Statfda: A User-Friendly Web Application to Analyse Curves Data from a Functional Perspective

Abstract

Using classical statistical techniques to analyze curves data whose behavior is clearly continuous might cause inexact results. To avoid this tricky situation, a suitable option is to carry out a functional data analysis that takes the time-dependent structure into account. However, the statistical modeling can be complex and difficult to understand and therefore, there is a great need of providing new tools to apply this methodology in any area of knowledge. In this sense, an open access web application is introduced in the current work. It serves as a comprehensible and effective tool whose greatest competitive advantage is that programming skills or conceptual knowledge in math are not required to use it. This interactive application has a user-friendly format and offers different well-defined functional techniques such as exploratory analysis, functional principal component analysis and functional regression models. Its functionalities are illustrated making use of a public dataset of biomechanics.