An anomaly identification method, apparatus, and storage medium based on time series features

Abstract

This application discloses an anomaly identification method, apparatus, and storage medium based on time series features, used to reduce misjudgments of abnormal moments in mine roadway space. The method involves: acquiring rock mass structure data and rock mass environment data of the mine roadway space; generating interlayer stability parameters based on the rock mass structure data and rock mass environment data; generating a window data matrix using the collected data set of mine characteristic indicators; generating an interlayer crack disturbance index set based on the collected data corresponding to the interlayer stability parameters and vibration-type features, and replacing the interlayer crack disturbance index set in the window data matrix; rearranging the elements of structural stability features and mine environment features in the window data matrix according to the interlayer crack disturbance index set; generating anomaly weights for each type of mine characteristic indicator in the window data matrix; generating an anomaly weighting matrix based on the anomaly weights and the window data matrix; and determining the anomaly occurrence time in the mine roadway space based on the anomaly weighting matrix.

Images