Visual Neural Decomposition to Explain Multivariate Data Sets

TL;DR

This paper introduces a neural network-based visualization method that helps data analysts understand complex relationships in high-dimensional data sets by making the model's inner workings interpretable and scalable to hundreds of variables.

Contribution

It presents a novel neural network visualization approach with a new regularization technique for better interpretability in high-dimensional data analysis.

Findings

Effective visualization of variable relationships in large data sets

Neural network models with improved interpretability

Successful application to artificial and real-world data

Abstract

Investigating relationships between variables in multi-dimensional data sets is a common task for data analysts and engineers. More specifically, it is often valuable to understand which ranges of which input variables lead to particular values of a given target variable. Unfortunately, with an increasing number of independent variables, this process may become cumbersome and time-consuming due to the many possible combinations that have to be explored. In this paper, we propose a novel approach to visualize correlations between input variables and a target output variable that scales to hundreds of variables. We developed a visual model based on neural networks that can be explored in a guided way to help analysts find and understand such correlations. First, we train a neural network to predict the target from the input variables. Then, we visualize the inner workings of the resulting model to help understand relations within the data set. We further introduce a new regularization term for the backpropagation algorithm that encourages the neural network to learn representations that are easier to interpret visually. We apply our method to artificial and real-world data sets to show its utility.