Report full LMTP analysis-weight concentration — margot_lmtp_analysis_weight

`margot_lmtp_analysis_weight_report()` multiplies the baseline design weight by the cumulative joint exposure-and-censoring ratio at each longitudinal node. Zero weights remain in every distribution and Kish effective-sample-size denominator. The function reports aggregate concentration without attaching a positivity or estimator-stability judgement.

Usage

margot_lmtp_analysis_weight_report(
  baseline_weights,
  joint_ratios,
  regularised_joint_ratios = NULL,
  wave_labels = NULL,
  policy_id = NULL,
  top_shares = c(0.01, 0.05, 0.1),
  regularisation = NULL
)

Arguments

baseline_weights: Non-negative baseline design weights, one per participant.
joint_ratios: A non-negative matrix of per-node joint exposure-and-censoring ratios, with one participant per row and one longitudinal node per column.
regularised_joint_ratios: Optional matrix after the registered numerical regularisation, with the same dimensions as `joint_ratios`.
wave_labels: Optional longitudinal-node labels. Column names from `joint_ratios` are used when available.
policy_id: Optional policy identifier recorded as provenance.
top_shares: Fractions used for concentration summaries.
regularisation: Optional aggregate description of the registered numerical regularisation.

Value

An object of class `margot_lmtp_analysis_weight_report` containing per-wave summaries, top-weight shares, aggregate zero and missingness causes, regularisation comparisons, and provenance. The return value contains no participant-level weights.