I created a workflow that reads a csv file, pre-processes the documents and trains a model to predict the sentiment of movie reviews.
I get a good roc curve and an excellent final score.
At this point, how can I predict the sentiment of a new text??
- List item
if you have created a ROC curve that means you are already applying the model to previously unseen data, right? If you have built the ROC curve on your training data, that is a very bad approach as it tells you nothing about the model quality. So assuming you have applied the model to unseen data for your ROC curve, you’d have to use exactly the same approach to predict the sentiment of a new text.
Hi there @Reggina56,
if you are interested in model deployment maybe this link can help:
This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.