scikit-fda: Interactive Visualization and Analysis Tools for Functional Data
Author
Sánchez Romero, ÁlvaroAdvisor
Suárez González, AlbertoEntity
UAM. Departamento de Ingeniería InformáticaDate
2021-05Subjects
Functional Data Analysis; Visualization; Visualization, interactivity; InformáticaEsta obra está bajo una licencia de Creative Commons Reconocimiento-NoComercial-SinObraDerivada 4.0 Internacional.
Abstract
In this work interactive tools have been developed for the visualization and analysis of functional
data in the scikit-fda library [1]. The scikit-fda library for the analysis of functional data is integrated in
SciPy [2], an open-source ecosystem for mathematics, science and engineering, developed in Python.
Functional Data Analysis (FDA) is the branch of statistics that studies random variables that depend on
a continuous parameter; or what is the same, random functions like curves or surfaces. FDA could for
example study the growth on the height of a person during his youth (in this case being the continuum
time). To be able to obtain graphical information from this data, new graphics had being developed to
obtain details of curves, atypical values, parameterization of functions and others.
To start, a comparative analysis of the visualization and data analysis libraries available will be
done. During this study attention will be paid to the interactive tools that are given to the user. With
the objective of establishing the context in which the project has been developed, a brief description of
the structure and functionality of scikit-fda will be given. After that, there will be a discussion about the
main visualization tools developed: a shape outlier detector, a method to parameterize functions and a
graphic that allows to compare a sample of data with two distributions. Besides that, I have also worked
in a module that allows you to interact with all the different graphics thanks to widgets or with your own
mouse. For this it has been necessary to investigate the best solutions and the most effcient ones that
allowed the plots to work in every graphical user interface (GUI).
Furthermore, this document will talk about the software steps that had to be followed to manage to
design and implement the project. It will comment on the documentation, types of tests used and work
mechanisms followed. GitHub is the tool used to share our progress and where the code is currently
available for everybody [3].
My objectives with this degree work are creating a set of visualization methods that help the users,
besides creating an easier, more standard and homogeneous interface for the already existing methods
and the new ones.
Files in this item
Google Scholar:Sánchez Romero, Álvaro
This item appears in the following Collection(s)
Except where otherwise noted, this item's license is described as https://creativecommons.org/licenses/by-nc-nd/4.0/
Related items
Showing items related by title, author, creator and subject.
-
Functional data analysis: interpolation, registration, and nearest neighbors in scikit-fda
Marcos Manchón, Pablo
2019-06 -
scikit-fda: Principal Component Analysis for Functional Data
Hong, Yujian
2020-05