Summarise calendar-development statistics (Mean, Median, Weighted)

S3 method for summary() on Calendar objects. Computes calendar-period summary statistics for cumulative loss ratios (ratio) and per-period loss ratios (incr_ratio).

Where summary.Triangle() aggregates by (groups, dev) (cohort x development), this method aggregates by (groups, calendar) (calendar period) so the resulting table is indexed by calendar diagonals rather than development periods.

Usage

# S3 method for class 'Calendar'
summary(object, ...)

Arguments

object: An object of class Calendar.
...: Unused; included for S3 compatibility.

Value

A data.table of class "CalendarSummary" with one row per (groups, calendar) combination, containing:

n_cohorts: Number of observations in the cell.
ratio_mean: Mean of cumulative loss ratios.
ratio_median: Median of cumulative loss ratios.
ratio_wt: Weighted cumulative loss ratio (sum(loss) / sum(premium)).
incr_ratio_mean: Mean of per-period loss ratios.
incr_ratio_median: Median of per-period loss ratios.
incr_ratio_wt: Weighted per-period loss ratio (sum(incr_loss) / sum(incr_premium)).

The returned object preserves the attributes groups, calendar, and grain.

Examples

if (FALSE) { # \dontrun{
cal <- as_calendar(
  df,
  groups   = "coverage",
  calendar = "cy_m",
  loss     = "incr_loss",
  premium  = "incr_premium"
)
smr  <- summary(cal)
head(smr)
} # }