.
1.3 Problem Statement
Given a set of tweets containing multiple features and varied opinions, the objective is to extract expressions of opinion describing a target feature and classify it as positive or negative.
1.4 Motivation
Sentiment analysis and opinion minion [14] is open research field with manifold real life applications. Blogs, Forum, Twitter, Facebook and other resources on internet are put to use by humans for expressing their opinions. The social media has bought the people around the world closer; communication is one click away. Before social media there was expensive short messaging service (SMS) provided by telecommunication companies with domestic and international charges. Today the short messaging has evolved from just sending messages to single person to sending messages to multiple people at cheapest price. This service is provided by many websites but Twitter was the one which pioneered it. Today twitter has hundreds of millions users who post nearly half a billions tweets every day i.e. approximately thousands of tweets for every second. Tweets are not only posted in English language but also in different local languages of the world. These data are precious to business intelligence where the company wants to know "why isn't consumer buying our laptops?", "why the competitors products are outselling our products". Thus a concrete system to process above mentioned queries is the need of the hour.
1.5 Objective:
Classify every tweet in either as positive sentiment or negative sentiment using different Machine Learning techniques and check which classifier performs the best.
.
