User‐driven Feature Space Transformation

Abstract

Interactive visualization systems for exploring and manipulating high-dimensional feature spaces have experienced a substantial progress in the last few years. State-of-art methods rely on solid mathematical and computational foundations that enable sophisticated and flexible interactive tools. Current methods are even capable of modifying data attributes during interaction, highlighting regions of potential interest in the feature space, and building visualizations that bring out the relevance of attributes. However, those methodologies rely on complex and non-intuitive interfaces that hamper the free handling of the feature spaces. Moreover, visualizing how neighborhood structures are affected during the space manipulation is also an issue for existing methods. This paper presents a novel visualization-assisted methodology for interacting and transforming data attributes embedded in feature spaces. The proposed approach relies on a combination of multidimensional projections and local transformations to provide an interactive mechanism for modifying attributes. Besides enabling a simple and intuitive visual layout, our approach allows the user to easily observe the changes in neighborhood structures during interaction. The usefulness of our methodology is shown in an application geared to image retrieval.

Adaptable Radial Axes Plots for Improved Multivariate Data Visualization

“…Recently, several dimensionality reduction methods have appeared in the literature where users can interactively modify visualizations by updating the coordinates of points, or the distances or neighboring relationships between them [PEP*11, MFNP13, MWT14, WTH15]. While these techniques are useful in order to observe relationships between the elements of a data set, they do not show information related to its original attributes.…”

Visual Interaction with Dimensionality Reduction: A Structured Literature Analysis

“…Furthermore, the quality of the DR pipeline can potentially be visualized, either separately, or embedded in the low dimensional representation. Some DR types calculate or identify errors, and in combination with other machine learning methods, additional quality information might be obtained (e.g., the precision of a classifier [42]). Furthermore, different DR pipeline variants (e.g., pre-defined DR configurations or automatically built recommendations) can be visualized [27].…”

Facing the high-dimensions: Inverse projection with radial basis functions

“…Recent MP techniques emphasize the importance of fast computation [1,2] and user-interactivity through the use of control points to steer the projection [1,3,14]. Control points have also been recently proposed, for instance, as a means to transform a multidimensional feature space based on user input provided on the projection layout [15].…”