select the next item by maximum Fisher information

Usage

compute_next_item_rsm(theta, bank, out = NULL, D = 1, randomesque = 1)

Arguments

theta: current ability estimate
bank: matrix of item parameters
out: vector of the indices of the items already administered
D: metric constant can be 1 or 1.702
randomesque: number of most informative items to draw at random from
1 is pure maximum information and 3 to 5 spreads item exposure

Value

item index of the selected item in the bank
info information of the selected item at theta
available indices of the items still available

Note

this function should return the same result as
catR::nextItem(itemBank,model="RSM",theta,out,criterion="MFI") when
randomesque is 1
under the rating scale model this is equivalent to picking the item whose
lambdaj sits closest to theta because all information curves are identical
in shape so the maximum is a pure nearest neighbour lookup on lambdaj

Examples

bank<-matrix(c(-0.560,-0.230,1.559,0.071,0.129,
               rep(1.715,5),rep(0.461,5),rep(-1.265,5),rep(-0.687,5)),
             nrow=5,
             dimnames=list(c(1,2,3,4,5),
                           c("lambdaj","delta1","delta2","delta3","delta4")))
catR::nextItem(bank,model="RSM",theta=0,criterion="MFI") # with catR installed
#> $item
#> [1] 4
#> 
#> $par
#> lambdaj  delta1  delta2  delta3  delta4 
#>   0.071   1.715   0.461  -1.265  -0.687 
#> 
#> $info
#> [1] 3.061434
#> 
#> $criterion
#> [1] "MFI"
#> 
#> $randomesque
#> [1] 1
#> 
#> $name
#> [1] "4"
#> 
#> $prior.prop
#> [1] NA
#> 
#> $post.prop
#> [1] NA
#> 
#> $cb.prop
#> [1] NA
#> 
compute_next_item_rsm(theta=0,bank)
#> $item
#> [1] 4
#> 
#> $info
#> [1] 3.061434
#> 
#> $available
#> [1] 1 2 3 4 5
#>