stap_glm: Bayesian generalized spatial-temporal aggregated...

In rstap: Spatial Temporal Aggregated Predictor Models via 'stan'

Description

Generalized linear modeling with spatial temporal aggregated predictors using prior distributions for the coefficients, intercept, spatial-temporal scales, and auxiliary parameters.

Usage

stap_glm(formula, family = gaussian(), subject_data = NULL,
  distance_data = NULL, time_data = NULL, subject_ID = NULL,
  max_distance = NULL, max_time = NULL, weights, offset = NULL,
  model = TRUE, y = TRUE, contrasts = NULL, ..., prior = normal(),
  prior_intercept = normal(), prior_stap = normal(),
  prior_theta = log_normal(location = 1L, scale = 1L),
  prior_aux = exponential(), adapt_delta = NULL)

stap_lm(formula, family = gaussian(), subject_data = NULL,
  distance_data = NULL, time_data = NULL, subject_ID = NULL,
  max_distance = NULL, max_time = NULL, weights, offset = NULL,
  model = TRUE, y = TRUE, contrasts = NULL, ..., prior = normal(),
  prior_intercept = normal(), prior_stap = normal(),
  prior_theta = log_normal(location = 1L, scale = 1L),
  prior_aux = exponential(), adapt_delta = NULL)

Arguments

formula	Same as for glm. Note that in-formula transformations will not be passed ot the final design matrix. Covariates that have "scale" in their name are not advised as this text is parsed for in the final model fit.
family	Same as glm for gaussian, binomial, and poisson families.
subject_data	a data.frame that contains data specific to the subject or subjects on whom the outcome is measured. Must contain one column that has the subject_ID on which to join the distance and time_data
distance_data	a (minimum) three column data.frame that contains (1) an id_key (2) The sap/tap/stap features and (3) the distances between subject with a given id and the built environment feature in column (2), the distance column must be the only column of type "double" and the sap/tap/stap features must be specified in the dataframe exactly as they are in the formula.
time_data	same as distance_data except with time that the subject has been exposed to the built environment feature, instead of distance
subject_ID	name of column(s) to join on between subject_data and bef_data
max_distance	the inclusion distance; upper bound for all elements of dists_crs
max_time	inclusion time; upper bound for all elements of times_crs
offset, weights	Same as glm.
model	logical denoting whether or not to return the fixed covariates model frame object in the fitted object
y	In stap_glm, logical scalar indicating whether to return the response vector. In stan_glm.fit, a response vector.
contrasts	Same as glm, but rarely specified.
...	Further arguments passed to the function in the rstap to specify iter, chains, cores, refresh, etc.
prior	The prior distribution for the regression coefficients. prior should be a call to one of the various functions provided by rstap for specifying priors. The subset of these functions that can be used for the prior on the coefficients can be grouped into several "families": Family Functions Student t family normal, student_t, cauchy Hierarchical shrinkage family hs, hs_plus Laplace family laplace, lasso Product normal family product_normal See the priors help page for details on the families and how to specify the arguments for all of the functions in the table above. To omit a prior —i.e., to use a flat (improper) uniform prior— prior can be set to NULL, although this is rarely a good idea. Note: If prior is from the Student t family or Laplace family, and if the autoscale argument to the function used to specify the prior (e.g. normal) is left at its default and recommended value of TRUE, then the default or user-specified prior scale(s) may be adjusted internally based on the scales of the predictors. See the priors help page and the Prior Distributions vignette for details on the rescaling and the prior_summary function for a summary of the priors used for a particular model.
prior_intercept	The prior distribution for the intercept. prior_intercept can be a call to normal, student_t or cauchy. See the priors help page for details on these functions. To omit a prior on the intercept —i.e., to use a flat (improper) uniform prior— prior_intercept can be set to NULL. Note: The prior distribution for the intercept is set so it applies to the value when all predictors are centered. If you prefer to specify a prior on the intercept without the predictors being auto-centered, then you have to omit the intercept from the formula and include a column of ones as a predictor, in which case some element of prior specifies the prior on it, rather than prior_intercept. Regardless of how prior_intercept is specified, the reported estimates of the intercept always correspond to a parameterization without centered predictors (i.e., same as in glm).
prior_stap	prior for spatial-temporal aggregated predictors. Note that prior is set on the standardized latent covariates.
prior_theta	prior for the spatial-temporal aggregated predictors' scale. Can either be a single prior or a prior nested within a list of lists.
prior_aux	The prior distribution for the "auxiliary" parameter (if applicable). The "auxiliary" parameter refers to a different parameter depending on the family. For Gaussian models prior_aux controls "sigma", the error standard deviation. For negative binomial models prior_aux controls "reciprocal_dispersion", which is similar to the "size" parameter of rnbinom: smaller values of "reciprocal_dispersion" correspond to greater dispersion. For gamma models prior_aux sets the prior on to the "shape" parameter (see e.g., rgamma), and for inverse-Gaussian models it is the so-called "lambda" parameter (which is essentially the reciprocal of a scale parameter). Binomial and Poisson models do not have auxiliary parameters. prior_aux can be a call to exponential to use an exponential distribution, or normal, student_t or cauchy, which results in a half-normal, half-t, or half-Cauchy prior. See priors for details on these functions. To omit a prior —i.e., to use a flat (improper) uniform prior— set prior_aux to NULL.
adapt_delta	See the adapt_delta help page for details.

Details

The stap_glm function is similar in syntax to stan_glm except instead of performing full bayesian inference for a generalized linear model stap_glm incorporates spatial-temporal covariates

Value

A stapreg object is returned for stap_glm.

A stapfit object (or a slightly modified stapfit object) is returned if stan_glm.fit is called directly.

References

Gelman, A. and Hill, J. (2007). Data Analysis Using Regression and Multilevel/Hierarchical Models. Cambridge University Press, Cambridge, UK.

Muth, C., Oravecz, Z., and Gabry, J. (2018) User-friendly Bayesian regression modeling: A tutorial with rstanarm and shinystan. The Quantitative Methods for Psychology. 14(2), 99–119. https://www.tqmp.org/RegularArticles/vol14-2/p099/p099.pdf

See Also

stapreg-methods and glm.

The various vignettes for stap_glm at https://biostatistics4socialimpact.github.io/rstap/articles and the preprint article.

Examples

fit_glm <- stap_glm(formula = y ~ sex + sap(Fast_Food),
                   subject_data = homog_subject_data[1:100,], # for speed of example only
                     distance_data = homog_distance_data,
                     family = gaussian(link = 'identity'),
                     subject_ID = 'subj_id',
                     prior = normal(location = 0, scale = 5, autoscale = FALSE),
                     prior_intercept = normal(location = 25, scale = 5, autoscale = FALSE),
                     prior_stap = normal(location = 0, scale = 3, autoscale = FALSE),
                     prior_theta = log_normal(location = 1, scale = 1),
                     prior_aux = cauchy(location = 0,scale = 5),
                     max_distance = max(homog_distance_data$Distance),
                     chains = 1, iter = 300, # for speed of example only
                     refresh = -1, verbose = FALSE) 

rstap documentation built on May 1, 2019, 9:21 p.m.