Least-Absolute-Value Prediction Lines in Closed Form
对于给定的X和Y联合分布，研究使‖Y – h(X)‖期望最小的线性函数h，并给出常能导出显式解的特征形式，通过多个例子展示。
Abstract Abstract For a given joint distribution of X and Y, the linear function h that minimizes the expectation of ‖Y – h(X)‖ is characterized in a form that often leads to explicit solutions, as various examples demonstrate. Key Words: Regression L 1 loss functionAbsolute deviations