Abstract
We show that extending the Gaussian distribution to the domain of graphs corresponds to truncated Gaussian distributions in Euclidean spaces. Based on this observation, we derive a maximum likelihood method for estimating the parameters of the Gaussian on graphs. In conjunction with a naive Bayes classifier, we applied the proposed approach to image classification. © 2011 Springer-Verlag Berlin Heidelberg.
Cite
CITATION STYLE
Jain, B. J., & Obermayer, K. (2011). Maximum likelihood for gaussians on graphs. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 6658 LNCS, 62–71. https://doi.org/10.1007/978-3-642-20844-7_7
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
