Kernel Domain Description with Incomplete Data: Using Instance-Specific Margins to Avoid Imputation.

Adam Gripton Weiping Lu

Published in: ICPR (2010)

Keyphrases

incomplete data
missing values
missing data
multiple imputation
learning bayesian networks
em algorithm
incomplete data sets
bayesian networks
domain description
feature selection
feature space
maximum likelihood
machine learning
general purpose
training set
support vector
artificial intelligence