Evaluations-encompassing computational evaluations, benchmarks and user studies-are essential tools for validating the performance and applicability of graph and network layoutalgorithms (also known as graph drawing)...
详细信息
Evaluations-encompassing computational evaluations, benchmarks and user studies-are essential tools for validating the performance and applicability of graph and network layoutalgorithms (also known as graph drawing). These evaluations not only offer significant insights into an algorithm's performance and capabilities, but also assist the reader in determining if the algorithm is suitable for a specific purpose, such as handling graphs with a high volume of nodes or dense graphs. Unfortunately, there is no standard approach for evaluating layoutalgorithms. Prior work holds a 'Wild West' of diverse benchmark datasets and data characteristics, as well as varied evaluation metrics and ways to report results. It is often difficult to compare layoutalgorithms without first implementing them and then running your own evaluation. In this systematic review, we delve into the myriad of methodologies employed to conduct evaluations-the utilized techniques, reported outcomes and the pros and cons of choosing one approach over another. Our examination extends beyond computational evaluations, encompassing user-centric evaluations, thus presenting a comprehensive understanding of algorithm validation. This systematic review-and its accompanying website-guides readers through evaluation types, the types of results reported, and the available benchmark datasets and their data characteristics. Our objective is to provide a valuable resource for readers to understand and effectively apply various evaluation methods for graph layout algorithms. A free copy of this paper and all supplemental material is available at ***, and the categorized papers are accessible on our website at . We examine the diverse methodologies used in evaluating graph and network layoutalgorithms, covering computational evaluations, benchmarks, and user studies, detailing their techniques, outcomes, and trade-offs. The aim is to guide readers in understanding and applying evaluation methods effectively
Any graph drawing can be characterised by a range of computational aesthetic metrics. For example, a given drawing might be described as having eight crossings, a mean angular resolution of 0.34, and an edge orthogona...
详细信息
ISBN:
(纸本)9798350393811;9798350393804
Any graph drawing can be characterised by a range of computational aesthetic metrics. For example, a given drawing might be described as having eight crossings, a mean angular resolution of 0.34, and an edge orthogonality value of 0.72. However, without knowing the distribution of these metrics it is hard to compare the quality of drawings of different graphs, nor know whether a given drawing is typical or an outlier within the space of all possible drawings. This paper explores the range and distribution of ten normalised graph drawing layout metrics, based on graphs created by six graph generation algorithms and drawings created by six popular layoutalgorithms. We include the "Rome" and "North" graph repositories in our analysis. Our exploration of the multi-dimensional aesthetics space allows for comparisons between the graph drawing algorithms, highlighting those that cover larger or smaller volumes of the aesthetics space. We calculate the correlation coefficients between the metrics, indicating those that may conflict with each other (negatively correlated), and those that may be redundant (positively correlated). Our results will be useful as the basis for simulated annealing or gradient descent layoutalgorithms, for identifying the best layoutalgorithms for producing a specified combination and range of aesthetics, and for informing experimental controls in human empirical studies.
Visualizing relational data is crucial for understanding complex connections between entities in social networks, political affiliations, or biological interactions. Well-known representations like node-link diagrams ...
详细信息
Visualizing relational data is crucial for understanding complex connections between entities in social networks, political affiliations, or biological interactions. Well-known representations like node-link diagrams and adjacency matrices offer valuable insights, but their effectiveness relies on the ability to identify patterns in the underlying topological structure. Reordering strategies and layoutalgorithms play a vital role in the visualization process since the arrangement of nodes, edges, or cells influences the visibility of these patterns. The BioFabric visualization combines elements of node-link diagrams and adjacency matrices, leveraging the strengths of both, the visual clarity of node-link diagrams and the tabular organization of adjacency matrices. A unique characteristic of BioFabric is the possibility to reorder nodes and edges separately. This raises the question of which combination of layoutalgorithms best reveals certain patterns. In this paper, we discuss patterns and anti-patterns in BioFabric, such as staircases or escalators, relate them to already established patterns, and propose metrics to evaluate their quality. Based on these quality metrics, we compared combinations of well-established reordering techniques applied to BioFabric with a well-known benchmark data set. Our experiments indicate that the edge order has a stronger influence on revealing patterns than the node layout. The results show that the best combination for revealing staircases is a barycentric node layout, together with an edge order based on node indices and length. Our research contributes a first building block for many promising future research directions, which we also share and discuss. A free copy of this paper and all supplemental materials are available at https://***/9mt8r/?view_only=b70dfbe550e3404f83059afdc60184c6.
