Radial basis neural tree model for improving waste recovery process in a paper industry

Abstract

In this article, we propose a novel hybridization of regression trees (RTs) and radial basis function networks, namely, radial basis neural tree model, for waste recovery process (WRP) improvement in a paper industry. As a by‐product of the paper manufacturing process, a lot of waste along with valuable fibers and fillers come out from the paper machine. The WRP involves separating the unwanted materials from the valuable ones so that the recovered fibers and fillers can be further reused in the production process. This job is done by fiber‐filler recovery equipment (FFRE). The efficiency of FFRE depends on several crucial process parameters, and monitoring them is a difficult proposition. The proposed model can be useful to find the essential parameters from the set of available data and to perform prediction task to improve WRP efficiency. An idea of parameter optimization along with regularity conditions for the universal consistency of the proposed model is given. The proposed model has the advantages of easy interpretability and excellent performance when applied to the FFRE efficiency improvement problem. Improved waste recovery will help the industry to become environmentally friendly with less ecological damage apart from being cost‐effective.

Bayesian neural tree models for nonparametric regression

“…We present two variants of BNT models where each consists of a Bayesian (frequentist) implementation of a tree‐based component for feature selection purposes, and a frequentist (Bayesian) implementation of a neural network component for prediction purposes (see Figure 1). Such hybridisation of decision trees and neural networks in entirely frequentist settings were first proposed and theoretically justified in Chakraborty, Chattopadhyay & Chakraborty (2020); Chakraborty, Chakraborty & Murthy (2019); Chakraborty, Chakraborty & Chattopadhyay (2019). In this work, we extend those approaches but consider frequentist and Bayesian versions of the component models.…”

“…Thus, a hybrid (or ensemble) formulation of trees and neural networks can leverage their strengths and overcome their limitations while applying individually. Several such hybrid models blending CART and ANNs have been discussed in the literature (Utgoff 1989;Sethi 1990;Sirat & Nadal 1990;Kijsirikul & Chongkasemwongse 2001;Micheloni et al 2012;Chakraborty, Chattopadhyay & Chakraborty 2018;Chakraborty, Chakraborty & Chattopadhyay 2019;Chakraborty, Chakraborty & Murthy 2019;Vanli et al 2019;Chakraborty, Chattopadhyay & Chakraborty 2020;Chakraborty & Chakraborty 2020a,b), and have been useful for improving the prediction accuracy of the individual models. These hybrid models, however, only consider frequentist implementations of their components.…”

“…2012; Chakraborty, Chattopadhyay & Chakraborty 2018; Chakraborty, Chakraborty & Chattopadhyay 2019; Chakraborty, Chakraborty & Murthy 2019; Vanli et al . 2019; Chakraborty, Chattopadhyay & Chakraborty 2020; Chakraborty & Chakraborty 2020a,b), and have been useful for improving the prediction accuracy of the individual models. These hybrid models, however, only consider frequentist implementations of their components.…”