Comparing Association Rules and Deep Neural Networks on Medical Data

Date

2019-12

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

Deep neural networks are today's most popular tool for building predictive models across various different disciplines. A decade ago, the most popular predictive modeling technique was association rule mining. In this work, we carefully compare these two techniques in an effort to identify a more effective model with which to predict heart disease, a multi-prediction problem. Both techniques require significant knowledge, manual tuning, and experimentation to determine optimal parameters. Our goal was to build a predictive model that is at least as good as the best association rules across our entire data set. Promising results were obtained for some examples, while others still remain unclear. Making predictive models with medical data continues to be a challenging problem to solve that requires more attention from the scientific community.

Description

Keywords

Neural networks, Neurosciences, Heart disease, Association rules

Citation