Sentiment Analysis of Customers

Abstract

There is an economic and political need to learn and know more and more information about customers, and the social media has recently become the most powerful tool for interaction with people. Customers became users in the social media, who express their opinion and share it with not only companies but also other users. Since a huge amount of reviews, opinions and comments appeared, there is a necessity to extract, aggregate, and analyse them; these are the aims of sentiment analysis. In this paper the technological details of sentiment analysis are presented. The problem types and their solution with text mining are described. Text mining is based on data mining, but steps of text preprocessing with tokenisation, stemming, filtering is an additional important phase before the data mining procedure. The correctness of sentiment analysis solutions can be measured by different validation methods. At the end of the paper the final conclusion is presented.