Nonlinear regression in environmental sciences by support vector machines combined with evolutionary strategy

A carbon risk prediction model for Chinese heavy-polluting industrial enterprises based on support vector machine

“…Although in two-dimensional linear separable cases, there are many possible linear classifier which can correctly classify into two types. Beyond that, SVM also tries to find an optimal hyperplane as accurately as possible so as to maximize the margin of classification intervals between both sides, namely minimize the deviation of the prediction to the utter most [36][37][38].…”

Evolutionary approach for crowdsourcing quality control

“…It is used to solve many real world problems. Hu and Di Paolo [8] use genetic algorithms to solve the problem of air traffic control in multirunway systems; Szlapczynski and Szlapczynska [15] apply evolutionary algorithms and some assumptions of game theory to solve ship encounter situations; Lima et al [12] propose a hybrid algorithm which combines support vector regression with evolutionary strategy for predictive models in the environmental sciences; Zhang et al [19] develop an approach and prototype for selecting optimal material constituent compositions. There are more developments and applications of evolutionary computing that are not covered in this related work.…”

“…It is the process of incrementally improving quality or achieving a goal in a given environment, and it is widely used in many different applications [8,15,12,19]. Theoretically, it is so powerful that it has the ability to tackle any search space provided that initialization and variation operators are available [6].…”

Regression trees for modeling geochemical data—An application to Late Jurassic carbonates (Ammonitico Rosso)

“…Over the last decades a large amount of machine learning methods have emerged. Among the most widely used are decision trees (Breiman, 1984;Leibovici et al, 2011;Qi and Zhu, 2011;Zhang et al, 2009), artificial neural networks (Baykan and Yilmaz, 2010;Bue and Stepinski, 2006;Canty, 2009;Dubois et al, 2007;Mas and Flores, 2008;Pavel et al, 2011), support vector machines (Lima et al, 2012;Mountrakis et al, 2011;Petropoulos et al, 2012;Yu et al, 2012;Zuo and Carranza, 2011) and classifier ensembles (Breiman, 1996;Rodriguez-Galiano et al, 2012a, just to mention a few. These methods start from very diverse conceptual bases, although all of them have a series of shared advantages: (i) ability to learn complex patterns, considering nonlinear relationships between explanatory and dependent variables; (ii) generalization ability, hence applicable to incomplete or noisy databases; (iii) possibility to incorporate a priori information; and (iv) integration of different types of data in the analysis due to the absence of assumptions about the data used (e.g.…”