融合多源数据与IV-BO-XGB模型的神木市滑坡易发性评价OA

Landslide hazard susceptibility assessment plays an important role in the prevention and management of geological disasters.Aiming at the problem that the traditional model relies on subjective experience to carry out superparametric optimization,which leads to the limitation of the identification accuracy of high-risk areas and the generalization ability of the model,this study takes Shenmu City as an example to build a landslide susceptibility evaluation method system with the synergistic effect of"dynamic deformation monitoring sample optimization superparametric optimization".The Bayesian optimization(BO)algorithm was used to optimize the super parameters of random forest(RF)and extreme gradient boosting tree(XGboost),and coupled models of information value(IV)-Bayesian optimization-random forest(IV-BO-RF)and information value-Bayesian optimization-extreme gradient boosting(IV-BO-XGB)were constructed to evaluate the landslide susceptibility in Shenmu City.The results show that the average accuracy of the model is improved by 3.47％～4.28％after the optimization of BO algorithm,in which the area under curve(AUC)value of IV-BO-XGB is 0.960,the proportion of disaster points in extremely high prone areas is 71.05％,and the misjudgment rate is only 4.89％,which has better generalization ability on the whole.Through the collaborative innovation of introducing dynamic factors,optimizing sample distribution,and adjusting parameters of intelligent algorithms,it breaks through the limitations of traditional methods in dynamic feature capture and model parameter optimization,and provides a reference for landslide disaster prevention and control in Shenmu City.