fim_logit: A function to calculate FIM entries

Description Usage Arguments Details

View source: R/fim_logit.R

Description

This function calculates the finsher information matrix of a 1st order logit model.

Usage

1
fim_logit(dataNumeric, betaDotX, xIndices, yIndex)

Arguments

dataNumeric

Numerical format of original data, where variable values start from 0.

betaDotX

Inner product of parameters beta and input variables X.

xIndices

A vector of indices of the input variables.

y

The index of the output/target variable.

Details

FIM is a square matrix with dimensions equal to |beta| = |xIndices| + 1 for binary logit models. rowID and colID in FIM corresponds to rowID-1 and colID-1 in xIndices, because the first row and column of FIM is the 2nd derivative w.r.t. beta_0. There is no need to add a column of 1s in front of dataNumeric, because the function start filling in entries from 2nd row and 2nd column. The 1st row and column are filled in separately.


kelvinyangli/mbmml documentation built on June 29, 2020, 3:12 a.m.