sentiment analysis

User 2570 | 11/14/2015, 4:15:11 PM

how perfect is Graph create in performing SA from social media data(twitter & Facebook)

Comments

User 1592 | 11/14/2015, 5:51:11 PM

Accuracy of our sentiment analysis algorithm is highly dependent on your data. Here are some resources to get started: http://blog.dato.com/sentiment-analysis-in-five-lines-of-python http://blog.dato.com/practical-text-analysis-using-deep-learning


User 19 | 11/17/2015, 2:16:11 AM

I agree with Danny: it depends on your data set. You can find details on how the default model was trained here: https://dato.com/learn/userguide/sentiment_analysis/sentiment-analysis.html

The best solution is to train a new sentiment classifier on your own data set. Labeled data can be expensive to collect, however. Sometimes people use the presence of positive (or negative) emoticons as a proxy for a positive (or negative) sentiment.

Hope that helps, and please let us know if you have further questions. Chris


User 2570 | 11/17/2015, 10:46:26 AM

Thank you so much, looking forward to it, in case of any problem will surely let you know