Related Experiment Videos
[Principal components analysis]
J A López del Val1, J P Alonso Pérez de Agreda
1Unidad Docente de Medicinia Familiar y Comunitaria de Zaragoza.
Objective:
This study wants to identify factors or components latent besides health indicators from Spanish regions, and its graphic output.
Design:
Observational study.
Setting:
Twenty eight indicators from regions were used: mortality, morbidity, communicable and no-communicable diseases, diet, dwelling and sanitary resources. Measurement was made between 1980-1988.
Interventions:
Principal component analysis has been applied to the indicators, reducing data dimension.
Measurement And Main Results:
Eight factors have been extracted, which explain 90% of the original information. This analysis, as can be seen from communnalities, represents correctly the set of original variables. The factors with more easily interpretation were: shortage sanitary resources, develop diet, mortality, chronic diseases and accidental.
Conclusions:
Only reduction data dimension could be justify the use of principal component analysis. Behind the agrupation of variables is mostly the socioeconomic background.