About

This paper proposes a learning scheme using convolutional networks trained directly on tabular datasets for binary classification, focusing on learning data structure without meta-features. Experiments show it excels in identifying patterns and, when applied to real-world datasets, suggests optimal classifiers.

Stars

Watchers

Forks