Slope Stability and Safety Assessment Based on Random Forest Enhanced under Multi-Strategy Pelican Optimization

To improve the prediction accuracy of slope stability and prevent slope failure accidents, this study proposes a slope stability prediction model based on an improved pelican optimization algorithm optimized random forest (Improved Pelican Optimization Algorithm optimized Random Forest, IPOA-RF). First, according to 431 slope cases, the slope height, slope angle, unit weight, cohesion, internal friction angle, and pore water pressure ratio were selected as the main predictive features. Second, due to the issue of excessive hyperparameters in the traditional random forest (RF) model, the IPOA algorithm was employed to optimize the RF parameters using an optimal-guidance strategy, mutation operator, and dynamically adjusted search mechanism. Finally, compared with five other optimization algorithms, the proposed IPOA algorithm exhibited superior parameter optimization ability and convergence performance in ten benchmark test functions. The designed IPOA-RF model achieved an average prediction accuracy of 85.1%, approximately 10.4% higher than that of the traditional RF model (74.7%). The results demonstrate that the IPOA-RF model can rapidly and accurately identify slope stability conditions, effectively overcoming the limitations of conventional methods. This model not only provides an innovative solution for slope stability assessment but also offers technical support for enhancing the safety and operational efficiency of practical slope engineering projects.