基于BP神经网络的洪涝灾害承灾体脆弱性评估OACSTPCD

To reduce the damage of flood disasters to the bearing body,Beijing City as an example,an assessment framework for vulnerability of flood disaster bearing body considering exposure,sensitivity and adaptive capacity was developed.The entropy-weighted TOPSIS algorithm was applied to objectively weight the vulnerability indicators of flood disaster bearing body and obtain the standardized vulnerability index.The initial weights and vulnerability indices were used as input and output datasets,respec-tively,for training a BP neural network to obtain optimized weights.Furthermore,the ArcGIS technique was combined to assess the vulnerability of flood disaster bearing body,and the natural break method was used to classify the vulnerability into 4 levels.The results showed that:①Population density,economic density,urban POI density,vegetation coverage and drainage pipe network sig-nificantly influenced the vulnerability of flood disaster bearing body in Beijing City.②The vulnerability of flood disaster bearing body in Beijing city gradually decreased from southeast to northwest in spatial distribution,with higher vulnerability levels in the central urban area and lower vulnerability levels in the peripheral areas.The research results have certain guiding significance for reducing the vulnerability of flood disaster bearing body in Beijing City,and the weight optimization model and vulnerability as-sessment model can also be applied to the other cities.