How do we rank the tweets after classification?

We only build models to predict if the tweet will be liked or reshared or commented. What I am so confused about is how we would rank them. It seems like we need to use the engagement scores for each tweet. However, how should we calculate the engagement scores? It seems different from the user engagement scores mentioned in the early sessions.


Hi @luoluoluo

The term predict here doesn’t mean only predicting on the basis of the training model. The engagement score would be added in the training model whenever a tweet is liked, shared, or commented. Only the name of the term is changed here along with some logic but what you are thinking would be done in the training model.
If you still face any confusion. feel free to ask.

Happy learning,