Random forest regression evaluation model of regional flood disaster resilience based on the whale optimization algorithm

Abstract

This study proposes a flood disaster resilience evaluation model based on an improved random forest model, which is used to solve the fuzziness problem in resilience evaluations. The model uses the whale optimization algorithm (WOA) to determine the key parameters in the traditional random forest regression (RFR) model and combines the evaluation index set constructed by the Driving forces-Pressure-State-Impact-Response (DPSIR) model to output the resilience index of the study area. This approach has certain advantages in solving the spatiotemporal distribution problem of disaster resilience and can analyze the temporal and spatial variability of the research area and the key driving factors. Taking the Jiansanjiang Administration of Heilongjiang Province of China as an example, the model was used to analyze the resilience of flood disasters in 15 farms under the jurisdiction of the region from 2002 to 2016. The results showed that the level of resilience to flood disasters in the Jiansanjiang Administration was generally increasing at a growth rate of 4.175/10a. In addition, the level of flood resilience was spatially different as shown by the high level of resilience in the southwest and low level in the northeast. The degree of differentiation between farms increased between 2006 and 2011 and decreased between 2012 and 2016. The study also found that economic indicators and population indicators have a greater impact on the assessment results. Compared with the stochastic forest regression model optimized by particle swarm optimization (PSO-RFR) and the RFR model, the WOA-RFR model has outstanding advantages in fitting accuracy, generalization performance. The rationality coefficient and stability coefficient of the WOA-RFR are 0.964 and 0.976, respectively, which have reached a high level. The proposed WOA-RFR model can be used to perform regional disaster resilience evaluation, provide stable technical support and establish a scientific basis for regional disaster prevention and mitigation to ensure regional production safety and sustainable development.