Existing graph layout algorithms are usually not able to optimize all the aesthetic properties desired in a graphlayout. To evaluate howwell the desired visual features are reflected in a graphlayout, many readabili...
详细信息
Existing graph layout algorithms are usually not able to optimize all the aesthetic properties desired in a graphlayout. To evaluate howwell the desired visual features are reflected in a graphlayout, many readabilitymetrics have been proposed in the past decades. However, the calculation of these readabilitymetrics often requires access to the node and edge coordinates and is usually computationally inefficient, especially for dense graphs. Importantly, when the node and edge coordinates are not accessible, it becomes impossible to evaluate the graphlayouts quantitatively. In this paper, wepresent a novel deep learning-based approach to evaluate the readability of graphlayouts by directly using graph images. Aconvolutional neural network architecture is proposed and trained on a benchmark dataset of graph images, which is composed of synthetically generated graphs and graphs created by sampling fromreal large networks. Multiple representative readabilitymetrics (including edge crossing, node spread, and group overlap) are considered in the proposed approach. Wequantitatively compare our approach to traditional methods and qualitatively evaluate our approach by showing usage scenarios and visualizing convolutional layers. This paper is a first step towards using deep learning based methods to quantitatively evaluate images fromthe visualization field.
Experimental design is important, but is often under-supported, in systems biology research. To improve experimental design, we extend the visualization of complex sphingolipid pathways to study biosynthetic origin in...
详细信息
ISBN:
(纸本)9781424418145
Experimental design is important, but is often under-supported, in systems biology research. To improve experimental design, we extend the visualization of complex sphingolipid pathways to study biosynthetic origin in SphinGOMAP. We use the ganglio-series sphingolipid dataset as a test bed and the Java Universal Network/graph Framework (JUNG) visualization toolkit. The result is an interactive visualization tool and data model for experimental design in lipid systems biology research. We improve the current SphinGOMAP in terms of interactive visualization by allowing (j) choice of four different network layouts, (ii) dynamic addition/deletion of on-screen molecules and (iii) mouse-over to reveal detailed molecule data. Future work will focus on integrating various lipid-relevant data systematically i.e. SphinGOMAP biosynthetic data, Lipid Bank molecular data (Japan) and Lipid MAPS metabolic pathway data (USA). We aim to build a comprehensive and interactive communication platform to improve experimental design for scientists globally in high-throughput lipid systems biology research.
tIn this article, the authors explore the use of graphlayout algorithims for visualizing proximity matrices such as those obtained in cultural domain analysis. Traditionally, multidimensional scaling has been used fo...
详细信息
tIn this article, the authors explore the use of graphlayout algorithims for visualizing proximity matrices such as those obtained in cultural domain analysis. Traditionally, multidimensional scaling has been used for this purpose. The authors compare the two approaches to identify, conditions when each approach is effective. As might be expected, they find that multidimensional scaling shines when the data are of low dimensionality and are compatible with the defining characteristics of Euclidean distances, such as symmetry and triangle inequality constraints. However, when one is working with data that do not fit these criteria, graphlayout algorithims do a better job of communicating the structure of the data. In addition, graph layout algorithms lend themselves to interactive use, which can yield a deeper and more accurate understanding of the data.
Information visualisation systems which generate diagrams representing discrete relational information must consider potential users if they are to he effective. Many algorithms which render an abstract graph structur...
详细信息
Information visualisation systems which generate diagrams representing discrete relational information must consider potential users if they are to he effective. Many algorithms which render an abstract graph structure as a diagram are valued for their conformance to aesthetic criteria (e.g, reducing the number of edge crossings, maximising symmetry), or for computational efficiency. They are not usually judged on their ability to produce diagrams that maximise human performance. This paper presents the results of experiments investigating the relative worth (from an HCI point of view) of graph drawing aesthetics and algorithms using a single graph. The results indicate that while some individual aesthetics affect human performance, it is difficult to say that one algorithm is 'better' than another from a relational understanding point of view. Designers of automatic layoutalgorithms, and the systems which embody such algorithms, can benefit from this study and this human-centred approach, by adapting their methods to focus on user concerns, rather than computational ones. (C) 2000 Elsevier Science B.V. All rights reserved.
暂无评论