中文摘要：

为实现对飞机维修人员安全性的有效预测,构建了维修人员安全性评价指标。通过专家打分法获得了维修人员安全性评价指标的原始训练数据和测试数据,分别采用BP-GDM模型、BP-LM模型和GA-BP-LM模型对数据进行了训练和预测,仿真结果表明：BP-GDM模型预测精度最差,且训练耗时较长,而GA-BP-LM模型不仅可以获得较高的预测精度,且耗时最短,BP-LM模型居中。此外GA-BP-LM模型相对误差值波动幅度最小,平均相对误差仅为3.33%,达到了预测的精度要求,为飞机维修人员安全性预测提供了有效手段。

英文摘要：

In order to scientific predict the safety of the aircraft maintenance technicians,the safety evaluation index is established.Different kinds of aircraft maintenance technicians safety evaluation index originality training data and test data are gained by expert mark method,used the BP-GDM model,BP-LM model and GA-BP-LM model to train and predict,the simulation result show that： the forecast precision of BP-GDM model is the worst and consume lots of time,but GA-BP-LM gained the best forecast precision and the lest simulation time,the BP-LM was placed in the middle.What＇s more GA-BP-LM model had the smallest opposite error fluctuate,and the average opposite error value is only 3.33%,which matched well with the expect requirement,the model provide effective method for predict of aircraft maintenance technicians safety.