Abstract
In this paper, a fast tunnel cable state identification method based on a sparse online hybrid Gaussian multi-classification algorithm is proposed for real-time monitoring of the state of tunnel cables and their changes. The method first reduces the size of the training set using the data subset approximation method, then uses the hybrid Gaussian multi-classification algorithm to train the tunnel cable state fast identification model, and finally performs the parameter update online when necessary. The proposed method takes less time in the model training phase, and the trained tunnel cable state fast identification model can maintain good recognition accuracy under complex operating conditions and can be updated quickly.

This publication has 19 references indexed in Scilit: