Hi !
I am trying to adjust lessen4 model to training sentiment of twitter data. I use a 1m kaggle twitter training data sample but I only get 74% accuracy without really increasing accuracy during training:
Hi !
I am trying to adjust lessen4 model to training sentiment of twitter data. I use a 1m kaggle twitter training data sample but I only get 74% accuracy without really increasing accuracy during training:
@ulat It might because of various reasons, one being a lot of noise in data. Improprer feature engineering etc.,
Can you post the full code ?