In this paper we study minimax and adaptation rates in general isotonic regression. For uniform deterministic and random designs in with and noise, the minimax rate for the risk is known to be bounded from below by when the unknown mean function is nondecreasing and its range is bounded by a constant, while the least squares estimator (LSE) is known to nearly achieve the minimax rate up to a factor where is sample size, in the lattice design and in the random design. Moreover, the LSE is known to achieve the adaptation rate when is piecewise constant on hyperrectangles in a partition of . Due to the minimax theorem, the LSE is identical on every design point to both the max-min and min-max estimators over all upper and lower sets containing the design point. This motivates our consideration of estimators which lie in-between the max-min and min-max estimators over possibly smaller classes of upper and lower sets, including a subclass of block estimators. Under a -th moment condition on the noise, we develop risk bounds for such general estimators for isotonic regression on graphs. For uniform deterministic and random designs in with , our risk bound for the block estimator matches the minimax rate when the range of is bounded and achieves the near parametric adaptation rate when is -piecewise constant. Furthermore, the block estimator possesses the following oracle property in variable selection: When depends on only a subset of variables, the risk of the block estimator automatically achieves up to a poly-logarithmic factor the minimax rate based on the oracular knowledge of .
View on arXiv