multivariate linear regression (mlr)

Motor oils have different physicochemical properties, namely viscosity, viscosity index, flash point, pour point, etc. Viscosity is one of the important properties of motor oils since all the properties of industrial lubricants are referred to as their viscosities. The changes in viscosity with variation in temperature are regarded as the viscosity index. The greater the viscosity index, the lower the chances of the viscosity of motor oil with temperature and vice versa. According to the importance of viscosity index in lubricants and because the viscosity index of lubricants is dependent on the chemical composition of motor oils, thus in this study, a simple spectroscopic technique like Fourier Transform InfraRed (FT-IR) spectroscopy was used to analyze the Behran motor oils. The important wavenumbers that affect the viscosity indices were identified by using the Genetic Algorithm (GA) as a variable selection method. By using this method, some functional groups like Alkyl halides, Alkene, Nitro, Acid, Alkane, Alkyne, and Alcohol were recognized that affect the viscosity index of motor oils. Modeling the viscosity index of motor oils was done by Multivariate Linear Regression (MLR) method. Various data preprocessing techniques like Mean Centering and Auto-scaling were operated before the MLR and GA-MLR techniques. The results of modeling were evaluated by using different parameters like regression coefficients (R2) and Root Mean Square Error (RMSE). The values of R2 and RMSE, obtained by the GA-MLR were 0.998 and 0.954 respectively.