sentiment analysis using weka

[Madhusudhanan Sambath]

hi to all
i need help,how can i convert the txt file data set to arff format to
give input for weka to do some sentiment analysis work.
i have downloaded product review data set from amazon, epinion web site .
all dataset are in txt file only.how can i convert to arff file and do
sentiment classification using weka software

i saw tutorial for movie review.but i am not able to convert the other
type of review as arff file(since we can give inpust as arff file for
weka)

kindly guide and help me.also suggest some book for text
preprocessing,feature selection methods , make model to give input for
classification algorithm using weka software

thanks in advance

thanks and regards
S.Madhusudhanan
research scholar,
Anna university, chennai.india

[Sam]

My suggestion is to pull out features you think will be helpful and then do all your preprocessing in Python. Make sure you're outputting something that's WEKA-compliant (tabs in the right places, etc.) and then write the whole thing to a .arff file. I don't think there's a pre-existing txt-to-arff module available, but I'm not certain. As for features--that's where the linguistics comes in! You could certainly try looking at the various machine-readable sentiment dictionaries that are available if you need some help/inspiration.

[saifee vohra]

Hello,
You can use rapidminer for sentiment analysis.In that you don't need to convert in .arff format, because it accepts txt files,CSV files.
Try look at following link..

http://www.corequant.com/?p=1

Hope this helps.

On Fri, Mar 15, 2013 at 8:28 PM, Sam <sam....@gmail.com> wrote:

My suggestion is to pull out features you think will be helpful and then do all your preprocessing in Python. Make sure you're outputting something that's WEKA-compliant (tabs in the right places, etc.) and then write the whole thing to a .arff file. I don't think there's a pre-existing txt-to-arff module available, but I'm not certain. As for features--that's where the linguistics comes in! You could certainly try looking at the various machine-readable sentiment dictionaries that are available if you need some help/inspiration.

--
You received this message because you are subscribed to the Google Groups "nltk-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nltk-users+...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Madhusudhanan Sambath]

thanks for ur reply