Is there a Machine Learning Ranking algorithm that can rank documents for a query using as a training dataset to the algorithm a dataset having:

  • some features of the queries,
  • some features of the documents,
  • a boolean telling which document was selected by a user among the documents presented for a query.

The main point here is that the training data don't have a target variable that is pointing to a hierarchical relevance of the documents, instead for each query only one document out of all of them is considered to be relevant.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.