Algorithms and error estimations for monotone regression on partially preordered sets (Q2373448)

Monotone (or isotopic) regression is used in many fields, including data analysis (multidimensional scaling, describing ordinal data by cardinal methods, etc.), archaeology, and isotonic likelihood estimation. In many cases, the monotonicity is only defined for a partial instead of a total preorder. No efficient algorithm is known which solves the general problem in a finite number of steps. In this paper, some error estimations are given for an approximate solution of the optimum. In addition, some results concerning monotone regression and the treatment of missing values are presented.