Model space visualization for multivariate linear trend discovery

Abstract

Discovering and extracting linear trends and correlations in datasets is very important for analysts to understand multivariate phenomena. However, current widely used multivariate visualization techniques, such as parallel coordinates and scatterplot matrices, fail to reveal and illustrate such linear relationships intuitively, especially when more than 3 variables are involved or multiple trends coexist in the dataset. We present a novel multivariate model parameter space visualization system that helps analysts discover single and multiple linear patterns and extract subsets of data that fit a model well. Using this system, analysts are able to explore and navigate in model parameter space, interactively select and tune patterns, and refine the model for accuracy using computational techniques. We build connections between model space and data space visually, allowing analysts to employ their domain knowledge during exploration to better interpret the patterns they discover and their validity. Case studies with real datasets are used to investigate the effectiveness of the visualizations.

On Computationally-Enhanced Visual Analysis of Heterogeneous Data and Its Application in Biomedical Informatics

“…Interactive systems have also been used to help create decision trees [98] (see Figure 4). Guo et al [70] enable the interactive exploration of multivariate model parameters. They visualize the model space together with the data to reveal the trends in the data.…”

“…[51], [52], [53], [54], [55], [56] Groups & Classification [57] [58], [59] [60], [61], [62], [63], [64], [65], [66], [67], [68], [69], [70], [71], [72], [73], [74] [75], [76], [77], [78], [79], [80] Dependence & Prediction [81], [82], [46] [83], [84], [85], [86], [87], [88], [89] [90], [91], [92] being analyzed. The results are then presented to the user through different visual encodings that are often accompanied by interaction.…”

Visualization System Requirements for Data Processing Pipeline Design and Optimization

“…Each colored cell indicates a requirement that is important for a given case study/paper, and an 'x' indicates that the requirement remains unmet by users' current tools. [25] req req req req req Chemical process models [26] x req req x Economic modelling [27] req req xx req req x req Aircraft engine design [28] req req x req req Phylogenetic trees [29,30] x req x x req req Molecular evolution [31,32] x x req x req x * Raidou et al [75] iCoCooN req req req * Ruppert et al [46] --x req x Luboshik et al [38] --x req req req Bruckner et al [10] --x req x req * Beham et al [43] Cupid req req xx Konyha et al [39] --x req x Pretorius et al [12] Paramorama req x req Afzal et al [36] RVF x req req x Bergner et al [13] ParaGlide xx req Torsney-Weir et al [33] Tuner x req x x * Padua et al [11] --x req x * Bögl et al [42] TiMoVA req req req x x req Spence et al [73] --req req Berger et al [9] --req x xx Piringer et al [40] HyperMoVal xx req req req Matkovic et al [17] --req x Coffey et al [34] --x x req Matkovic et al [50] --x Potter et al [49] Ensemble-Vis x req Booshehrian et al [19] Vismon req x req x Brecheisen et al [52] --x req x Unger et al [54] --req req req Amirkhanov et al [74] --x Marks et al [48] Design Galleries req x Waser et al [18] World Lines req x x * Martins et al [35] --x req x * Wu [53] --xx x Guo et al [41] --req…”

“…Sometimes the pipeline is wellestablished (e.g., [38], [39]), but in other situations users need to choose between algorithms (e.g., the 3D image segmentation case study in Sec. 4.2), improve the sophistication of an existing pipeline [17], [40], make a pipeline robust to the characteristics of the input data [12], or design the pipeline from scratch [27], [41], [42]. The latter is particularly true in exploratory analysis, where users are analyzing a new form of data or looking for new patterns (e.g., [9], [10], [11], [43]).…”

A visual analytics framework for spatio-temporal analysis and modelling

“…Thus, Guo, Z., et al (2009) suggest techniques that help an analyst to discover single and multiple linear trends in multivariate data. Hao et al (2011) describe an approach to building peak-preserving models of single time series.…”