China
Africa
Explore chapters and articles related to this topic
Distribution Shapes
Published in Wendy L. Martinez, Angel R. Martinez, Jeffrey L. Solka, Exploratory Data Analysis with MATLAB®, 2017
Wendy L. Martinez, Angel R. Martinez, Jeffrey L. Solka
The bagplot is a bivariate box-and-whiskers plot developed by Rousseeuw, Ruts, and Tukey [1999]. This is a generalization of the univariate boxplot discussed previously. It uses the idea of the location depth of an observation with respect to a bivariate data set. This extends the idea of ranking or ordering univariate data to the bivariate case, so we can find quantities analogous to the univariate quartiles. The bagplot consists of the following components:A bag that contains the inner 50% of the data, similar to the IQR;A cross (or other symbol) indicating the depth median (described shortly);A fence to determine potential outliers; andA loop that indicates the points that are between the bag and the fence.
Using BART to Perform Pareto Optimization and Quantify its Uncertainties
Published in Technometrics, 2022
Akira Horiguchi, Thomas J. Santner, Ying Sun, Matthew T. Pratola
Figure 6 displays bagplots of the 100 values of and for each simulation scenario (PS plots and similar scenario results can be found in the Supplement). A bagplot extends the common boxplot for two-dimensional outputs and contains three main features analogous to the common univariate median, the box, and the whiskers on a conventional boxplot (Rousseeuw, Ruts, and Tukey 1999). For visual clarity, we include only two of these features: the depth median, which is the point with the highest possible halfspace depth, and the “bag,” which is a polygon that encloses 50% of the points around the depth median.
Batch process control and monitoring: a Dual STATIS and Parallel Coordinates (DS-PC) approach
Published in Production & Manufacturing Research, 2018
Miriam Ramos-Barberán, Miriam Vanessa Hinojosa-Ramos, José Ascencio-Moreno, Francisco Vera, Omar Ruiz-Barzola, María Purificación Galindo-Villardón
The charting procedure begins with the construction of control regions, which are based on bagplots (Rousseeuw, Ruts, & Tukey, 1999). The bagplot represents a bivariate generalization of the univariate boxplot in which ‘depth median’ plays the role of the robust centroid and it is surrounded by a convex polygon that defines the control region for the referred charts (Figure 3).