ø Word of mouth on wed: customers reviews, twitter, forums and social media.
Company can use this sentimental analysis to:
Improves marketing strategy: Most companies use social media and public forums to promote their products. Customers see the entire information about the product and in return, the company is able to find out the customer perception of their brand based on the sentimental analysis.
Develop new product: Sentimental analysis can read the target customers opinion on how they want the product. Based on this, company can create new products with new features.
Improve customer service: If customers complain about the brand, like products errors and late deliveries, sentimental analysis can easily identify the customers negative responses and solve the problem.
Sales revenue: Overall sales can be increased by successfully evaluating marketing campaign and improving product quality. All these can be improved by sentimental analysis.
Sentiment Analysis can be performed in three levels.
Ø The first is the document level, which is concerned with extracting the general opinion expressed in a document. In this level, it is important that the document addresses only one entity.
Ø The second level is the sentence level, where each individual sentence in a document is classified separately.
Ø The third level is the aspect level, which is concerned to identify exactly which aspects of the entity the author liked or disliked.
Difficulties in the sentimental analysis:
Most of the customers post out opinion about the brand but sometimes it is difficult to find out that statement is positive or negative to machine-based learning system because human language is sometimes sarcastic which is complex for machine learning to interpret.
BY
NAKKA SAI KISHORE
