Feature included
1.) Handling typical Negation (Eg. not good) 2.) Use of contractions with negations (Eg. wasn't good) 2.) Use of BoosterWords or Degree Modifiers to alter the sentiment intensity score (very good) 3.) Emoticons such as :) and :D 4.) Exclamation and Question marks emphasis 5.) Emphasis on Capitlization of Words (Eg. very GOOD) 6.) Handling of Internet Slang Words 7.) Handling Initialisms and Acronyms such as 'lol' 8.) Emphasis of punctuation marks (GOOD!!!) 9.) Handling Idioms Phrases.
Including these all features improves the polarity classifying of tweets as positive, negative and neutral.