scientific article; zbMATH DE number 3883432

From MaRDI portal
Publication:3217458





[For the entire collection see Zbl 0531.00015.] The risk of plug-in classification rule is investigated for the case when conditional probability densities of observations are unknown and are estimated by the supervised training sample. The asymptotic expansions for the risk are derived which are true for arbitrary regular distribution families both for parametric and nonparametric (Rosenblatt- Parzen, k-NN) decision functions. Optimal decision functions, minimizing the risk, are found by the derived expansions.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3217458)