一种改进自适应步长人工蜂群优化随机森林的滑坡风险性评价方法

A Landslide Risk Assessment Method with Improved Adaptive Step Size Artificial Bee Colony Optimization Random Forest

融合群智能全局优化人工蜂群(artificial bee colony,ABC)算法与随机森林(random forest,RF)模型,提出一种改进自适应步长人工蜂群(improved adaptive step size artificial bee colony,IASSABC)算法优化RF模型的滑坡风险性评价方法。针对ABC算法存在的后期收敛速度慢与早熟的问题,融合全局最优解和领域半径参数的思想,发展了IASSABC算法并用于优化调整RF模型参数,以提升RF模型的泛化性能与抗过拟合能力,降低RF模型陷入局部最优解的风险。实际案例表明,本文方法兼具引入最优解的自适应步长人工蜂群(adaptive step size artificial bee colony,ASSABC)算法具有的强针对性与方向性优点,以及引入领域半径参数的改进人工蜂群(improved artificial bee colony,IABC)算法具有的全局参数最优解强搜索能力。相较于传统的ABC、ASSAB、IABC模型,IASSABC优化RF模型在滑坡风险性评价中具有更强的参数最优解搜索能力和更高的评价准确性。

In this study,we develop a landslide risk assessment method with improved adaptive step size artificial bee colony(IASSABC)optimization random forest(RF)model by integrating the swarm intelligence global optimization artificial bee colony(ABC)algorithm with the RF model.Aiming at the shortcomings of ABC algorithm,such as slow convergence speed and premature maturity,the IASSABC model is developed by integrating the ideas of global optimal solution and domain radius parameter,and is used to optimize and adjust the parameters of the RF model,so as to improve the generalization performance and overfitting ability of the RF model,and to reduce the risk of falling into the local optimal solution.The actual case of landslide risk assessment shows that the IASSABC model combines the advantages of strong targeting and directionality of the adaptive step size artificial bee colony(ASSABC),which introduces the optimal solution,and the strong search ability of global parameter optimal solution of the improved artificial bee colony(IABC),which introduces the domain radius parameter.Compared with the traditional ABC,ASSABC and IABC models,the IASSABC optimized RF model has a stronger ability to search for the optimal solution of the parameter in landslide risk evaluation,and has a higher evaluation accuracy.