dictionary based sentiment analysis with german text

nrwstudent

Hi there,

does someone know how to create a dictionary based sentiment analysis process for german texts?
I've just seen solutions for english texts using Wordnet.... i'd like to use the german SentiWS - does someone have

experience with that one ?

at the moment I've got an process that just counts the positive and negative words in the text but it doesn't weight the terms-

so that as a result "schlecht" (bad) and "fürchterlich" (dreadful) are regarded as equal. SentiWS weights the words but until now i

filtered the weights out, since I do not know how to deal with them in the process.

I'd be very thankful if someone helped.

Thanks in advance

Thomas_Ott

I dont have experience with the German Sentiwordnet but I believe you can "hack" a way to connect it to the RapidMiner operators using an Open File operator. In the past, that's how some users have done it before we had the Wordnet Extension. I would search these forums or the InterWebz for something.

MartinLiebig

Hi,

someone from Katharina Morik's group at TU Dortmund was working on using some other german dictionary, propably within RM. I sadly forgot the details. But you may contact her.

~Martin

ep020

Hi,

I also work on a sentiment analysis in German. And I also need to count the positiv and negativ words. Can you please send mir your process.

It would be very helpful.

Thanks!

Thomas_Ott

@ep020it would be best if you want to get someone's attention to use the the '@' symbol and their name to get their attention.

ep020

@nrwstudent