Kernel classification rules from missing data

From MaRDI portal





The author studied nonparametric kernel classification rules when data are missing from training sets. Two treatments of missing data are investigated: 1. Ignore the missing data and use the regular kernel classification rule on the known data. 2. Estimate the missing data by a weighted sum of known data and replace the missing data by the estimated ones and then use the regular classification rule. Convergence of missclassification probabilities is studied. Simulation and comparison studies are carried out in this paper.











This page was built for publication: Kernel classification rules from missing data

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4277156)