Determination of reinforcement degrees in constructing large-scale structures by using multiclass support vector machines

Recently, as the birthrate declines and the population ages, labor shortage is becoming an issue in the construction industry, especially in the field of building and maintaining large-scale infrastructure. This paper focuses on determination of the reinforcement degrees at construction sites of a largescale structure, which has traditionally required the experience and skill of experts. We formulate the decision as a recognition problem based on the various risk scores observed at each section of the construction site, which can resolve the inconsistency in the structure data by using risk scores of the current and previous sections, and we introduce multiclass SVMs that exploit the monotonicity of the data. Finally, we apply the proposed methods to data at various real construction sites of one kind of large-scale structure, where we train the SVMs on known datasets obtained at many known sites of the structure, and then evaluate the classification rates of the obtained classifiers to the unknown dataset of the structure. (C) 2022 Elsevier B.V. All rights reserved.

Automated summary

This paper focuses on the labor shortage issue in the construction industry, particularly in the field of large-scale infrastructure. By recognizing the risk scores at different sections of the construction site, the paper addresses the inconsistency in structure data and introduces multiclass SVMs to improve classification accuracy.