APPLICATION OF ENSEMBLE LEARNING METHODS TO ESTIMATION OF ADHERED SALT FOR CONCRETE BRIDGES

Abstract

Application of the machine learning methods is attempted to estimate the adhered salt on concrete bridges. In order to select the best evaluation model, performance of four algorithms belonging to the ensemble learning in the machine learning is discussed with examining the explanatory variables and data handling methods of machine learning. Based on the obtained parameters, the amount of adhered salt is then evaluated, and the validity of the developed simple method is verified.