Zscore {BayesLCA}

Evaluating Class Membership of Binary Data

Description

For a fitted model of class blca, and binary data X, the probability of class membership for each data point is provided.

Usage

Zscore(X, fit = NULL, itemprob = NULL, classprob = NULL)

Arguments

X

A binary data matrix. X must have the same number of columns as the data that fit was applied to.

fit

An object of class blca.

itemprob

A matrix of item probabilities, conditional on class membership.

classprob

A vector denoting class membership probability.

Details

Calculation of the probability of class membership for a data point relies on two parameters, class membership and item probability. These may be supplied directly to Zscore, or alternatively, a blca object containing both parameters can be used instead.

Value

A matrix of equal rows to X and with G, the number of classes, columns, where each row is a score denoting the probability of class membership. Each row should therefore sum to 1.

Note

Zscore.internal has the same functionality as Zscore, but is only intended for internal use.

Author(s)

Arthur White

Examples

set.seed(1)
type1 <- c(0.8, 0.8, 0.05, 0.2)
type2 <- c(0.2, 0.2, 0.05, 0.8)
x<- rlca(250, rbind(type1,type2), c(0.5,0.5))

fit <- blca.em(x, 2)
fit$Z ## Unique data types
Zscore(x, fit=fit) ## Whole data set
Zscore(c(0, 1, 1, 0), fit=fit) ## Not in data set
Zscore(x, itemprob=rbind(type1,type2), classprob=c(0.5,0.5))