Multivariate Adaptive Regression Splines (MARS) Heuristic Model: Application of Heavy Metal Prediction

Abstract

In the last two decades, soft computing modeling such as Artificial Intelligence (AI) approaches have gained a massive attention by the information technology researchers. Nowadays, AI models are improving human abilities in several areas of engineering and science problems. In this paper, we investigate the proficiency of modern heuristic approach called Multivariate adaptive regression splines (MARS) in prediction regression problem. The experimental data set of heavy metal is selected as a case study. The predictive model is conducted based on several physiochemical inputs variables. In order to inspect the MARS model accuracy, Multiple Linear Regression (MLR) model is chosen to compare the results. Here, the root mean square error (RMSE), mean absolute percentage error, and coefficient of determination are examined the accuracy of the models. The finding of the investigated model (i.e., MARS) exhibited a noticeable improving in the prediction accuracy in comparison with MLR.

A Comparative Study of Data-Driven Modeling Methods for Soft-Sensing in Underground Coal Gasification

“…The application and comparison of the efficiency of various learning methods, i.e., NN, Adaptive Regression Splines (MARS) or Support Vector Machines (SVMs) in the UCG data prediction has not yet been the subject of an extensive study, but similar applications in steel making processess and biomass gasification are registered (e.g., [22,23]).…”