The two new item selection methods extended the posterior weighted Kullback-Leibler (PWKL) for CDM by changing the disterministic selection algorithm to include stochastic components either in the item selection index or in the item selection procedure. The exposure rate of items and the usage of underexposed items were compared through simulations.
I have some questions for the paper:
(1) How to keep the maximum exposure rate through equation 10? Since there is no enough description/discussion for the equation.
(2) The importance parameter belta is really important! But how to decide the value of b is arbitrary. What’s the difference if I choose belta 2 and 2.5, for example?
(3)It seemed from the current studies that both of the two new methods are good. But, which is better? in what kind of context?
(4) Compare to Barrada et al. (2008), the new methods in the current studies are incorporate with restriction on the maximum exposure rate. The effect of controlling overexposure is apparent in the long test. But how about in the short test?