Tensor-based Kernel Machines with Structured Inducing Points for Large and High-Dimensional Data

This repository allows for the reproduction of the numerical experiments in Tensor-based Kernel Machines with Structured Inducing Points for Large and High-Dimensional Data.

Requirements

Install the GPLM toolbox for results related to the UCI benchmark. All former UCI datasets can be downloaded from the webpage of Andrew Wilson. The SUSY dataset can be downloaded from the UCI Machine Learning Repository.

Training and Evaluation

To train and test the models in the paper on the various datasets, it is sufficient to execute the corresponding homonymous scripts. All the necessary functions are located in the appropriate folder.