Modeling of lattice parameters of the γ and γ′ phases of the superalloys with base nickel by using a multiple linear regression analysis

In the present work the lattice parameters of the γ and γ′ phases of the nickel base superalloys have been modeled by using a multiple linear regression (MLR) analysis. The most significant parameters were selected by means of a procedure of genetic algorithm, as result linear combinations were obtained from 2 to 13 variables. Both models were validated by computing the Q2 statistic (cross-validation coefficient). The values of Q2 were determined by two tests of cross-validation: (1) with a datum forming part of the test set and (2) with a third of the datum forming part of the test set, this procedure is repeated until each independent datum has been omitted at least once. The values of Q2 for the obtained optimal models are approximately 0.97 and 0.95 for the phase γ and γ′, respectively, indicating a good correlation.