I'm curious if I can load my saved trained model and predict different features and target, is that possible?

Because I want to deploy ML model into Streamlit web app and users can choose features and target whatever they want, just something like this

If not possible, are there any tricks or ideas to predict something according to the user's wishes for feature and target selection, especially for Streamlit deployment? Thanks!

CodePudding user response：

You can use your saved model for prediction, If you want to move to cloud then you can retrain the model, save the saved model(say h5) in the cloud(say AWS) then you can fetch predictions from in there. Plus it would give you the flexibility to constantly train the model and used the trained model.