In this study, we perform the automatic classification of the signals generated by mining activities using supervised learning algorithms. The target mine of this study is a limestone mine in Danyang, South Korea. First, we select appropriate attributes in the time and frequency domains to characterise the signals. We manually label the recorded data into four categories: blasting, cleaning, drilling and noise. The labelled signals are divided into the training and test data sets. We train and test the 22 models provided by MATLAB and find the most accurate model. To enhance the reliability, we adopt cross-validation strategy during training and repeat training and test for 10 cases which have different composition of training and test data sets. As a result, it is proved that the bagging tree with ensemble is the most appropriate model for our problem. The accuracy of this model is approximately 98.7 %. We evaluate the trained model by using the cross-plot of attributes and confusion matrix of trained results.


Article metrics loading...

Loading full text...

Full text loading...


  1. Cheon, D.S., Park, E.S., Jung.Y.B., Park, C. & Synn, J.H.
    (2008). Monitoring technique using acoustic emission and microseismic event. Tunnel & underground space, 18(1), 1–9.
    [Google Scholar]
  2. Ge, M.
    (2005). Efficient mine microseismic monitoring. International Journal of Coal Geology, 64(1–2), 44–56.
    [Google Scholar]
  3. Ishida, T.
    (1999). An Introduction to Acoustic Emission of rock. Kinmiraisha Nagoya, p. 213.
    [Google Scholar]
  4. Koerner, R. M., McCabe, W. M., & Lord, A. E.
    (1981). Overview of acoustic emission monitoring of rock structures. Rock mechanics, 14(1), 27–35.
    [Google Scholar]

Data & Media loading...

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error