Origin and development of S.I.A.
Let us recall that this method of non-symmetrical
data analysis allows us to extract and structure knowledge in the form
of rules and general rules using a set of data that connects a population
of subjects (or objects) with a set of variables. This method stems from
the statistical modelling of the quasi-implication: if a given variable
or conjunction of variables "a" is noticed among the population,
then, in general, so will be variable "b". The variables referred
to can be of various types: binary, modal, numerical, interval, etc. Contrary
to the methods of symmetrical analysis based, for instance, on distance
or correlation, groups of acquired rules can lead to hypotheses of causality.
These groups are structured according to different supplementary approaches
(implicative graph, oriented hierarchy). The quantitative determination
of subjects or descriptors responsible for these structures is determined
by their contribution or their typicality. Finally, the results and their
interpretation can be readily visualized by means of the CHIC software
(Classification Hiérarchique Implicative and Cohésitive).
Orus P., Zemora L., Gregori P. (Eds) (2009) Teoria y Aplicaciones
del Analisis Estadistico Implicativo, Universitat Jaume-1, Castellon
(Espagne)
Régnier J.C., Gras R., Spagnolo F., Di Paola
B. (Eds) (2011) Analyse Statistique Implicative: Objet de recherche
et de formation en analyse des données, outil pour la recherche
multidisciplinaire. Prolongement des débats. QRDM Quaderni
di Ricerca in Didattica - GRIM ISSN on-line 1592-4424, Palerme:
Université de Palerme.
https://sites.unipa.it/grim/QRDM_20_Suppl_1.htm
|