"Where can I find a tutorial on using RapidMiner for sentiment analysis?"

Legacy User

Folks,

My interest is in doing sentiment analysis by keyword on web-based data.  My primary inputs are RSS-feeds.  No web crawling is needed as I have an continuous feed on topics that are of most interest.  From my review of the documentation and the examples, it unfortunately isn't clear how I can make this happen.

So, where can I find a tutorial or a how-to guide that will alllow me to build a sentiment analysis model?  

land

Hi,
do you have already downloaded the text plugin? You will need it in order to analyse texts. It provides several samples how to do this.

Greetings,
  Sebastian

Legacy User

Yes, the text plugin AND examples are installed.  Unfortunately, the examples are not as intuitive as I would like.
I am looking for a HOW_TO guide that provides a clearer overview of the steps needed to complete a sentiment analysis.  

My experience is that this kind of reference material is common in the open source world.

land

Hi,
beside the tutorials on http://sourceforge.net/project/showfiles.php?group_id=114160&;package_id=128128 and http://nemoz.org/joomla/content/blogcategory/21/53 there are no tutorials available as far as I know. Unfortunately one might assume, that the topic is far to complicated to summarize it in one single howto.
Currently we are planning to publish some guides and tutorials in the near future, but daily work consumes most of the available time.

Greetings,
  Sebastian

Legacy User

Hi Sebastian!

Thanks for the links but would it be possible to give me a list of the operators to use plsss?  :-*

I'm trying to:
1) feed ready articles to the RM
2) extract sentiments of related topics
3) visualise the sentiments of the topics

Thanks sooo much!
???

TobiasMalbrecht

Hi,

sorry, but I think we can not really help you here since the question is far to unspecific. To give you a short answer: for the text preprocessing you will need to use the operators from the text plugin - probably many of them. For the extraction of sentiments you will need some analysis approach. Either you have to label the data and apply some supervised learning method. Otherwise clustering might be a good starting point. Hence, it depends crucially on your data and your analysis goal, which operators you will actually need.

Agian: have a look into the examples in the text plugin for text (sentiment) specific applications and the RapidMiner tutorial to gain insight in the data mining steps.

Kind regards,
Tobias