基于关系抽取的物联网设备识别方法

IoT Device Identification Method Based on Relationship Extraction

在标语存在信息缺失的情况下,传统物联网设备识别方法难以实现细粒度识别,针对这一问题,提出一种基于关系抽取的物联网设备识别方法。该方法通过在物联网设备的品牌、类型、型号信息与关系抽取中头实体、关系、尾实体之间建立映射关系,将物联网设备识别问题转化为关系抽取问题,解决标语中关键字缺失问题,并且在特征提取时融入旋转式位置编码信息,增强对物联网设备标语的长文本处理能力,提高物联网设备细粒度识别的性能。实验结果表明:该方法在设备型号粒度上的精确率达到了92.5%,F1值达到88.8%,优于基线方法。

In the absence of complete information in banners,traditional methods for IoT device recog⁃nition face challenges in achieving fine-grained identification.To address this problem,an IoT device recognition method based on relation extraction is proposed.A mapping relationship is established be⁃tween the head entity,relation,and tail entity in the brand,type,and model information and relation extraction of IoT devices by this method.The IoT device recognition problem is transformed into a rela⁃tion extraction problem.As a resut,the problem of keyword missing in slogan is solved,and the rotat⁃ing position coding information is incorporated into the feature extraction,the long text processing abil⁃ity of IoT device slogan is enhanced,and the fine-grained recognition performance of IoT device is im⁃proved.The experimental results show that the precision of the proposed method is 92.5%and the F1 value is 88.8%,which are better than these of the baseline method.