I'm developing an app that automatically captions photos. It uses GPS among other things to create a human readable text.
There are cases where a photo has some data missing, (i.e GPS coordinates). I could still try to recover some data (last known GPS position), but that wouldn't be as accurate as the original data (exact GPS position). Therefore, the generated caption would also be slightly more incorrect than normal.
What's better in this case? Not generate a caption at all, or generate a slightly incorrect one?