Inference for infinitely many future values using statistical calibration

Many real problems involves a quantity of interest x which is expensive or difficult to measure, a surrogate quantity y which is cheaper or easy to measure, and that y and x are related by, say, a linear regression model. For example, x is the true alcohol level in blood stream while y is the reading on a breathalyzer, of a driver, or x is the number of one-penny coins in a bag while y is the net weight of the bag. Calibration data can be collected to establish the model between y and x, which can then be used to make inference about the infinitely many x’s based the corresponding observed y’s in future. Recently progress in this area (cf. Liu et al, 2016, Han et al, 2016) allows many possibilities of further research.

Informal inquiries can be made to Prof Wei Liu ([Email Address Removed]) with a copy of your curriculum vitae and cover letter. All general enquiries should be directed to [Email Address Removed].