Back

A New Model to Determine the Personal Reference Interval of Tests in Laboratory Medicine: A Preliminary Study

Coskun, A.; Serteser, M.; Cavusoglu, C.; Kilercik, M.; Unsal, I.

2020-02-11 pathology
10.1101/2020.02.07.20020446 medRxiv
Show abstract

"The authors have withdrawn this manuscript because they discovered errors in the equations that no longer support the findings given in the paper. Therefore, the authors do not wish this work to be cited as reference for the project. If you have any questions, please contact the corresponding author."

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.