About the Practice Problem: Identify the Sentiments category

Use this category for the discussion related to contest: Practice Problem: Identify the Sentiments category Feel free to share your approach and ask your questions here.

For more information, visit:

1 Like

The data provided for the training file contains lot of noisy data I think.

Have tried one of the best text classification language model hoping that without any fine tuning the model the accuracy would be reaching top 10 of the competition, but was not.

When investigated there were lot of noisy data,

For example below are classified as 0 in my opinion it is 1 (a negative)

Apple is soo crap. Would never get an iphone have enough crap woth ipods. Wiped the whole lot.

Helping my lovely 90 year old neighbor with her iPad this morning has just made me realise that ‘I’ don’t actually need an I pad!

Lost all of my contacts due to the fact of my iPhone breaking

Sweet, my phones wifi won’t even let me turn it on so it can search for wifi. Second time this has happened to me. #apple

Seriously hating apple right now … I have and iphone 4 so no IOS 8 for me … #apple #mehhh

Yeah, I got the same result. Don’t worry, that’s how they have classified the data. How much accuracy did you get on the test set?

© Copyright 2013-2019 Analytics Vidhya