Abstract
Suppose an unknown isotonic function f, defined on the unit cube in m-dimensional space Rm, is represented only by a data sequence contaminated by noise. Assuming independence and bounds on the first and second absolute central moments of the errors, we show that every estimator of f based on discrete isotonic L1 approximation of the data is strongly consistent.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have