Stability assessment of hard rock pillars using two intelligent classification techniques: A comparative study. (September 2017)

• Record Type: Journal Article
• Authors: Ghasemi, Ebrahim; Kalhori, Hamid; Bagherpour, Raheb
• Is Part Of: Tunnelling and underground space technology. Volume 68(2017)
• Page Start: 32
• Page End: 37
• Publication Date: 2017-09
• Subjects: Pillar stability -- Underground hard rock mines -- J48 algorithm -- Support vector classification (SVC); Tunneling -- Periodicals; Underground construction -- Periodicals; Tunnels -- Periodicals; Underground areas -- Periodicals; 624.193
• DOI: 10.1016/j.tust.2017.05.012
• Languages: English
• ISSNs: 0886-7798

Abstract

Highlights: Two models were proposed for pillar stability prediction in underground hard rock mines. J48 and SVC algorithms were employed for building and evaluating models. The results showed that the models can predict pillar stability in a reliable manner. The J48 model exhibited better performance than SVC model. Abstract: One of the most challenging safety problems in underground hard rock mines is pillar stability during mining operation. This paper presents an assessment of J48 and SVC application for pillar stability prediction in underground hard rock mines. Based on a database compiled from various hard rock mines and using these algorithms, two stability graphs are developed. The performance assessment of models indicates that both models can predict pillar stability with acceptable accuracy. In comparison with logistic regression model, the prediction capability of J48 and SVC models is better, but the J48 model shows superiority over two other models.