The application of pattern recognition for the prediction of thecoal mining water irruption

摘要

Application of pattern recognition is discussed in the predictionof the water irruption from the floor of the underground coal mine. Byexhaustive analysis of the hydrogeological data acquired from the coalmine, three data features are selected as the most useful out of sixproposed features. Over 90% of prediction accuracy is achieved withBayes classifier. For the reliability of the decision function, anergodic stability test of the function is developed that shows that thereliability of the decision function is over 89%. The prediction showsthat the pattern recognition method is applicable