Relevance vector machine with tuning based on self-adaptive differential evolution approach for predictive modelling of a chemical process

Abstract

In the past decade, relevance vector machines have gained the attention of many researchers, and this machine learning technique is a Bayesian sparse kernel method, both for classification and regression problems. In general, the choice of appropriate learning hyperparameters is a crucial step in obtaining a well-tuned model. To overcome this issue, we apply a self-adaptive differential evolution algorithm. In this paper, we propose a relevance vector machine for regression combined with a novel self-adaptive differential evolution approach for predictive modelling of phosphorus concentration levels in a steelmaking process with real data. We compared the performance of proposed relevance vector machine (RVM) with other machine learning techniques, such as random forest (RF), artificial neural network (ANN), K-nearest neighbors (K-NN), and also with statistical learning techniques as, Beta regression model and multiple linear regression model. The RVM has performance better than RF, ANN, K-NN, and statistical techniques used. Our study indicates that RVM models are an adequate tool for the prediction of the phosphorus concentration levels in the steelmaking process.