Induction and pruning of classification rules for prediction of microseismic hazards in coal mines

详细信息    查看全文

• 作者：Marek Sikora ; ^a ; ^b ; ^{marek.sikora@polsl.pl}
• 关键词：Rules induction ; Rules quality measures ; Classification ; Microseismic hazard prediction
• 刊名：Expert Systems with Applications
• 出版年：2011
• 出版时间：June 2011
• 年：2011
• 卷：38
• 期：6
• 页码：6748-6758
• 全文大小：771 K

文摘

The paper presents results of application of a rule induction and pruning algorithm for classification of a microseismic hazard sate in coal mines. Due to imbalanced distribution of examples describing states “hazardous” and “safe”, the special algorithm was used for induction and rule pruning. The algorithm selects optimal parameters‘ values influencing rule induction and pruning based on training and tuning sets. A rule quality measure which decides about a form and classification abilities of rules that are induced is the basic parameter of the algorithm. The specificity and sensitivity of a classifier were used to evaluate its quality. Conducted tests show that the admitted method of rules induction and classifier’s quality evaluation enables to get better results of classification of microseismic hazards than by methods currently used in mining practice. Results obtained by the rules-based classifier were also compared with results got by a decision tree induction algorithm and by a neuro-fuzzy system.