How to deal with data that is not in the dataset in ML.NET machine learning model? RRS feed

  • Question

  • I had created an answer grade prediction model in ML.NET(tried different regression algorithms and multi class classification algorithm). I have ID, question, answer and grade in my dataset. Inputs to the model are question and answer. Model predicts the grade for the answer. If we provide an input(answer) that is in the dataset, then the model predicts the grade accurately. But when the answer is completely incorrect and not matching with any of the answers in the dataset, then the model is not predicting the grade accurately.

    Any ideas on how to deal with this case?
    Tuesday, August 20, 2019 11:11 AM

All replies