cocoResid: Residual Based Model Assessment Procedure

In coconots: Convolution-Closed Models for Count Time Series

Residual Based Model Assessment Procedure

Description

Calculates the (Pearson) residuals of a fitted model for model evaluation purposes.

Usage

cocoResid(coco, val.num = 1e-10)

Arguments

coco	An object of class "coco
val.num	A non-negative real number that halts the calculation once the cumulative probability reaches 1-val.num

Details

The Pearson residuals are computed as the scaled deviation of the observed count from its conditional expectation given the relevant past history, including covariates, if applicable. If a fitted model is correctly specified, the Pearson residuals should exhibit mean zero, variance one, and no significant serial correlation.

Value

a list that includes the (Pearson) residuals, conditional expectations, conditional variances, and information on the model specifications.

Author(s)

Manuel Huth

coconots documentation built on Aug. 22, 2025, 5:17 p.m.