It seems the model we created in this assignment did not correctly classify the first sentence

Course Q&A Natural Language Processing NLP with Sequence Models
1

image
image779×392 48.6 KB

After I passed all the exercises, and use the first sentence to test the model I created.
It seems it still predicted the sentence “The Movie was almost good” as positive.

I thought it should be more accurate to give a negative sentiment.

Any Ideas about that?

2

Hi Amazing_Patrick,

Two thoughts:

‘Almost good’ is not necessarily the same as ‘bad’, and does not have to be considered negative. It could mean exactly what it indicates: the movie is acceptable, it’s just not completely good, it’s almost good (i.e. it’s in any case not bad). So it’s not a strange idea that the model predicts a positive sentiment.

Also, the models keep getting better as you progress in the course, so I would not consider this outcome too significant.