A Nested Hierarchy of Localized Scatterplots

Eisemann, Martin; Albuquerque, Georgia; Magnor, Marcus

doi:10.1109/sibgrapi.2014.14

Cited by 9 publications

(5 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Visualizing the surrogate models to approximate the behaviors of the original models, either globally or locally, is another branch of related works. [59][60][61][62][63][64][65][66] Rule-based visualizations have also been deployed for the interpretation of complex neural networks. [67][68][69][70] Nevertheless, these models differ due to the lack of inherent decisions that could be extracted directly from the bagged and boosted decision trees.…”

Section: Tree-and Rule-based Model Visualizationmentioning

confidence: 99%

VisRuler: Visual analytics for extracting decision rules from bagged and boosted decision trees

Chatzimparmpas

Martins

Kerren

2023

Information Visualization

View full text Add to dashboard Cite

Bagging and boosting are two popular ensemble methods in machine learning (ML) that produce many individual decision trees. Due to the inherent ensemble characteristic of these methods, they typically outperform single decision trees or other ML models in predictive performance. However, numerous decision paths are generated for each decision tree, increasing the overall complexity of the model and hindering its use in domains that require trustworthy and explainable decisions, such as finance, social care, and health care. Thus, the interpretability of bagging and boosting algorithms—such as random forest and adaptive boosting—reduces as the number of decisions rises. In this paper, we propose a visual analytics tool that aims to assist users in extracting decisions from such ML models via a thorough visual inspection workflow that includes selecting a set of robust and diverse models (originating from different ensemble learning algorithms), choosing important features according to their global contribution, and deciding which decisions are essential for global explanation (or locally, for specific cases). The outcome is a final decision based on the class agreement of several models and the explored manual decisions exported by users. We evaluated the applicability and effectiveness of VisRuler via a use case, a usage scenario, and a user study. The evaluation revealed that most users managed to successfully use our system to explore decision rules visually, performing the proposed tasks and answering the given questions in a satisfying way.

show abstract

Section: Tree-and Rule-based Model Visualizationmentioning

confidence: 99%

VisRuler: Visual analytics for extracting decision rules from bagged and boosted decision trees

Chatzimparmpas

Martins

Kerren

2023

Information Visualization

View full text Add to dashboard Cite

show abstract

“…Users can apply an interactive regression analysis on a local portion of the data and immediately see the best fitting model on the plot. Eisemann et al [EAM14] describe interactive visualisation of distinct patterns of data within a given scatterplot (a hierarchy of localised scatterplots), which allows the user to explore dense areas in a scatterplot. In this paper, a combination of these approaches is used to enhance the visualisation of local patterns and facilitate the exploration of the dataset.…”

Section: Visualisation Of Local Patternsmentioning

confidence: 99%

Interactive Visual Exploration of Local Patterns in Large Scatterplot Spaces

Chegini

Shao

Gregor

et al. 2018

Computer Graphics Forum

View full text Add to dashboard Cite

Analysts often use visualisation techniques like a scatterplot matrix (SPLOM) to explore multivariate datasets. The scatterplots of a SPLOM can help to identify and compare two-dimensional global patterns. However, local patterns which might only exist within subsets of records are typically much harder to identify and may go unnoticed among larger sets of plots in a SPLOM. This paper explores the notion of local patterns and presents a novel approach to visually select, search for, and compare local patterns in a multivariate dataset. Model-based and shape-based pattern descriptors are used to automatically compare local regions in scatterplots to assist in the discovery of similar local patterns. Mechanisms are provided to assess the level of similarity between local patterns and to rank similar patterns effectively. Moreover, a relevance feedback module is used to suggest potentially relevant local patterns to the user. The approach has been implemented in an interactive tool and demonstrated with two real-world datasets and use cases. It supports the discovery of potentially useful information such as clusters, functional dependencies between variables, and statistical relationships in subsets of data records and dimensions.

show abstract

“…Turkay et al propose to visualize and brush statistical aggregates on selected dimensions as well [TFH11]. Eisemann et al present a visualization of hierarchy of scatter plots [EAM14]. Inside one panel, a subset of data points is shown for other axes as nested plot.…”

Section: Multi-dimensional Visualizationmentioning

confidence: 99%

Enhancing Scatterplots with Multi‐Dimensional Focal Blur

Staib

Grottel

Gumhold

2016

Computer Graphics Forum

View full text Add to dashboard Cite

Pressure Temperature Wind speed a) b) c) d) Figure 1: Hurricane Isabel data set shown in different plots. Left: scatterplot of x×y using our blur method. Top row: zoomed-in views of the two marked regions. a) and c) use our method. b) and d) use transparent splats for comparison. Bottom row: histograms of three relevant dimensions and the selected focus coordinates marked with red vertical lines. (cf. Sec. 5.2) AbstractScatterplots directly depict two dimensions of multi-dimensional data points, discarding all other information. To visualize all data, these plots are extended to scatterplot matrices, which distribute the information of each data point over many plots. Problems arising from the resulting visual complexity are nowadays alleviated by concepts like filtering and focus and context. We present a method based on depth of field that contains both aspects and injects information from all dimensions into each scatterplot. Our approach is a natural generalization of the commonly known focus effects from optics. It is based on a multidimensional focus selection body. Points outside of this body are defocused depending on their distance. Our method allows for a continuous transition from data points in focus, over regions of blurry points providing contextual information, to visually filtered data. Our algorithm supports different focus selection bodies, blur kernels, and point shapes. We present an optimized GPU-based implementation for interactive exploration and show the usefulness of our approach on several data sets.

show abstract

A Nested Hierarchy of Localized Scatterplots

Cited by 9 publications

References 25 publications

VisRuler: Visual analytics for extracting decision rules from bagged and boosted decision trees

VisRuler: Visual analytics for extracting decision rules from bagged and boosted decision trees

Interactive Visual Exploration of Local Patterns in Large Scatterplot Spaces

Enhancing Scatterplots with Multi‐Dimensional Focal Blur

Contact Info

Product

Resources

About