基于BAS-ELM的地震经济损失预测

Prediction of earthquake economic losses based on BAS-ELM

为提高地震经济损失预测的准确性和有效性,提出了基于天牛须算法(beetle antennae search,BAS)优化极限学习机(extreme learning machine,ELM)的地震经济损失预测模型。以1996—2014年破坏性地震经济损失为样本数据,选取震级、震中烈度、人口密度和人均GDP等4个影响指标作为模型的输入向量,直接经济损失为输出向量,同时利用BAS优化ELM模型变量,从而消除了随机变量对预测结果的影响,最终建立基于BAS-ELM的地震经济损失预测模型。将建好的BAS-ELM模型用于测试样本的预测,并同其它模型进行了比较。结果表明:BAS-ELM的预测准确率为97.244%,具有更好的预测精度。

In order to improve the accuracy and effectiveness of earthquake economic losses prediction,a prediction model for earthquake economic losses based on ELM optimized by BAS was proposed.Using the economic losses of destructive earthquakes from 1996 to 2014 as sample data,four influencing factors such as magnitude,epicenter intensity,population density,and per capita GDP were selected as input vectors,and direct economic losses were used as output vectors.BAS was used to optimize the model variables,thereby the influence of random variables was eliminated on the prediction results.Finally,earthquake economic loss prediction model based on BAS-ELM was established.Then BAS-ELM model was used to predict the test samples and compared with other models.The result shows that the prediction accuracy of BAS-ELM is 97.244%,which has better prediction accuracy.