Grade Classification Of Pontaneous Combustion Tendency Of Sulfide Ores With Improved Gaussian Processes

At present, kemel methods has brought increasing concerns and become a hot spotof research on account of its function to solve nonlinear problems and insufficiency ofsample quantity in the field of grade classification of spontaneous combustion tendencyof sulfide ores. Based on kemel methods, Gaussian Process Classifier can illustrateclassification problems with entire Bayesian formula and structure a probabilistic modeldefinitely so that the results would be easily explained.This paper applies the predicted method of spontaneous combustion tendency tomend the defects in classification of spontaneous combustion tendency of sulfide ores. Itutilizes public data set to build a Gaussian Process classification model based onExpectation Propagation method and Laplace method, which is compared with SupportVector Machine, furthermore, it indicates that the Gaussian process classification modelbased on EP could acquire the optimal classification result. The paper analysis thedisadvantages of Gaussian Process and Discriminate Analysis, and puts forward amodified Gaussian Process Classifier using latent variable models trained withdiscriminative priors over the latent space, which can learn a discriminative latent spacefrom a small training set. It ultimately realizes the sufficient excavation in rankinformation of spontaneous combustion tendency of sulfide ores, and provides a newapproach to solve the rank classification of spontaneous combustion tendency of sulfideores...