Prediction of Methane Outbreaks in Coal Mines from Multivariate Time Series Using Random Forest

Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 9437)


In recent years we have experienced unprecedented increase of use of sensors in many industrial applications. Examples of such are Health and Usage Monitoring Systems (HUMS) for vehicles, so-called intelligent buildings, or instrumentation on machinery in order to monitor performance, detect faults and gain insights in operational aspects. Modern sensors are capable of not only generating large volumes of data but as well transmitting that data through network and storing it for further analysis. Unfortunately, that collected data requires further analysis in order to provide useful information to the decision makers who want to reduce costs, improve safety, etc. Such analysis proved to be a challenge, as there are no generic methodologies that allow for automating data analysis and in practice costs required to analyze data are prohibitively high for many practical applications. This paper is a step in a direction of developing generic methods for sensor data analysis – it describes an application of a generic method that can be applied to arbitrary set of multivariate time series data in order to perform classification or regression tasks. The presented application relates to prediction of methane concentrations in coal mines based on time series data from various sensors. The method was tested within the framework of IJCRS’15 data mining competition and resulted in the winning model outperforming other solutions.


  1. 1.
    Zagorecki, A.: A Versatile Approach to Classification of Multivariate Time Series Data. In: The Proceedings of the 2015 Federated Conference on Computer Science and Information Systems (2015, to appear)Google Scholar
  2. 2.
    Meina, M., Janusz, A., Rykaczewski, K., Ślęzak, D., Celmer, B., Krasuski, A.: Tagging firefighter activities at the emergency scene: summary of AAIAâĂŹ15 data mining competition at knowledge pit. In: Proceedings of the 2015 Federated Conference on Computer Science and Information Systems (2015)Google Scholar
  3. 3.
    Janusz, A., Krasuski, A., Stawicki, S., Rosiak, M., Slezak, D., Nguyen, H.S.: Key risk factors for Polish State Fire Service: a data mining competition at knowledge pit. Federated Conference on Computer Science and Information Systems (FedCSIS) 2014, pp. 345–354 (2014) doi:10.15439/2014F507
  4. 4.
    Hall, M., Frank, E., Holmes, G., Pfahringer, B., Reutemann, P., Witten, I.H.: The WEKA data mining software: an update. SIGKDD Explorations, vol. 11(1), pp. 10–18 (2009)Google Scholar
  5. 5.
    Hall, M.A.: Correlation-based Feature Subset Selection for Machine Learning. Hamilton, New Zealand (1998)Google Scholar
  6. 6.
    Breiman, L.: Random forests. Mach. Learn. 45(1), 5–32 (2001)MATHCrossRefGoogle Scholar

Copyright information

© Springer International Publishing Switzerland 2015

Authors and Affiliations

  1. 1.Defence Academy of the United KingdomShrivenhamUK
  2. 2.Centre for Simulation and AnalyticsCranfield UniversityBedfordUK

Personalised recommendations