Simulation of the Compressive Strength of Cemented Tailing Backfill through the Use of Firefly Algorithm and Random Forest Model

Cemented tailings backfill is widely used in worldwide mining areas, and its development trend increasing due to the technical economic benefits. However, there no reliable simple machine learning model for prediction of compressive strength. In present study, research process use artificial intelligence algorithms predict strength cemented tailing was conducted, overcoming shortcomings traditional empirical formulas. Experimental tests measure were conducted construct dataset learning. Five input parameters (tailing cement ratio, percentage fine tailings, type, curing time, solid water ratio) considered design laboratory tests. The firefly algorithm (FA) tune random forest (RF) hyperparameters, it adopted combine RF improve accuracy efficiency backfill. By comparing predicted actual results, reliability proposed are confirmed. Tailing ratio time two most important