Optimizing the Information Retrieval Trade-off in Data Visualization Using $\alpha$-Divergence

TL;DR

This paper introduces a systematic approach to optimize the trade-off between precision and recall in data visualization by using $oldsymbol{ extalpha}$-divergence, eliminating the need for manual tuning and providing a statistical framework for optimal parameter estimation.

Contribution

It proposes a novel cost function based on $oldsymbol{ extalpha}$-divergence that maximizes the geometric mean of precision and recall, with a new statistical method to estimate the optimal $oldsymbol{ extalpha}$ for data visualization.

Findings

The $oldsymbol{ extalpha}$-divergence-based cost function effectively balances precision and recall.

The EDA distribution enables rigorous estimation of the optimal $oldsymbol{ extalpha}$.

Experimental results confirm the method's ability to identify the best trade-off for various datasets.

Abstract

Data visualization is one of the major applications of nonlinear dimensionality reduction. From the information retrieval perspective, the quality of a visualization can be evaluated by considering the extent that the neighborhood relation of each data point is maintained while the number of unrelated points that are retrieved is minimized. This property can be quantified as a trade-off between the mean precision and mean recall of the visualization. While there have been some approaches to formulate the visualization objective directly as a weighted sum of the precision and recall, there is no systematic way to determine the optimal trade-off between these two nor a clear interpretation of the optimal value. In this paper, we investigate the properties of $\alpha$-divergence for information visualization, focusing our attention on a particular range of $\alpha$ values. We show that the minimization of the new cost function corresponds to maximizing a geometric mean between precision and recall, parameterized by $\alpha$. Contrary to some earlier methods, no hand-tuning is needed, but we can rigorously estimate the optimal value of $\alpha$ for a given input data. For this, we provide a statistical framework using a novel distribution called Exponential Divergence with Augmentation (EDA). By the extensive set of experiments, we show that the optimal value of $\alpha$, obtained by EDA corresponds to the optimal trade-off between the precision and recall for a given data distribution.