Optimization of Polynomial Functions on the NuSVR Algorithm Based on Machine Learning: Case Studies on Regression Datasets

Abstract

Purpose: Experimental studies are usually costly, time-consuming, and resource-intensive when it comes to investigating prospective corrosion inhibitor compounds. Machine learning (ML) based on the quantitative structure-property relationship model (QSPR) has become a massive method for testing the effectiveness of chemical compounds as corrosion inhibitors. The main challenge in the ML method is to design a model that produces high prediction accuracy so that the properties of a material can be predicted accurately. In this study, we examine the performance of polynomial functions in the ML-based NuSVR algorithm in evaluating the regression dataset of corrosion inhibition efficiency of pyridine-quinoline compounds.Methods: Polynomial functions for NuSVR algorithm-based ML.Result: The outcomes demonstrate that the NuSVR model's prediction ability is greatly enhanced by the application of polynomial functions. Originality: The combination of polynomial functions and deep machine learning based NuSVR algorithms to increase the accuracy of predictive models.

Prediction of Corrosion Inhibition Efficiency Based on Machine Learning for Pyrimidine Compounds: A Comparative Study of Linear and Non-linear Algorithms

“…In order to mitigate potential biases and variance problems, we employed the K-Fold cross-validation technique, thereby bolstering the reliability of our model evaluation. The model's performance was evaluated using a set of metrics including MSE, RMSE, MAE, MAPE, and the R-Square [2][3][4][5][6][7][8][9], which that metrics are quantify the model's predictions accuracy. A lower value for these metrics indicates a higher level of precision in the model [20,21].…”

“…Utilizing corrosion inhibitor technologies, which can decelerate corrosion rates in metals such as steel, iron, and aluminum, has decreased these expenses by as much as 35% [6]. Although Density Functional Theory (DFT) has the potential to assess corrosion inhibitors, conventional experimental techniques for corrosion treatment continue to be expensive and timeconsuming [7][8][9].…”

“…This has led to improved machine learning processes, including classification, clustering, and model development. Machine learning (ML) techniques that utilize quantitative structure-activity relationships (QSAR) or quantitative structure-property relationships (QSPR) have shown promise in the field of corrosion inhibitors, both in terms of efficiency and effectiveness [8][9][10][11]. Prior studies have examined different algorithmic models, including Partial Least Squares (PLS), Multiple Linear Regression (MLR), Random Forest (RF), Autoregressive with Exogenous Inputs (ARX), Support Vector Machine (SVM), and similar models [12][13][14][15].…”

Perbandingan Algoritma Multilinear Regression dan Decision Tree Regressor dalam Memprediksi Efisiensi Penghambatan Korosi Piridazin

“…Performa penghambatan laju korosi oleh inhibitor organik dapat mencapai 92 hingga 98 persen (Akrom, 2022). Kemampuan untuk membentuk lapisan molekuler yang melindungi permukaan logam dari kontak langsung dengan pemicu korosif merupakan mekanisme utama di balik efektivitas senyawa organik (Budi et al, 2023). Senyawa turunan piridazin menunjukkan potensi yang menarik karena memiliki gugus fungsi yang beragam, termasuk sulfur, nitrogen, dan atom oksigen dalam kerangka struktur molekulnya, yang berpengaruh baik dalam mendukung kemampuan adsorpsi inhibitor dalam permukaan baja (Hameed et al, 2020).…”

Implementasi Fungsi Polinomial pada Algoritma Gradient Boosting Regressor: Studi Regresi pada Dataset Obat-Obatan Kadaluarsa Sebagai Material Antikorosi

“…Implementasi fungsi polynomial terbukti meningkatkan kinerja algoritma NuSVR secara signifikan [18]. Oleh karena itu, pada penelitian ini mengimplementasikan fungsi polynomial untuk meningkatkan akurasi model dari algoritma GBR.…”