Utilizing gradient boosted machine for the prediction of damage to residential structures owing to blasting vibrations of open pit mining

Abstract

Owing to the complex features of blasting vibration damage assessment systems, a gradient boosted machine (GBM) model is developed for the classification of residential structure damage (RSD) due to blasting vibrations of open pit mining. Twelve indicators are defined as the indices for the prediction of RSD in the proposed model. These are: peak particle velocity, dominant frequency, dominant frequency duration, distance, maximum safe charge per delay, compressive strength of mortar joints, ratio of brick area to house area, height of residential house, roof structures, beam-column frames, quality of construction, and site conditions. The GBM model is achieved by training 108 sets of measured data of blasting vibration. A 10-fold cross-validation procedure was applied to determine the optimal parameter values during modeling, and an external testing set was employed to validate the prediction performance of the model. Two performance measures – classification accuracy rate and Cohen’s kappa – have been employed. The analysis of accuracy together with kappa for the dataset demonstrate that the GBM model has high credibility as it achieves a comparable median classification accuracy rate and Cohen’s kappa values of 91.7% and 0.875 for the prediction of RSD, respectively.