An Improved DWKNN-Based Rock Fall Distance Prediction Model

  • Qingjie Qi, Shuai Huang, Wengang Liu, Yingjie Liu, Shufeng Zhai

Abstract

In this study, we present a rock fall distance prediction model by improving the DWKNN algorithm. In this model, a method to measure the distance of falling rocks by slope height is proposed, which is a relatively simple and practical approach; the rock fall weight, slope height, slope angle, equivalent friction coefficient, horizontal bounce height and vertical bounce height are chosen as the main influencing predictors of the slope rock fall distance. Then, we apply the prediction model to the prediction of rock fall distance. The training samples are collected from 50 typical cases of rock fall while the predicting samples are 10, and the test results are determined by cross-training for 5 times. By comparison with other three popular methods, our proposed prediction model is more accurate. At last, the accuracy of our proposed model is verified by a series of laboratory tests, and the experimental results show that our proposed model is reliable.

Keywords- DWKNN algorithm, rock fall distance, prediction model, laboratory test

Published
2021-02-01
Section
Articles