So I'm a newbie into Data Science field. I've created a model using Gradient Boosting to predict a house price. To evaluate my model, I've splitted my dataset into train and test (0.8 Train / 0.2 Test) and got 0.94 R^2 for train and 0.89 for test. If I want to deploy this model, for business purpose, do I have to make a new model with train and test data together? If I do that, couldn't my model have a completly different result, like overfitting? Or should I use the model that I've trained from train dataset and create some kind of 'model.joblib', import to google cloud and create an API from train based model?
1 Answer
Below are the steps followed.
- Split data to train and OOT.
- Model on train data.
- Evaluate on Test set.
- Pickle the model
- Create an API.
Use the model that you have trained from train dataset and pickle,and create an API