基于随机森林算法的英语自动翻译设备运行态势监测方法  

Monitoring of English Automatic Translation Equipment Based on Random Forest

作　　者：郭春玲[1] GUO Chunling(International School of Xi’an Siyuan University,Xi’an 710038,China)

机构地区：[1]西安思源学院,西安710038

出　　处：《自动化与仪器仪表》2023年第1期178-183,共6页Automation & Instrumentation

基　　金：横向课题《大学英语翻译技巧与实践》(SYHX-2019001)。

摘　　要：针对传统英语翻译服务机器人故障诊断准确率低,导致机器人设备运行监测效果变差,安全性降低的问题。基于随机森林和梯度提升树算法,将两者相结合得到RF-GBDT故障特征选择算法;然后基于GRU神经网络,提出一种改进的故障诊断混合模型,通过此模型实现翻译设备故障准确诊断和运行态势监测。试验结果表明,从39维向量至29维向量的特征选择中,提出的RF-GBDT算法运算效率提高了30%及以上。算法应用发现,提出的RF-GBDT算法的故障诊断率最高可达92.5%,相较于未进行特征选择的算法,本算法可有效提升故障诊断率。对比于其他故障诊断模型,提出的GRU混合模型的诊断准确率高达94.3%,故障诊断精度明显更高,诊断效果更好,可提升英语翻译机器人的安全性。In view of the low fault accuracy of traditional English translation service robot, the operation monitoring effect and safety of robot equipment are reduced.Based on random forest and gradient lifting tree algorithm, RF-GBDT fault feature selection algorithm is combined, and then an improved fault diagnosis hybrid model based on GRU neural network is proposed to achieve accurate translation equipment fault diagnosis and operation situation monitoring.The experimental results show that the operational efficiency of the proposed RF-GBDT algorithm increases by 30% or more from 39 to 29 dimensions.The algorithm application shows that the fault diagnosis rate of the proposed RF-GBDT algorithm can reach 92.5%. This algorithm can effectively improve the fault diagnosis rate compared with the algorithm without feature selection.Compared with other fault diagnosis models, the diagnosis accuracy of the GRU hybrid model proposed in this paper is as high as 94.3%, the fault diagnosis accuracy is significantly higher, and the diagnosis effect is better, which can improve the safety of the English translation robot.

关 键 词：随机森林算法 英语自动翻译 特征选择 故障诊断 GRU 

分 类 号：TP392[自动化与计算机技术—计算机应用技术]