Conceptual Clustering with Systematic Missing Values

https://doi.org/10.1016/B978-1-55860-247-2.50065-6Get rights and content

Abstract

Human error and equipment glitches may create random missing values in data sets, but patterns of “non-observation” of feature values imply systematic missing values that may be relevant in recognizing certain domain phenomenon. Previous conceptual clustering schemes have dealt with random missing values and simple cases of systematic missing values. This paper presents a modification to the ITERATE algorithm that can derive meaningful structure from data sets with random and systematic missing values, even if the systematic missing values form a large portion of the data set.

References (0)

Cited by (6)

View full text