Comparative study of predicting the molecular diffusion coefficient for polar and non-polar binary gas using neural networks and multiple linear regressions / Naima Melzi, Latifa Khaouane, Yamina Ammi, Salah Hanini, Maamar Laidi, Hamid Zentou.
Sažetak

In the current study, an artificial neural network (ANN) and multiple linear regressions (MLR) have been used to develop predictive models for the estimation of molecular diffusion coefficients of 1252 polar and non-polar binary gases at multiple pressures over a wide range of temperatures and substances. The quality and reliability of each method were estimated in terms of the correlation coefficient (R), mean squared errors (MSE), root mean squared error (RMSE), and in terms of external validation coefficients (Q2ext). The comparison between the artificial neural network (ANN) and the multiple linear regressions (MLR) revealed that the neural network models showed a good predicting ability with lower errors (the roots of the mean squared errors in the total database were 0.1400 for ANN1 and 0.1300 for ANN2), and (root mean squared errors in the total databases were 0.5172 for MLR1 and 0.5000 for MLR2).; U ovoj studiji primijenjene su umjetna neuronska mreža (ANN) i model višestruke linearne regresije (MLR) za razvoj prediktivnih modela za procjenu koeficijenata molekularne difuzije 1252 polarnih i nepolarnih binarnih plinova pri višestrukim tlakovima u širokom rasponu temperatura i tvari. Kvaliteta i pouzdanost svake metode procijenjeni su pomoću korelacijskog koeficijenta (R), srednjih kvadratnih pogrešaka (MSE), korijena srednje kvadratne pogreške (RMSE) te koeficijenata vanjske validacije (Q2ext). Usporedba između umjetne neuronske mreže (ANN) i višestrukih linearnih regresija (MLR) otkrila je da modeli neuronske mreže pokazuju dobru sposobnost predviđanja s nižim pogreškama (korijeni srednjih kvadratnih pogrešaka u ukupnoj bazi podataka bili su 0,1400 za ANN1 i 0 (1300 za ANN2 a pogreške korijena srednje vrijednosti u ukupnim bazama podataka bile su 0,5172 za MLR1 i 0,5000 za MLR2).