Elsevier

Fuzzy Sets and Systems

Volume 107, Issue 3, 1 November 1999, Pages 255-275
Fuzzy Sets and Systems

Marriage of fuzzy sets and multiple correspondence analysis: Examples with subjective interval data and biomedical signals

https://doi.org/10.1016/S0165-0114(97)00317-5Get rights and content

Abstract

The multiple correspondence analysis (MCA) is a descriptive and multidimensional method which can investigate several empirical situations, each situation being described by a categorical variable set. This paper shows how the fuzzy sets principle can be used to transform raw continuous data into categorical. The transformation is considered in two main stages: data characterizing and data coding. The data characterizing is performed to build analysis variables from complex empirical variables, such as multidimensional signals, using fuzzy windowing of the time and/or the space axes. The analysis variables are indicators summarizing the information within the so obtained windows. The data coding is performed to build homogeneous analysis variables, i.e. variables that are based on a qualitative scale using fuzzy windowing. Both methodological and practical aspect are considered in this paper through two examples. The first example considers the comparative analysis of force and force derivative signals in several load lifting conditions. The second example considers the analysis and the modelling of individual agreements between a graphical view showing two bargraphs and the assertion the height of the first bar is large and the height of the second is large, the agreement being given through an interval.

References (0)

Cited by (26)

  • An innovative integrated modelling of safety data using multiple correspondence analysis and fuzzy discretization techniques

    2020, Safety Science
    Citation Excerpt :

    This transformation process also results in loss of valuable information because the degree of contribution to discretized values are not considered (Hu et al., 2007) which can be alleviated by using fuzzy coding (Zerrin and Greenacre, 2011). Representation of information in terms of linguistic terms in fuzzy coding approach will help in overcoming the drawbacks of traditional data mining based discretization approach (Loslever and Bouilland, 1999). Discretization is carried out in two important steps, data characterization, and data codification.

  • Variable compensation during the sit-to-stand task among individuals with severe knee osteoarthritis

    2017, Annals of Physical and Rehabilitation Medicine
    Citation Excerpt :

    Contrary to a classical binary approach, in which only one value can represent a modality for a variable (e.g., low, average, high: 1, 0, 0), the fuzzy approach permits a sliding scale of membership values (e.g., low, average, high: 0.7, 0.3, 0) as the probability of belonging to a modality. This feature improves the handling of data imprecision caused by changing quantitative data into qualitative data [23,36]. Thus, in this study, biomechanical variables were coded and normalized by using 3 triangular fuzzy membership functions: low, average and high [37].

  • Identification of gait patterns in individuals with cerebral palsy using multiple correspondence analysis

    2013, Research in Developmental Disabilities
    Citation Excerpt :

    The fuzzy window coding normalizes the data from different natures, it simplifies the knowledge extraction process and it increases data interpretability (Chau, 2001a). Contrary to a classical-binary approach in which only one value can represent a modality for a parameter (e.g., Low, Average, High – 1, 0, 0), the fuzzy approach permits different membership values (e.g., Low, Average, High – 0.7, 0.3, 0) as the probability of belonging to a modality and can deal with data imprecision (Bouilland & Loslever, 1998; Loslever & Bouilland, 1999). Thus, the biomechanical parameters were coded and normalized using three triangular fuzzy membership functions related to the following three modalities – Low, Average and High – as used by Armand et al. (2007) and Sagawa et al. (2012).

  • Associations between gait and clinical parameters in patients with severe knee osteoarthritis: A multiple correspondence analysis

    2013, Clinical Biomechanics
    Citation Excerpt :

    To interpret this relationship, multivariate statistical analyses have been proposed. These analyses have been employed to adapt the multidimensionality of clinical and biomechanical data to optimise the extraction of relevant information (for reviews see (Bouilland and Loslever, 1998; Chau, 2001a, 2001b; Loslever and Bouilland, 1999)). Given the complexity of knee OA, a multivariate statistical analysis may be helpful to increase knowledge of the gait in patients with knee OA.

  • Biplots of fuzzy coded data

    2011, Fuzzy Sets and Systems
View all citing articles on Scopus
View full text