Monument project 2022-12-24 to 2023-01-06
to : Martin Holmes
Minutes: 240
This is logging work done over the vacation.
The first job was to write a merge process which can take a wife and husband record pair and merge them into a single record. This is done, and works OK for the tests I’ve set up so far in XSpec (which I’m getting better at using), but there is still one currently remaining bug in the merge process which is leading to duplicate child records; that must be fixed before we can actually draw any conclusions from the stats (see below).
I set up a system where two trial merges are done, one generated using
all the potential matches, whatever their confidence level, and one using
only the 1.0 confidence matches; then there is a statistical process that
is run on these so we can get some idea of where we’re at in terms of
the merge process. We also broadened the scope of the initial harvesting
process on the advice of NA, so that we include people who don’t have a
formal <residence> element specifying their uprooting
but who do have the phrase forcibly uprooted
in their bio note.