Resumo

DE-LOS-COBOS-SILVA, Sergio G; GODDARD-CLOSE, John  e  GUTIERREZ-ANDRADE, Miguel A. Fuzzy regression vs. ordinary least squares regression: case study. Rev. Mat [online]. 2011, vol.18, n.1, pp. 33-48. ISSN 1409-1433.

The objective of this paper is to disseminate the technique of fuzzy regression and to  give a practical example of its use. To this end, classical regression is compared to  several fuzzy regression models on a problem concerning the consumer confidence  index with respect to the dollar rate, the latter taken as the independent variable. A  brief introduction is given to each of the different methodologies employed. The results  obtained using the regression algorithms, one with ordinary least squares and another  two with fuzzy regression, are presented. The instances generated using the official  historical data for the problem are given and the numerical results obtained with the  regression methods are reported.

Palavras-chave : linear regression; fuzzy regression; fuzzy linear programming.