1. The proposed method V2 did not perform well as expected. The author did not give clear reasons about it. I guess the weighting function is not estimated well adaptively. It seems the weights are not balanced among dimensions (see Table 8).
2. The final estimates of weights were used to compute the final scores. Because of weights were not balanced, the final scores may be not reliable.
3. The prominent findings may be that the Segall’s method tends to balance items among every dimension, similar to Ag method.
4. We can see from table 2, the difficult parameters were poorly estimated. But no reasons were given to it. The results of CAT are not reliable.
5. The test reliability seems overestimate the true reliability (correlation between true theta and estimated theta), especially when test length is larger.