A tool developed in Matlab for multiple correspondence analysis of fuzzy coded data sets: Application to morphometric skull data

作者：Pinti Antonio^*; Rambaud Fabienne; Griffon Jean Louis; Ahmed Abdelmalik Taleb

来源：Computer Methods and Programs in Biomedicine, 2010, 98(1): 66-75.

DOI：10.1016/j.cmpb.2009.09.009

• Multiple correspondence analysis
• Principal component analysis
• Fuzzy sets
• Biplot
• Skulls dataset

Multiple Correspondence factorial Analysis is a multivariate method for the exploratory study of multidimensional contingency tables. Its use can be extended to the analysis of a table of fuzzy coded data resulting from a distribution into fuzzy windows defined by linguistic properties. There are few existing software tools that allow performing this type of analysis on a data table; furthermore these tools are not interactive and do not allow defining and representing fuzzy windowing. This paper presents a software tool, developed with Mat lab, to compute and represent results from multiple correspondence factorial analyses. Pre-defined membership functions can be selected by the user according to the distribution histograms of the data.
This paper presents an application example of this program onto a data table of morphometric parameters of 150 male skulls throughout 5 periods of Egyptian civilization. The results are compared to those of a principal component analysis, which is more often used for the study of experimental data. Our program allows a rapid description of the morphological evolution of skulls over time, notably thanks to a linguistic description of each variable, whereas the results of the latter method are less obvious to observe and require a deeper analysis in order to arrive at the same conclusions.

• 出版日期2010-4