|
| << vorige | volgende >> |
Tijdschrift beschrijving
Alle jaargangen van het bijbehorende tijdschrift
Alle afleveringen van het bijbehorende jaargang
Alle artikelen van de bijbehorende aflevering
Details van artikel 5 van 7 gevonden artikelen
| Titel: |
Linear Near-Neighbor Classifiers for Correlated Categories |
| Auteur: |
Devore, Jay L. |
| Verschenen in: |
Communications in statistics |
| Paginering: | Jaargang 4 (1975) nr. 10 pagina's 891-905 |
| Jaar: |
1975 |
| Inhoud: |
We consider a classification problem in which θ̵ = θ1, θ2, … is a sequence of unknown categories, with each θi. = 0 or 1, and X̵ = X1, X2, … is a sequence of observed random variables. The conditional distribution of Xi. given θi. is assumed normal with variance 1 and mean 2θi - 1. We assume that the θ̵ sequence is a realization of a two-state stationary Markov chain. Thus there is information about θi. contained not only in Xi. but in all other Xj's. Here we study various properties of classification rules for θi. based on linear functions of the Xj's with |j-i| small. When the values of the Markovian parameters are unspecified, it is natural to consider using a plug-in rule based on substitution of parameter estimates. The large sample behavior of parameter estimates and plug-in rules is investigated, with particular attention to the probability that the plug-in rule yields a smaller risk than the usual minimax procedure appropriate for uncorrelated categories. |
| Uitgever: |
Taylor & Francis |
| Bronbestand: |
Elektronische Wetenschappelijke Tijdschriften |
Details van artikel 5 van 7 gevonden artikelen
| << vorige | volgende >> |