2013 | OriginalPaper | Buchkapitel
Class Based Weighted K-Nearest Neighbor over Imbalance Dataset
verfasst von : Harshit Dubey, Vikram Pudi
Erschienen in: Advances in Knowledge Discovery and Data Mining
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
K
-Nearest Neighbor based classifier classifies a query instance based on the class labels of its neighbor instances. Although
k
NN has proved to be a ubiquitous classification tool with good scalability, but it suffers from some drawbacks. The existing
k
NN algorithm is equivalent to using only local prior probabilities to predict instance labels, and hence it does not take into account the class distribution around neighborhood of the query instance, which results into undesirable performance on imbalanced data. In this paper, a modified version of
k
NN algorithm is proposed so that it takes into account the class distribution in a wider region around the query instance. Our empirical experiments with several real world datasets show that our algorithm outperforms current state-of-the-art approaches.