ggm_compare_estimate: GGM Compare: Estimate

Description Usage Arguments Details Value Note References Examples

View source: R/ggm_compare_estimate.default.R

Description

Compare partial correlations that are estimated from any number of groups. This method works for continuous, binary, ordinal, and mixed data (a combination of categorical and continuous variables). The approach (i.e., a difference between posterior distributions) was described in \insertCiteWilliams2019;textualBGGM.

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
ggm_compare_estimate(
  ...,
  formula = NULL,
  type = "continuous",
  mixed_type = NULL,
  analytic = FALSE,
  prior_sd = 0.5,
  iter = 5000,
  impute = TRUE,
  progress = TRUE,
  seed = 1
)

Arguments

...

Matrices (or data frames) of dimensions n (observations) by p (variables). Requires at least two.

formula

An object of class formula. This allows for including control variables in the model (i.e., ~ gender). See the note for further details.

type

Character string. Which type of data for Y ? The options include continuous, binary, ordinal, or continuous. See the note for further details.

mixed_type

Numeric vector. An indicator of length p for which varibles should be treated as ranks. (1 for rank and 0 to use the 'empirical' or observed distribution). The default is currently to treat all integer variables as ranks when type = "mixed" and NULL otherwise. See note for further details.

analytic

Logical. Should the analytic solution be computed (default is FALSE)? This is only available for continous data. Note that if type = "mixed" and analytic = TRUE, the data will automatically be treated as continuous.

prior_sd

The scale of the prior distribution (centered at zero), in reference to a beta distribtuion (defaults to 0.50). See note for further details.

iter

Number of iterations (posterior samples; defaults to 5000).

impute

Logicial. Should the missing values (NA) be imputed during model fitting (defaults to TRUE) ?

progress

Logical. Should a progress bar be included (defaults to TRUE) ?

seed

An integer for the random seed.

Details

This function can be used to compare the partial correlations for any number of groups. This is accomplished with pairwise comparisons for each relation. In the case of three groups, for example, group 1 and group 2 are compared, then group 1 and group 3 are compared, and then group 2 and group 3 are compared. There is a full distibution for each difference that can be summarized (i.e., summary.ggm_compare_estimate) and then visualized (i.e., plot.summary.ggm_compare_estimate). The graph of difference is selected with select.ggm_compare_estimate).

Controlling for Variables:

When controlling for variables, it is assumed that Y includes only the nodes in the GGM and the control variables. Internally, only the predictors that are included in formula are removed from Y. This is not behavior of, say, lm, but was adopted to ensure users do not have to write out each variable that should be included in the GGM. An example is provided below.

Mixed Type:

The term "mixed" is somewhat of a misnomer, because the method can be used for data including only continuous or only discrete variables. This is based on the ranked likelihood which requires sampling the ranks for each variable (i.e., the data is not merely transformed to ranks). This is computationally expensive when there are many levels. For example, with continuous data, there are as many ranks as data points!

The option mixed_type allows the user to determine which variable should be treated as ranks and the "emprical" distribution is used otherwise. This is accomplished by specifying an indicator vector of length p. A one indicates to use the ranks, whereas a zero indicates to "ignore" that variable. By default all integer variables are handled as ranks.

Dealing with Errors:

An error is most likely to arise when type = "ordinal". The are two common errors (although still rare):

Imputing Missing Values:

Missing values are imputed with the approach described in \insertCitehoff2009first;textualBGGM. The basic idea is to impute the missing values with the respective posterior pedictive distribution, given the observed data, as the model is being estimated. Note that the default is TRUE, but this ignored when there are no missing values. If set to FALSE, and there are missing values, list-wise deletion is performed with na.omit.

Value

A list of class ggm_compare_estimate containing:

Note

Mixed Data:

The mixed data approach was introduced \insertCite@in @hoff2007extending;textualBGGM (our paper describing an extension to Bayesian hypothesis testing if forthcoming). This is a semi-paramateric copula model based on the ranked likelihood. This is computationally expensive when treating continuous data as ranks. The current default is to treat only integer data as ranks. This should of course be adjusted for continous data that is skewed. This can be accomplished with the argument mixed_type. A 1 in the numeric vector of length pindicates to treat that respective node as a rank (corresponding to the column number) and a zero indicates to use the observed (or "emprical") data.

It is also important to note that type = "mixed" is not restricted to mixed data (containing a combination of categorical and continuous): all the nodes can be ordinal or continuous (but again this will take some time).

Interpretation of Conditional (In)dependence Models for Latent Data:

See BGGM-package for details about interpreting GGMs based on latent data (i.e, all data types besides "continuous")

Additional GGM Compare Methods

Bayesian hypothesis testing is implemented in ggm_compare_explore and ggm_compare_confirm \insertCiteWilliams2019_bfBGGM. The latter allows for confirmatory hypothesis testing. An approach based on a posterior predictive check is implemented in ggm_compare_ppc \insertCitewilliams2020comparingBGGM. This provides a 'global' test for comparing the entire GGM and a 'nodewise' test for comparing each variable in the network \insertCiteWilliams2019;textualBGGM.

References

\insertAllCited

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
# note: iter = 250 for demonstrative purposes

# data
Y <- bfi

# males and females
Ymale <- subset(Y, gender == 1,
                   select = -c(gender,
                               education))[,1:10]

Yfemale <- subset(Y, gender == 2,
                     select = -c(gender,
                                 education))[,1:10]

# fit model
fit <- ggm_compare_estimate(Ymale,  Yfemale,
                           type = "ordinal",
                           iter = 250,
                           prior_sd = 0.25,
                           progress = FALSE)

###########################
### example 2: analytic ###
###########################
# only continuous

# fit model
fit <- ggm_compare_estimate(Ymale, Yfemale,
                            analytic = TRUE)

# summary
summ <- summary(fit)

# plot summary
plt_summ <- plot(summary(fit))

# select
E <- select(fit)

# plot select
plt_E <- plot(select(fit))

BGGM documentation built on Aug. 20, 2021, 5:08 p.m.