A Comparative Study of Six Hybrid Prediction Models for Uniaxial Compressive Strength of Rock Based on Swarm Intelligence Optimization Algorithms

Uniaxial compressive strength (UCS) is a significant parameter in mining engineering and rock engineering. The laboratory test time-consuming economically costly. Therefore, developing reliable accurate UCS prediction model through easily obtained parameters good way. In this paper, we set five input compare six hybrid models based on BP neural network swarm intelligence optimization algorithms–bird algorithm (BSA), grey wolf (GWO), whale (WOA), seagull (SOA), lion (LSO), firefly (FA) with the accuracy of two single without optimization–BP random forest algorithm. Finally, above eight were evaluated compared by root mean square error (RMSE), absolute percentage (MAPE), coefficient determination (R 2 ), a10 index to obtain most suitable model. It indicated that best FA-BP model, RMSE value 4.883, MAPE 0.063, R 0.985, an 0.967. Furthermore, normalized mutual information sensitivity analysis shows point load effective UCS, respectively.