0

I want to help my friend to analyze Posts on Social Networks (Facebook, Twitter, Linkdin and etc.) as well as several weblogs and websites.

When it comes to Data Analyzing, I need to analyze the feeling of people about a specific subject e.g. The US Election. My idea is finding the sentences containing a set of the words/phrases (or tags) and looking for a set of adjectives or icons e.g :). I need a machine learning algorithm to reduce the mistakes (understand sarcastic posts, find new trends, and maybe more) and some toolkits to help me.

  • Which model of UCI works for me? or any other possible model.

I have also found NLTK and MontyLingua and TweetNLP. I tried to check NLTK but it is not as good as I need. I need a toolkit which can process a sentence according to the whole meaning of that not words. I think this Stackoverflow post is great for this prat of my question but It will be great if anybody give a clue for the fastest with the least usage of resources.

Community
  • 1
  • 1
Alin
  • 350
  • 2
  • 13
  • 1
    You can find the complete answer on duplicated question, but as a brief answer, as far as i know and as i have experience with `NLTK` you can use it to find the type of words (`Pos-tagger` or ... ) like adjectives or verbs, then you need a mathematical calculation maybe using probability to guess the emotions! and for math part you can use `Numpy` and `Scipy` – Mazdak Apr 05 '15 at 09:40
  • 1
    Also `Stanford NLP Parser` could be a good choice for such tasks!that `nltk` support it too! http://nlp.stanford.edu/software/lex-parser.shtml – Mazdak Apr 05 '15 at 09:47

0 Answers0