This paper proposes an hybrid Artificial Neural Network (ANN) with Self-Organizing Map (SOM) and modified Adaptive Coordinates (AC) for multivariate dimension reduction and data structures exploration. SOM, being a pr...
详细信息
ISBN:
(纸本)9781424453306
This paper proposes an hybrid Artificial Neural Network (ANN) with Self-Organizing Map (SOM) and modified Adaptive Coordinates (AC) for multivariate dimension reduction and data structures exploration. SOM, being a prominent unsupervised learning algorithm, is often used for multivariate data visualization. However, SOM only preserved input space inter-neurons distances and not in the output space because of SOM rigid grid. SOM grid provides little information for visual exploration of the clustering tendency of the multivariatedata. Modified AC is therefore proposed to remove SOM's map rigidity and provides better data topology preserved visualization. Empirical study of the hybrid yielded promising topology preserved visualizations for synthetic and benchmarking datasets.
We argue that runtime program transformation, partial evaluation, and dynamic compilation are essential tools for automated generation of flexible, highly interactive graphical interfaces. in particular, these techniq...
详细信息
ISBN:
(纸本)9780897919173
We argue that runtime program transformation, partial evaluation, and dynamic compilation are essential tools for automated generation of flexible, highly interactive graphical interfaces. in particular, these techniques help bridge the gap between a high-level, functional description and an efficient implementation. To support our claim, we describe our application of these techniques to a functional implementation of n-Vision, a real-time visualization system that represents multivariate relations as nested 3D interactors, and to Auto Visual, a rule-based system that designs n-Vision visualizations from high-level task specifications, n-Vision visualizations are specified using a simple functional language. These programs are transformed into a cached dataflow graph. A partial evaluator is used on particular computation-intensive function applications, and the results are compiled to native code. The functional representation simplifies generation of correct code, and the program transformations ensure good performance. We demonstrate why these transformations improve performance and why they cannot be done at compile time.
The scatterplot matrix is one of the most common methods used to project multivariatedata onto two dimensions for display. While each off-diagonal plot maps a pair of non-identical dimensions. there is no prescribed ...
详细信息
ISBN:
(纸本)0819461008
The scatterplot matrix is one of the most common methods used to project multivariatedata onto two dimensions for display. While each off-diagonal plot maps a pair of non-identical dimensions. there is no prescribed mapping for the diagonal plots. In this paper, histograms, 1D plots and 2D plots are drawn in the diagonal plots of the scatterplots matrix. In 1D plots, the data are assumed to have. order, and they are projected in this order. In 2D plots, the data are assumed to have spatial information, and they are projected onto locations based oil these spatial attributes using color to represent the dimension value. The plots and the scatterplots are linked together by brushing. Brushing oil these alternate visualizations will affect the selected data in the regular scatterplots, and vice versa. Users can also navigate to other visualizations, such as parallel coordinates and glyphs, which are also linked with the scatterplot matrix by brushing. Ordering and spatial attributes can also be used as methods of indexing and organizing data. Users call select all ordering span or a spatial region by interacting with 1D plots or with 2D plots, and then observe the characteristics of the selected data subset. 1D plots and 2D plots provide the ability to explore the ordering and spatial attributes, while other views are for viewing the abstract data. In a sense. we are linking what are traditionally seen as scientific visualization methods with methods from the. information visualization and statistical graphics fields. We validate the usefulness of this integration by providing two case studies, time series data analysis and spatial data analyses.
This paper presents unsupervised algorithms for discovering previously unknown subspace trends in high-dimensional data sets without the benefit of prior information. A subspace trend is a sustained pattern of gradual...
详细信息
This paper presents unsupervised algorithms for discovering previously unknown subspace trends in high-dimensional data sets without the benefit of prior information. A subspace trend is a sustained pattern of gradual/progressive changes within an unknown subset of feature dimensions. A fundamental challenge to subspace trend discovery is the presence of irrelevant data dimensions, noise, outliers, and confusion from multiple subspace trends driven by independent factors that are mixed in with each other. These factors can obscure the trends in conventional dimension reduction & projection based datavisualizations. To overcome these limitations, we propose a novel graph-theoretic neighborhood similarity measure for detecting concordant progressive changes across data dimensions. Using this measure, we present an unsupervised algorithm for trend-relevant feature selection, subspace trend discovery, quantification of trend strength, and validation. Our method successfully identified verifiable subspace trends in diverse synthetic and real-world biomedical datasets. visualizations derived from the selected trend-relevant features revealed biologically meaningful hidden subspace trend(s) that were obscured by irrelevant features and noise. Although our examples are drawn from the biological domain, the proposed algorithm is broadly applicable to exploratory analysis of high-dimensional data including visualization, hypothesis generation, knowledge discovery, and prediction in diverse other applications.
Glyphs are graphical entities that convey one or more data values via attributes such as shape, size, color, and position. They have been widely used in the visualization of data and information, and are especially we...
详细信息
Glyphs are graphical entities that convey one or more data values via attributes such as shape, size, color, and position. They have been widely used in the visualization of data and information, and are especially well suited for displaying complex, multivariatedata sets. The placement or layout of glyphs on a display can communicate significant information regarding the data values themselves as well as relationships between data points, and a wide assortment of placement strategies have been developed to date. Methods range from simply using data dimensions as positional attributes to basing placement on implicit or explicit structure within the data set. This paper presents an overview of multivariate glyphs, a list of issues regarding the layout of glyphs, and a comprehensive taxonomy of placement strategies to assist the visualization designer in selecting the technique most suitable to his or her data and task. Examples, strengths, weaknesses, and design considerations are given for each category of technique. We conclude with some general guidelines for selecting a placement strategy, along with a brief description of some of our future research directions.
Numerical weather simulation data usually comprises various meteorological variables, such as precipitation, temperature and pressure. In practical applications, data generated with several different numerical simulat...
详细信息
Numerical weather simulation data usually comprises various meteorological variables, such as precipitation, temperature and pressure. In practical applications, data generated with several different numerical simulation models are usually used together by forecasters to generate the final forecast. However, it is difficult for forecasters to obtain a clear view of all the data due to its complexity. This has been a great limitation for domain experts to take advantage of all the data in their routine work. In order to help explore the multi-variate and multi-model data, we propose a stamp based exploration framework to assist domain experts in analyzing the data. The framework is used to assist domain experts in detecting the bias patterns between numerical simulation data and observation data. The exploration pipeline originates from a single meteorological variable and extends to multiple variables under the guidance of a designed stamp board. Regional data patterns can be detected by analyzing distinctive stamps on the board or generating extending stamps using the Boolean set operations. Experiment results show that some meteorological phenomena and regional data patterns can be easily detected through the exploration. These can help domain experts conduct the data analysis efficiently and further guide forecasters in producing reliable weather forecast.
暂无评论