Example reporting triangle with downward corrections

A synthetic reporting triangle demonstrating downward corrections at a specific delay. This represents a realistic case where data quality reviews at delay 2 consistently identify false positives or reclassify cases, resulting in net downward adjustments that produce negative values.

When estimated with preprocess = NULL, this triangle produces a PMF with negative entries and a CDF that is not strictly increasing, reflecting the downward correction process.

Usage

example_downward_corr_mat

Format

A matrix with 8 rows and 4 columns.

Rows represent reference dates (time points when events occurred). Columns represent reporting delays (0 to 3 days). Values represent counts, with negative values at delay 2 representing downward corrections. Lower-right triangle contains NA values (unobserved future reports).

Details

This example demonstrates relaxed assumptions for PMF and CDF when working with downward corrections:

• With preprocess = NULL, the PMF can have negative entries

• The CDF may not be strictly increasing

• This reflects real reporting processes with systematic downward corrections

See also

• estimate_delay() with preprocess = NULL to preserve negative entries

• preprocess_negative_values() to handle negatives by redistribution

Example datasets germany_covid19_hosp, syn_nssp_df, syn_nssp_line_list