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