This commit is contained in:
Pablo Antonio Lillo Cea 2026-05-08 10:17:14 +02:00
parent d302b8e1a2
commit c50b9917d3

View file

@ -23,15 +23,6 @@
# deaths, migration) correlated with CA row
# scores post-hoc
#
# Why restrict active variables to person-counts?
# CA chi-square distance has a clean compositional interpretation only when
# all active columns measure the same type of unit. When counts of persons,
# dwellings, farms, and vehicles share a row, the row total is meaningless
# and chi-square distance becomes a weighted average of incompatible profiles.
# Restricting active columns to person-counts ensures that chi-square distance
# directly compares "what fraction of the population has attribute X?" across
# municipalities, independently of population size.
#
# Outputs:
# data/processed/ca_exploratory.rds - CA result (FactoMineR object)
# data/processed/ca_var_groups.rds - variable group lists