Data visualization is the process by which data of any size or dimensionality is processed to produce an understandable set of data in a lower dimensionality, allowing it to be manipulated and understood more easily by people. The goal of our paper is to survey the performance of current high-dimensional data visualization techniques and quantify their strengths and weaknesses through relevant quantitative measures, including runtime, memory usage, clustering quality, separation quality, global structure preservation, and local structure preservation. To perform the analysis, we select a subset of state-of-the-art methods. Our work shows how the selected algorithms produce embeddings with unique qualities that lend themselves towards certain tasks, and how each of these algorithms are constrained by compute resources.
Download