An Incremental Dimensionality Reduction Method for Visualizing Streaming Multidimensional Data

Fujiwara, Takanori; Chou, Jia-Kai; Shilpika,; Xu, Panpan; Liu, Ren; Ma, Kwan‐Liu

doi:10.1109/tvcg.2019.2934433

Cited by 64 publications

(39 citation statements)

References 67 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Axis reordering may be more challenging with an increasing number of covariates 47 . Previous researches 48,49 showed dimension reduction as an alternative technique for multidimensional data visualization.…”

Section: Discussionmentioning

confidence: 99%

Exploratory Data Analysis of Adverse Birth Outcomes and Exposure to Oxides of Nitrogen Using Interactive Parallel Coordinates Plot Technique

Mitku¹,

Zewotir²,

North³

et al. 2020

Sci Rep

View full text Add to dashboard Cite

We propose that a parallel coordinates plot can be used to study multidimensional data particularly to explore discovery of patterns across the variables. this can assist researchers from the health sciences to visualize their cohort data with interactive data analysis. the study used data from Mother and child in the Environment birth cohort in Durban, South Africa for the period 2013 to 2017 retrospectively registered. In this paper, we demonstrate that the exploration of multidimensional data with parallel coordinates plot and use of brushing using different colours assists with the identification of relationships and patterns. parallel coordinates plot visualization facilitates the researcher's skills to find trends, identify outliers and perform quality checks in large multivariate data. We have identified trends in the data that provide directions for further research, and illustrated thereby the potential of parallel coordinates plot to explore patterns and relationships of prenatal oxides of nitrogen exposure with multidimensional birth outcomes. the study recognized the co-occurrence of adverse birth outcomes among infants and these infants had mothers with moderate to high level of nox exposure during pregnancy. Brushing using different colours facilitated the detection of patterns of relationships to perform basic and advanced statistical model-based analysis. Data exploration techniques are important approaches for displaying multidimensional and multivariate datasets 1 , and to demonstrate the different ways that complex interactions among variables can be identified in images 2,3. It is a method for identifying and confirming previously unknown and important patterns within data sets and an important philosophical approach to understanding one's data 4. Visual analytics is well developed for up to three variables (three-dimensional plots (3D plots)). Visual analytics for more than four covariates is complex and not well developed 5. One of the most challenging stages in multivariate data analysis is to identify patterns and associations between a set of interrelated variables. Most typical data exploratory techniques fail to visualize multidimensional data 6. The typical exploratory data analysis techniques, such as bar charts, line graphs, histograms and scatter plots suit one or two-dimensional data 7. However, visualization of complex multidimensional data is challenging for researchers. Box plots, scatter plots, heat maps, complex radial tree layout diagrams are among the most commonly used techniques to better explore multidimensional data 8. The box-and-whisker plot is also useful in bivariate analysis, as it allows comparisons between different groups that results from categorical grouping variables 9. Previous data exploration techniques relied mostly on scatter plot matrices 10. However, it is not generalizable beyond three dimensions and making multivariate associations requires a large amount of screen space. Due to this, the problem of representing multidimensional data is a difficult an...

show abstract

Section: Discussionmentioning

confidence: 99%

Exploratory Data Analysis of Adverse Birth Outcomes and Exposure to Oxides of Nitrogen Using Interactive Parallel Coordinates Plot Technique

Mitku¹,

Zewotir²,

North³

et al. 2020

Sci Rep

View full text Add to dashboard Cite

show abstract

“…In most cases, the visualization tools cover at least the target group of domain experts/practitioners [EGG*12, FMH16, FCS*20, GNRM08, HNH*12, KPN16]. Then, other target groups such as ML experts [JC17b, KJR*18, SSK10, WLN*17] and developers are in the focus of the authors [KFC16, Mad19, RL15b, YZR*18] (commonly together).…”

Section: In‐depth Categorization Of Trust Against Facets Of Interamentioning

confidence: 99%

The State of the Art in Enhancing Trust in Machine Learning Models with the Use of Visualizations

Chatzimparmpas

Martins

Jusufi

et al. 2020

Computer Graphics Forum

143

View full text Add to dashboard Cite

Machine learning (ML) models are nowadays used in complex applications in various domains, such as medicine, bioinformatics, and other sciences. Due to their black box nature, however, it may sometimes be hard to understand and trust the results they provide. This has increased the demand for reliable visualization tools related to enhancing trust in ML models, which has become a prominent topic of research in the visualization community over the past decades. To provide an overview and present the frontiers of current research on the topic, we present a State‐of‐the‐Art Report (STAR) on enhancing trust in ML models with the use of interactive visualization. We define and describe the background of the topic, introduce a categorization for visualization techniques that aim to accomplish this goal, and discuss insights and opportunities for future research directions. Among our contributions is a categorization of trust against different facets of interactive ML, expanded and improved from previous research. Our results are investigated from different analytical perspectives: (a) providing a statistical overview, (b) summarizing key findings, (c) performing topic analyses, and (d) exploring the data sets used in the individual papers, all with the support of an interactive web‐based survey browser. We intend this survey to be beneficial for visualization researchers whose interests involve making ML models more trustworthy, as well as researchers and practitioners from other disciplines in their search for effective visualization techniques suitable for solving their tasks with confidence and conveying meaning to their data.

show abstract

“…Similar to the classical PCA, cPCA has the "sign ambiguity" problem [13,23,35]. Because of this problem, arbitrary sign flipping in each (c)PC occurs when performing EVD.…”

Section: Optimal Sign Flipping Of Cpcs and Fcsmentioning

confidence: 99%

Supporting Analysis of Dimensionality Reduction Results with Contrastive Learning

Fujiwara

Kwon

2020

IEEE Trans. Visual. Comput. Graphics

Self Cite

View full text Add to dashboard Cite

Dimensionality reduction (DR) is frequently used for analyzing and visualizing high-dimensional data as it provides a good first glance of the data. However, to interpret the DR result for gaining useful insights from the data, it would take additional analysis effort such as identifying clusters and understanding their characteristics. While there are many automatic methods (e.g., density-based clustering methods) to identify clusters, effective methods for understanding a cluster's characteristics are still lacking. A cluster can be mostly characterized by its distribution of feature values. Reviewing the original feature values is not a straightforward task when the number of features is large. To address this challenge, we present a visual analytics method that effectively highlights the essential features of a cluster in a DR result. To extract the essential features, we introduce an enhanced usage of contrastive principal component analysis (cPCA). Our method, called ccPCA (contrasting clusters in PCA), can calculate each feature's relative contribution to the contrast between one cluster and other clusters. With ccPCA, we have created an interactive system including a scalable visualization of clusters' feature contributions. We demonstrate the effectiveness of our method and system with case studies using several publicly available datasets.

show abstract

An Incremental Dimensionality Reduction Method for Visualizing Streaming Multidimensional Data

Cited by 64 publications

References 67 publications

Exploratory Data Analysis of Adverse Birth Outcomes and Exposure to Oxides of Nitrogen Using Interactive Parallel Coordinates Plot Technique

Exploratory Data Analysis of Adverse Birth Outcomes and Exposure to Oxides of Nitrogen Using Interactive Parallel Coordinates Plot Technique

The State of the Art in Enhancing Trust in Machine Learning Models with the Use of Visualizations

Supporting Analysis of Dimensionality Reduction Results with Contrastive Learning

Contact Info

Product

Resources

About