Statistics API · phiz

These functions turn model results from phit and draws from CmdStanPHP into plots. Each returns an ordinary Plot or Compose, which any plot method can restyle. Fits are phit’s fit objects (Phit\Fit, from phit 0.6) or arrays in the same shape. Statistics describes the keys each one reads and shows each one.

Functions: coefplot() · residual_plot() · qq_plot() · fitted_plot() · scale_location_plot() · diagnostic_plots() · effect_plot() · predictions() · interval_plot() · marginal_means() · random_effects_plot()

`coefplot()`

coefplot(object|array $fits, mixed ...$options): Plot

Coefficient plot of one fit, or of several keyed by label (dodged side by side).

`residual_plot()`

residual_plot(object|array $fit): Plot

Residuals against fitted values, with a reference line at zero.

`qq_plot()`

qq_plot(object|array $fitOrValues, bool $standardize = false): Plot

A normal QQ plot of a fit’s residuals, or of a list of values; standardize scales them first.

`fitted_plot()`

fitted_plot(object|array $fit, ?array $observed = null): Plot

Observed against fitted values, with the identity line; observed defaults to fitted plus residuals.

`scale_location_plot()`

scale_location_plot(object|array $fit): Plot

The square root of the absolute residuals divided by sigma, against fitted values, as R’s plot(fit, 3).

`diagnostic_plots()`

diagnostic_plots(object|array $fit, mixed ...$options): Compose

residual_plot(), qq_plot(), scale_location_plot(), and fitted_plot() in a two-by-two grid.

`effect_plot()`

effect_plot(object|array $fit, mixed $data, string $x, mixed ...$options): Plot

Predicted values with confidence bands as x varies (by levels of $by), other variables held at typical values.

`predictions()`

predictions(object|array $fit, array $grid, mixed ...$options): array

The predicted rows behind effect_plot() for a grid of rows, without plotting them: each row with estimate, lower, upper, and std_error.

`interval_plot()`

interval_plot(array $draws, mixed ...$options): Plot

Posterior intervals from draws columns or a summary, as bayesplot’s mcmc_intervals.

`marginal_means()`

marginal_means(    object|array $fit,    mixed $data,    string $factor,    mixed ...$options,): Plot

Estimated marginal means of a factor with confidence intervals, as emmeans.

`random_effects_plot()`

random_effects_plot(object|array $fit, mixed $data = null, mixed ...$options): Plot

A caterpillar plot of a mixed model’s random effects with conditional intervals, as lme4’s dotplot(ranef()).