I am trying to make a sales forecast using neural networks. The model itself works, but my dataset still has many columns that contain information in string form, which currently cannot be taken into account. Is there a node for Keras that can be used to add nominal values as input to the model?
@Vincentsoy what you could also do is use target encoding to convert the strings to numbers. This could work (also for Keras) if you expect the relationship between the strings and the target would remain stable.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.