Synthetic Generation of High-Dimensional Datasets

Abstract

Abstract-Generation of synthetic datasets is a common practice in many research areas. Such data is often generated to meet specific needs or certain conditions that may not be easily found in the original, real data. The nature of the data varies according to the application area and includes text, graphs, social or weather data, among many others. The common process to create such synthetic datasets is to implement small scripts or programs, restricted to small problems or to a specific application. In this paper we propose a framework designed to generate high dimensional datasets. Users can interactively create and navigate through multi dimensional datasets using a suitable graphical user-interface. The data creation is driven by statistical distributions based on a few user-defined parameters. First, a grounding dataset is created according to given inputs, and then structures and trends are included in selected dimensions and orthogonal projection planes. Furthermore, our framework supports the creation of complex non-orthogonal trends and classified datasets. It can successfully be used to create synthetic datasets simulating important trends as multidimensional clusters, correlations and outliers.

Evaluation in the Crowd. Crowdsourcing and Human-Centered Experiments

“…In many cases, study designers can use automatic or visual-interactive data generation tools. For example, the PCDC System [23], SketchPadN-D [100] or the system developed by Albuquerque et al [4] allow for visualinteractive creation of multivariate data with specified properties. Random data generators, such as graph generators, can automatically create data with special properties (e.g., [25,106,7,6]).…”

Generative Data Models for Validation and Evaluation of Visualization Techniques

“…While such semi-automatic approaches are typically time-consuming, the user knows exactly which and how patterns are distributed in the data. The approach by Albuquerque et al [3] focuses on visual properties by representing user-defined structures as probability density functions. Afterward, the specified number of points is sampled according to these structures.…”

Interactive Random Graph Generation with Evolutionary Algorithms

“…Except for drawing cliques, matrices make it hard to predict how a node-link representation of the final graph would look like, with its measures also being almost impossible to control. Albuquerque et al [2] propose generators and sketching for 1D, 2D and 3D scatterplots to generate multivariate data sets. In general, such sketching interfaces do not create diverse data sets, but provide only one particular solution at a time.…”