Generalized RBF kernel for incomplete data

2016/12/05 by Struski, Łukasz, Śmieja, Marek, Tabor, Jacek
#FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML)

paper · doi:10.48550/arxiv.1612.01480

Abstract

We construct \bf genRBF kernel, which generalizes the classical Gaussian RBF kernel to the case of incomplete data. We model the uncertainty contained in missing attributes making use of data distribution and associate every point with a conditional probability density function. This allows to embed incomplete data into the function space and to define a kernel between two missing data points based on scalar product in L₂. Experiments show that introduced kernel applied to SVM classifier gives better results than other state-of-the-art methods, especially in the case when large number of features is missing. Moreover, it is easy to implement and can be used together with any kernel approaches with no additional modifications.

Related

• Cluster based RBF Kernel for Support Vector Machines
• Explicit Approximations of the Gaussian Kernel
• Probabilistic Kernel Support Vector Machines
• Spectral Clustering using PCKID - A Probabilistic Cluster Kernel for\n Incomplete Data
• Approximating Incomplete Kernel Matrices by the em Algorithm