A Coal Mine Gas Data Labeling Method Based on Single Class Support Vector Machine

Abstract

The present invention belongs to the safety neighborhood of coal mines, which involves a coal mines gas data marking method based on single -category support vector machines, including: real -time obtaining raw data sets, classification of raw data sets, and obtaining the ductal unbalanced data set; adopt useThe K Means cluster algorithm is processed to the second category unbalanced dataset to get the K sample pool; each sample pool includes an unsigned sample pool and a sample pool;Enter into the single -class support vector model model, predict the label label in the unsigned sample pool; the original data is marked according to the predicted labeling label;The class support vector machine actively learns, reducing the marking sample of coal mines gas data; the present invention adds density and distribution information in the process of active learning, making the selected sample more representative.

Description

technical field [0001] The invention belongs to the safety neighborhood of coal mine gas, and in particular relates to a method for marking coal mine gas data based on a single-class support vector machine. Background technique [0002] In the field of coal mine gas safety, the amount of data involved is very large. In order to conduct research on coal mine data, a large number of sample labels are required using traditional supervised learning methods. In order to reduce the marking cost, it is very necessary to use active learning to reduce the marking samples. Active learning is a machine learning algorithm that obtains the sample labels of all samples by marking some samples with a large amount of information in the sample pool. Traditional coal mine data labeling methods include uncertainty sampling methods and methods based on committee inquiries. In addition to active learning methods for balanced data, there are some active learning methods for specialized imbalan...

AI Technical Summary

Problems solved by technology

However, coal mine data is highly unbalanced data, which makes the initialization of the algorithm difficult and the importance of samples selected by active learning is reduced.

[0004] Using the unbalanced data active learning method to process coal mine data often leads to waste of labels, making marking time-consuming and laborious; how to further reduce the number of marking samples is very necessary for marking suitable for coal mine gas safety data

Images