Classifying tweets by sentiment using Weka

0
votes

I want to classify tweets into positive-negative-neutral. For that I have a CSV file containing tweets and sentiment [positive-negative-neutral] in each row. When I feed this file to Weka, it converts it to arff format. Then when I select the classifier [for eg. J48] I am able to see results only for "negative" tweets. It shows precision, recall etc but only for "negative" attribute. I want to see all the values considering all 3 sentiments.

Next, I want to train it for all negative, positive and neutral and then try the test csv file. Can someone please tell me how I can proceed?

1
javamachine-learningwekaclassification
Please show your sample data and sample output so it will be more clear - Muhammad Abrar
Note that this is not a programming question: you are asking how to use a tool that somebody else wrote. I would suggest contacting the weka developers, or posting on their mailing list (more info at cs.waikato.ac.nz/ml/weka) - Nathan S.

1 Answers

1
votes

As you haven't provided data, you are working on, and sample output, can not help very explicitly but yes you may get some help by following blog written by Jose Maria Gomez Hidalgo(Nihil Obstat) on sentiment analysis using Weka at http://jmgomezhidalgo.blogspot.in/2013/06/baseline-sentiment-analysis-with-weka.html

Hope it helped.