Jing Wang, University of Connecticut

• Title: Unweighted esitmation based on optimal sample under measurement constraints

• Abstract: Subsampling is a practical approach to extract information from massive data. However, when responses are expensive to measure, developing subsampling schemes is challenging. The estimating efficiency of the existing method under this scenario can be improved for using reweighted estimator. We proposed an unweighted estimator to obtain a more efficient estimator. Asymptotical results obtained via martingale techniques and numerical experiments verified the better performance of our method.