An improved DBSCAN mine water inrush spectrum identification method based on MVO

Abstract

The invention discloses an improved DBSCAN mine water inrush spectrum identification method based on MVO. The advanced multi-cosmos optimization (Multi-Verse Optimizer (MVO) algorithm is used to improve the unsupervised clustering algorithm DBSCAN, and the parameter Eps of DBSCAN algorithm is self-optimized. Finally, the optimized parameters are combined with DBSCAN algorithm to identify the laser-induced fluorescence spectrum of coal mine water inrush. In the process of parameter optimization of the improved algorithm, the value interval of the cosmic position of MVO algorithm is set according to the spatial distance of the water sample spectral data, and the manual input parameters are reduced to the maximum extent possible. The improved DBSCAN algorithm based on MVO not only saves tedious manual parameter optimization process, but also can output the value interval of the parameter Eps corresponding to the optimal recognition rate. In addition, the unsupervised learning algorithm isused to identify the spectrum of coal mine water inrush, which can minimize the false identification of unknown water source.

Description

technical field [0001] The invention belongs to the field of spectral identification of mine water inrush, in particular to an improved DBSCAN mine water inrush spectral identification method based on MVO. Background technique [0002] my country is a big energy consumption country, in which coal resources occupy a dominant position in my country's energy, and this status quo cannot be changed in a short period of time, so the safe mining of coal is of great significance to my country's energy security. In coal mining, mine water inrush seriously threatens the safe production of coal mines. With the increase of mining intensity and depth, coal mines are facing more and more threats of water inrush. After a water inrush occurs in a mine, quickly and accurately identifying the source of the water inrush so as to take the most appropriate response is one of the effective ways to reduce the water inrush loss. [0003] Laser-induced fluorescence technology has been successfully ...

AI Technical Summary

Problems solved by technology

In the field of pattern recognition and machine learning, for labeled data, the performance of supervised learning algorithms is always better than that of unsupervised learning algorithms, but supervised learning algorithms rely heavily on the category information and label information of data. For water inrush spectrum recognition, if The type of water inrush that occurs is not in the training set, and the use of supervised learning algorithms is prone to misidentification and misjudgment

[0004] The unsupervised learning algorithm DBSCAN is based on the density of sample data during clustering. Although it does not require the label and category information of training samples, using DBSCAN for mine water inrush spectrum identification can effectively reduce the misidentification of unknown water sources. However, DBSCAN The algorithm needs to set appropriate parameters when clustering. As far as the DBSCAN algorithm is concerned, a large number of experiments are required to set appropriate parameters, and the process is relatively cumbersome.

Images