"(Sentiment Analysis) How to Assign Weight to Words in Training Set"

ogzkusogzkus Member Posts: 2 Contributor I
edited June 11 in Help

Dear Fellow Rapidminer Users,

 

These days I am working on conducting sentiment analysis on social media data. In my training set I am using the words in order to train the algorithm and every word has a score which shows positivity/negativity. However, they have different level of positivity or negativity. For example:

happy - 4.8 positive
sad - 2.7 negative
brilliant - 4.98 positive (more positive then the word 'happy')

 

As it can be seen from the example words positivity/negativity level of the words are different. My question is that how can I assign weight to words in traning set instead of labeling them only positive or negative? Which kind of algorithm should I establish in order to conduct sentiment analysis within the indicated framework and do you think that will it be more detaily and efficient when it comes to sentiment analysis?

Answers

  • mschmitzmschmitz Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 2,060  RM Data Scientist

    Hi,

     

    have a look at Dictionary Based Sentiment Analysis in the operator toolbox extension. That should do the trick.

     

    Cheers,

    Martin

    - Head of Data Science Services at RapidMiner -
    Dortmund, Germany
    sgenzerogzkus
  • ogzkusogzkus Member Posts: 2 Contributor I

    Hey!

     

    Thank you for the answer. Is it possible to state a roadmap? Because, I am not so use to Rapidminer for this kind of complicated process.

     

    Thanks

  • mschmitzmschmitz Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 2,060  RM Data Scientist

    Hi,

     

    what do you mean by road map? An example how to use it? 

     

    If yes, have a look at the help text of the operator. It always provides a tutorial process on how to use it.

     

    Best,

    Martin

    - Head of Data Science Services at RapidMiner -
    Dortmund, Germany
    sgenzer
  • priyankajanjanapriyankajanjana Member Posts: 1 Contributor I

     

    Hiii sir!! Did you get the solution for the problem which you posted??If so,please suggest me.I am also searching for the same...

Sign In or Register to comment.