AN APPROACH for big data technologies in social media mining

Abstract

Since the advent of Web 2.0 there has seen a shift from 
publisher-generated to user-created content, with the latter dominating the 
content available through the Internet. A significant proportion of this arises 
out of users who post reviews on products and services or give opinions and 
views on a range of topics, particularly through SM (social media) sites. As 
searching and posting on these sites can be  an indicator of concerns, interests 
and intentions,  this data can be turned into commercial and social value by 
exploiting them in a timely manner, becoming in turn of strategic importance to 
companies, health organizations and government agencies. There is a growing 
interest in social media analysis for detecting new trends, user opinions and 
researching product or supplier reputation. Web browser companies sell the 
opportunity companies have for targeted marketing. SM analysis is also used to 
make predictions about market developments or sales revenue.
Opinion mining is an established field in computational linguistics. However 
opinion mining using SM has some implications on the techniques used due to the 
peculiarities of SM especially when slang, special characters, abbreviations or 
slang is used. This paper describes an approach for SM analysis using machine 
learning techniques to handle some of these issues.