Record Multiple Benchmark Results

Convenience function for recording several metrics from one model run.

Usage

record_benchmarks(
  conn,
  step_name,
  model_name,
  format,
  metrics,
  baselines = list(),
  n_train = NA,
  n_test = NA,
  date_range = c(NA, NA),
  gender = "all",
  notes = NULL
)

Arguments

conn

DBI connection (read-write)

step_name

Character. Pipeline step name

model_name

Character. Model identifier

format

Character. Cricket format

metrics

Named list of metric values (e.g., list(mlogloss = 1.23, accuracy = 0.45))

baselines

Named list of baseline values (same names as metrics)

n_train

Integer. Training set size

n_test

Integer. Test set size

date_range

Date vector of length 2

gender

Character. Gender filter

notes

Character. Optional notes

Value

Invisibly returns the run_id used