Regressor Levels for Bayesian Predictive Response
研究在满秩线性模型中如何选择预测变量向量，以最大化响应变量超过给定阈值的概率，采用贝叶斯框架整合辅助信息并允许对预测因子水平施加约束，适用于工业场景。
The predictor vector for a response variable in a full-rank linear model is chosen so as to maximize the probability that the response exceeds a given threshold. The problem is formulated in the Bayesian framework in a way that incorporates collateral information and allows constraints on the levels of the predictors. Industrial applications are indicated.