Glance at a(n) Mclust object

Glance accepts a model object and returns a tibble::tibble() with exactly one row of model summaries. The summaries are typically goodness of fit measures, p-values for hypothesis tests on residuals, or model convergence information.

Glance never returns information from the original call to the modeling function. This includes the name of the modeling function or any arguments passed to the modeling function.

Glance does not calculate summary measures. Rather, it farms out these computations to appropriate methods and gathers the results together. Sometimes a goodness of fit measure will be undefined. In these cases the measure will be reported as NA.

Glance returns the same number of columns regardless of whether the model matrix is rank-deficient or not. If so, entries in columns that no longer have a well-defined value are filled in with an NA of the appropriate type.

Usage

# S3 method for class 'Mclust'
glance(x, ...)

Arguments

x

An Mclust object return from mclust::Mclust().

...

Additional arguments. Not used. Needed to match generic signature only. Cautionary note: Misspelled arguments will be absorbed in ..., where they will be ignored. If the misspelled argument has a default value, the default value will be used. For example, if you pass conf.lvel = 0.9, all computation will proceed using conf.level = 0.95. Two exceptions here are:

• tidy() methods will warn when supplied an exponentiate argument if it will be ignored.

• augment() methods will warn when supplied a newdata argument if it will be ignored.

Value

A tibble::tibble() with exactly one row and columns:

BIC

Bayesian Information Criterion for the model.

df

Degrees of freedom used by the model.

logLik

The log-likelihood of the model. [stats::logLik()] may be a useful reference.

nobs

Number of observations used.

model

A string denoting the model type with optimal BIC

G

Number mixture components in optimal model

hypvol

If the other model contains a noise component, the value of the hypervolume parameter. Otherwise `NA`.

Examples

# load library for models and data
library(mclust)

# load data manipulation libraries
library(dplyr)
library(tibble)
library(purrr)
library(tidyr)

set.seed(27)

centers <- tibble(
  cluster = factor(1:3),
  # number points in each cluster
  num_points = c(100, 150, 50),
  # x1 coordinate of cluster center
  x1 = c(5, 0, -3),
  # x2 coordinate of cluster center
  x2 = c(-1, 1, -2)
)

points <- centers %>%
  mutate(
    x1 = map2(num_points, x1, rnorm),
    x2 = map2(num_points, x2, rnorm)
  ) %>%
  select(-num_points, -cluster) %>%
  unnest(c(x1, x2))

# fit model
m <- Mclust(points)

# summarize model fit with tidiers
tidy(m)
#> # A tibble: 3 × 6
#>   component  size proportion variance mean.x1 mean.x2
#>       <int> <int>      <dbl>    <dbl>   <dbl>   <dbl>
#> 1         1   101      0.335     1.12  5.01     -1.04
#> 2         2   150      0.503     1.12  0.0594    1.00
#> 3         3    49      0.161     1.12 -3.20     -2.06
augment(m, points)
#> # A tibble: 300 × 4
#>       x1     x2 .class .uncertainty
#>    <dbl>  <dbl> <fct>         <dbl>
#>  1  6.91 -2.74  1          3.98e-11
#>  2  6.14 -2.45  1          1.99e- 9
#>  3  4.24 -0.946 1          1.47e- 4
#>  4  3.54  0.287 1          2.94e- 2
#>  5  3.91  0.408 1          7.48e- 3
#>  6  5.30 -1.58  1          4.22e- 7
#>  7  5.01 -1.77  1          1.06e- 6
#>  8  6.16 -1.68  1          7.64e- 9
#>  9  7.13 -2.17  1          4.16e-11
#> 10  5.24 -2.42  1          1.16e- 7
#> # ℹ 290 more rows
glance(m)
#> # A tibble: 1 × 7
#>   model     G    BIC logLik    df hypvol  nobs
#>   <chr> <int>  <dbl>  <dbl> <dbl>  <dbl> <int>
#> 1 EII       3 -2402. -1175.     9     NA   300