Complementary View on Multivariate Data Structure Based on Kohonen’s SOM, Parallel Coordinates and t-SNE Methods

2017 ◽  
pp. 255-265
Author(s):  
Anna M. Bartkowiak ◽  
Radosław Zimroz
Keyword(s):  
Data Structure ◽  
Multivariate Data ◽  
Parallel Coordinates

scholarly journals Canonical correlation analysis in high dimensions with structured regularization

2021 ◽  
pp. 1471082X2110410
Author(s):  
Elena Tuzhilina ◽  
Leonardo Tozzi ◽  
Trevor Hastie
Keyword(s):  
Data Structure ◽  
Correlation Analysis ◽  
Canonical Correlation Analysis ◽  
Canonical Correlation ◽  
Multivariate Data ◽  
High Dimensional Data ◽  
High Dimensional ◽  
High Dimensions ◽  
Structured Regularization ◽  
Data Matrices

Canonical correlation analysis (CCA) is a technique for measuring the association between two multivariate data matrices. A regularized modification of canonical correlation analysis (RCCA) which imposes an [Formula: see text] penalty on the CCA coefficients is widely used in applications with high-dimensional data. One limitation of such regularization is that it ignores any data structure, treating all the features equally, which can be ill-suited for some applications. In this article we introduce several approaches to regularizing CCA that take the underlying data structure into account. In particular, the proposed group regularized canonical correlation analysis (GRCCA) is useful when the variables are correlated in groups. We illustrate some computational strategies to avoid excessive computations with regularized CCA in high dimensions. We demonstrate the application of these methods in our motivating application from neuroscience, as well as in a small simulation example.

scholarly journals Revealing Structure in Visualizations of Dense 2D and 3D Parallel Coordinates

2006 ◽  
Vol 5 (2) ◽  
pp. 125-136 ◽  
Author(s):  
Jimmy Johansson ◽  
Patric Ljung ◽  
Mikael Jern ◽  
Matthew Cooper
Keyword(s):  
High Precision ◽  
Transfer Functions ◽  
Multivariate Data ◽  
Data Sets ◽  
Parallel Coordinates ◽  
Data Set ◽  
Number Of Clusters ◽  
Analysis Process ◽  
2D And 3D

Parallel coordinates is a well-known technique used for visualization of multivariate data. When the size of the data sets increases the parallel coordinates display results in an image far too cluttered to perceive any structure. We tackle this problem by constructing high-precision textures to represent the data. By using transfer functions that operate on the high-precision textures, it is possible to highlight different aspects of the entire data set or clusters of the data. Our methods are implemented in both standard 2D parallel coordinates and 3D multi-relational parallel coordinates. Furthermore, when visualizing a larger number of clusters, a technique called ‘feature animation’ may be used as guidance by presenting various cluster statistics. A case study is also performed to illustrate the analysis process when analysing large multivariate data sets using our proposed techniques.

On the usability of three-dimensional display in parallel coordinates: Evaluating the efficiency of identifying two-dimensional relationships

2013 ◽  
Vol 13 (1) ◽  
pp. 29-41 ◽  
Author(s):  
Jimmy Johansson ◽  
Camilla Forsell ◽  
Matthew Cooper
Keyword(s):  
Response Time ◽  
Information Visualization ◽  
Multivariate Data ◽  
Three Dimensional ◽  
Two Dimensional ◽  
Parallel Coordinates ◽  
Formal Evaluation ◽  
Three Dimensional Display

In recent years, several different attempts have been made to extend the well-known technique of parallel coordinates using a three-dimensional display. This article presents an evaluation that investigates the performance of the three-dimensional parallel coordinates technique and compares it with standard, two-dimensional parallel coordinates for analysis of two-dimensional relationships. Three-dimensional parallel coordinates, based on parallel planes instead of parallel axes, have been used for many years within the information visualization community for a variety of applications. Despite its quite common use, no formal evaluation detailing its usefulness for different tasks has been conducted. The evaluation presented in this article is the first step towards determining the usefulness of this type of three-dimensional parallel coordinates. The study compared three-dimensional parallel coordinates, using two different axis configurations commonly seen in the literature, with standard two-dimensional parallel coordinates for identification of two-dimensional relationships between variables in multivariate data. This type of task and the relationships to be judged are known to be well supported by two-dimensional parallel coordinates and multi-relational three-dimensional parallel coordinates. The results show that for identification of two-dimensional relationships, two-dimensional parallel coordinates are superior to the three-dimensional extensions, in terms of both response time and accuracy. Subjective opinions were also in favour of two-dimensional parallel coordinates. This study adds to the much-needed body of work examining the usability of three-dimensional representations in information visualization and for what tasks and data a proposed method is or is not appropriate.

scholarly journals Augmenting the usability of parallel coordinate plot: The polyline glyphs

2017 ◽  
Vol 17 (2) ◽  
pp. 108-127 ◽  
Author(s):  
Tomasz Opach ◽  
Jan Ketil Rød
Keyword(s):  
Empirical Study ◽  
Information Acquisition ◽  
Multivariate Data ◽  
The Body ◽  
Visualization Technique ◽  
Parallel Coordinates ◽  
Visual Clutter ◽  
User Tasks ◽  
Insight Into

Polyline glyphs are minimized thumbnails of polylines from parallel coordinates. Since such glyphs may augment the usability of parallel coordinates, the authors investigate whether there are benefits to be derived from using polyline glyphs that are dynamically linked to parallel coordinates as opposed to the use of the latter visualization technique alone. They also identify user tasks that can be effectively solved if parallel coordinates dynamically linked to polyline glyphs are used. This study adds to the body of previous work a discussion on the features of the polyline glyphs that facilitate the exploration and understanding of multivariate data. Moreover, the authors conduct an empirical study in which parallel coordinates dynamically linked to polyline glyphs are used to solve four tasks. The main finding is that polyline glyphs can facilitate a better insight into the similarities between the multivariate signatures of data items and information acquisition if visual clutter hinders the use of parallel coordinates. The study also reveals that if visual clutter does not occur in parallel coordinates and the polylines from the latter can be differentiated, individuals tend not to use polyline glyphs to study multivariate signatures.

Sign in / Sign up

Export Citation Format

Share Document