A NEW FUZZY REGRESSION ALGORITHM

A new fuzzy regression algorithm is described and compared with conven tional ordinary and weighted least-squares and robust regression metho ds. The application of these different methods to relevant data sets p roves that the performance of the procedure described in this paper ex ceeds that of the ordinary least-squares method and equals and often e xceeds that of weighted or robust methods, including the two fuzzy met hods proposed previously (Otto, M.; Bandemer, H., Chemom. Intell. Lab. Syst, 1986, 1, 71. Hu, Y.; Smeyers-Verbeke, J.; Massart, D. L. Chemom , Intell, Lab. Syst, 1990, 8, 143), Moreover, we emphasize the effecti veness and the generality of the two new criteria proposed in this pap er for diagnosing the linearity of calibration lines in analytical che mistry.