Entry-type mining methods such as cut and fill methods play a significant role in the underground extraction of deposits and it is still a suitable method for extracting deposits with an irregular shape and with weak host rocks (especially hanging wall). The stability of the roof opening in this mining method is especially important for the safety of the working personnel. In recent years, various methods have been developed in order to design and eva‎luate the stability of span in Entry-type mining methods, which can be divided into three main categories: analytical, numerical and experimental. Experimental methods are often more useful than other methods due to their simplicity. Therefore, in this research, relying on a database consisting of 399 data from the records of cut and fill mining in Canada and using intelligent methods, new experimental models have been developed to eva‎luate the stability of span. In this thesis, decision tree-based algorithms such as CART, LMT, and J48 algorithms and artificial intelligence-based algorithms such as GEP algorithm and regression methods such as MARS and LR were used. In the development of these models, 85% of the data in the database has been used for model training and the remaining 15% for model testing. After developing the models, their performance was eva‎luated using various statistical indices. The results showed that the model developed using J48 algorithm has the highest performance among all models. It should also be mentioned that all the developed models are of the white box type, which the user can easily use.

ابراهيم قاسمي

لهراسب فرامرزي، سعيد مهدوي