bigmlogit: Multinomial logit model for big data sets

In bigmlogit: Estimation of models with limited dependent variables

Description

Estimation by maximum likelihood of the multinomial logit model, with alternative-specific and/or individual specific variables, for big data sets (millions of individuals and dozens of alternatives)

Usage

bigmlogit(formula, data, chunksize = 1E3, alternatives = NULL,
         mixed = NULL, fitted = NULL, ...)
## S3 method for class 'bigmlogit'
mean(x, ...)
## S3 method for class 'bigmlogit'
predict(object, data, alternatives = NULL, mixed = NULL, 
    print.level = 0, ...) 

Arguments

object	a bigmlogit object,
x	a bigmlogit object,
formula	a symbolic description of the model to be estimated,
data	the data: a connection to a data base,
chunksize	the size of the blocks of data,
alternatives	if NULL, the alternative-specific covariates data frame is stored in the data base ; otherwise, it is provided by this argument,
mixed	if NULL, each mixed covariate is stored as a table in the data base ; otherwise, they are provided by this argument which is a named list containing a data frame for every mixed variable,
print.level	the amont of information printed while computing the predictions,
fitted	the name of the data base which would contain the fitted values,
...	further arguments passed to maxLik.

Value

An object of class c("bigmlogit", "maxLik").

Author(s)

Yves Croissant

See Also

multinom from package nnet performs the estimation of the multinomial logit model with individual specific variables, mlogit from package mlogit with individual and alternative specific variables.

bigmlogit documentation built on May 2, 2019, 6:50 p.m.