Identifying Posterior Classification Probabilities in Insulating Fluids using Classification Algorithms

The posterior probability of the oil samples— mineral oil (MO), punga oil (PO), and modified punga oil (MPO)— are utilized in this study. The posterior probability of the oil samples is discovered using two classification techniques. Breakdown voltage (BDV), kinematic viscosity (KVIS), dielectric constant (DC), and moisture content were the dielectric parameters employed in this investigation (MC). The classification of clusters and visualization of their 2D interactive effect, visualization of the 3D decision surface at various clusters, examination of Gaussian mixture assumptions, the Q- Q plot, and consistency testing using the Mardia Kurtosis test for linear and quadratic discriminant analyses are the main counterparts of the work. The study’s findings show that using the observed dielectric properties of the oil samples, it may be possible to comprehend their subsequent roles.