Investigation of Best QSPR-Based Machine Learning Model to Predict Corrosion Inhibition Performance of Pyridine-Quinoline Compounds

Abstract

Corrosion is a major concern for the industrial and academic sectors because it causes significant losses in many fields. Currently, there is a great deal of interest in the topic of material damage control using organic chemicals. Pyridine and quinoline are potential corrosion inhibitors because they are non-toxic, inexpensive, and efficient in various corrosive conditions. Experimental studies in searching for candidate corrosion inhibitor candidates require a lot of time, cost, and labor intensive. Using a machine learning (ML) strategy based on a quantitative structure-property relationship (QSPR) model, we evaluate gradient boosting regressor (GBR), support vector regression (SVR), and k-nearest neighbor (KNN) algorithms as predictive models to investigate corrosion inhibition efficiency (CIE) of pyridine-quinoline compounds in this study. We found that the GBR model, when compared with the SVR and KNN models as well as models from the literature for the pyridine-quinoline compound dataset, has the best predictive performance based on the metric coefficient of determination (R2) and root mean square error (RMSE). Overall, our study provides a new perspective on how the ML model can estimate the effectiveness of corrosion prevention on iron surfaces by organic inhibitor compounds.