R/immer_jml_update_item_Rcpp.R

In immer: Item Response Models for Multiple Ratings

## File Name: immer_jml_update_item_Rcpp.R
## File Version: 0.463


immer_jml_update_item_Rcpp <- function( score_items, ItemScore, I, K, b, A, xsi, theta,
    N, dat_resp, max_incr, maxiter_update, conv_update, b_fixed, shortcut_index, weights,
    use_weights=FALSE)
{
    iterate <- TRUE
    iter <- 0
    eps <- 1E-20
    NX <- length(xsi)
    A_ <- as.vector(A)
    xsi_converged <- rep(0,NX)
    max_incr <- rep(5,NX)
    update <- as.vector(shortcut_index$update)
    update_weights <- as.vector(shortcut_index$update_weights)

    if (TRUE){
        update <- rep(1,length(update))
        update_weights <- weights
    }

    while(iterate){
        b0 <- b
        xsi0 <- xsi

        #-- derivatives with respect to b and xsi
        res <- immer_jml_update_item_derivatives( theta=theta, score_items=score_items, N=N,
                    K=K, I=I, dat_resp=dat_resp, b=b, A_=A_, xsi=xsi, max_incr=max_incr, b_fixed=b_fixed,
                    ItemScore=ItemScore, update=update, update_weights=update_weights )
        der2_xsi <- res$der2_xsi
        b <- res$b
        xsi <- res$xsi
        incr <- res$incr
        max_incr <- abs( xsi - xsi0)
        xsi_converged <- ( abs( xsi - xsi0 ) < conv_update )
        iter <- iter + 1
        b_change <- max( abs( b - b0) )
        if (iter >=maxiter_update){ iterate <- FALSE }
        if (b_change < conv_update){ iterate <- FALSE }
    }
    #--- output
    res <- list(b=b, xsi=xsi, xsi_der2=der2_xsi )
    return(res)
}