trendtestR (version 1.0.1)

diagnose_model_trend: Diagnose a fitted model using residual plots and statistical tests (ggplot2 only) / Modell-Diagnose mittels Residuenplots und statistischen Tests (nur ggplot2)

Description

Diagnose model fit for lm, glm, gam (mgcv), and zeroinfl (pscl) models using residual plots and tests. / Diagnose lineare Modelle (lm), generalisierte lineare Modelle (glm), GAMs von mgcv und Zero-Inflated-Modelle von pscl mit ggplot2.

Usage

diagnose_model_trend(
  model,
  value_col = "value",
  residual_type = "deviance",
  group_col = NULL,
  verbose = TRUE
)

Value

A list with ggplot2 plots and diagnostic test results. / Eine Liste mit ggplot2-Plots und diagnostischen Testergebnissen:

plots

A named list with residual plots ("residuals_vs_fitted", "qq", "scale_location"). / Eine Liste mit Residuenplots ("residuals_vs_fitted", "qq", "scale_location").

diagnostics

A named list of statistical test results ("Shapiro", "KS", "Levene", "GAM check"). / Eine Liste mit Ergebnissen statistischer Tests ("Shapiro", "KS", "Levene", "GAM check").

Arguments

model

A fitted model object (lm, glm, gam, or zeroinfl). / Ein angepasstes Modellobjekt (lm, glm, gam oder zeroinfl).

value_col

Name of the response variable (used in axis labels). / Name der Zielvariable (verwendet fuer Achsenbeschriftungen).

residual_type

Type of residuals to use ("deviance", "pearson", "response", etc.). / Art der Residuen ("deviance", "pearson", "response" usw.).

group_col

Optional. Grouping variable to color residual plots. / Optional. Gruppierungsvariable fuer Farbgebung in den Residuenplots.

verbose

Logical; whether to print diagnostic messages. / Logisch; ob Diagnosenachrichten ausgegeben werden sollen.

See Also

[explore_poisson_trend], [explore_continuous_trend], [explore_zinb_trend], [explore_trend_auto]

Examples

# Example for a linear model
set.seed(123)
df <- data.frame(x = rnorm(100), y = rnorm(100))
model_lm <- lm(y ~ x, data = df)
diagnose_model_trend(model_lm)

# Beispiel fuer ein GLM
df_glm <- data.frame(x = rnorm(100), y = rpois(100, lambda = 2))
model_glm <- glm(y ~ x, data = df_glm, family = poisson())
diagnose_model_trend(model_glm)