dsProbitRegr: Calculate Probit Regression on Server

In difuture-lmu/dsROCGLM: ROC and calibration analysis for DataSHIELD

Calculate Probit Regression on Server

Description

Distributed computing of the probit regression using DataSHIELD. The method gets the parts required for the fisher scoring in each iteration and aggreagtes it to a global update. Hence, the updating step is done at the analyst's machine while calculation of the required parts are done at the servers.

Usage

dsProbitRegr(
  connections,
  formula,
  data,
  w = NULL,
  stop_tol = 1e-08,
  iter_max = 25L,
  trace = FALSE
)

Arguments

connections	('DSI::connection') Connection to an OPAL server.
formula	('character(1L)') Formula used for the probit regression.
data	('character(1L)') Data as string.
w	('character(1L)') Weights for the probit regression. Must be a column in 'data'.
stop_tol	('numeric(1L)') Tolerance for the stop criteria ('abs(dev - dev_old) / (abs(dev) + 0.1)') for which the Fischer scoring stops (default is '1e-8').
iter_max	('integer(1L)') Maximal number of Fischer scoring iterations (default is '25L').
trace	('logical(1L)') If 'TRUE' (default), information about the progress is shown.

Value

List with estimated parameter, number of iterations, and the deviance at the stopping iteration.

Author(s)

Daniel S.

difuture-lmu/dsROCGLM documentation built on March 24, 2024, 1:07 p.m.