物联网中非结构化信息特征自动提取方法研究  

Research on automatic feature extraction method of unstructured information in Internet of things

作　　者：林晓农 LIN Xiao-nong(School of Computing and Information Science,Fuzhou Institute of Technology,Fuzhou 350506,Fujian,China)

机构地区：[1]福州理工学院计算与信息科学学院,福建福州350506

出　　处：《贵阳学院学报（自然科学版）》2021年第1期17-21,共5页Journal of Guiyang University：Natural Sciences

摘　　要：由于物联网中非结构化信息占比较大,且信息增速较快,为解决现有方法无法有效处理海量非结构化信息的特征,存在特征提取结果准确性不高与耗时较长的问题,提出物联网中非结构化信息特征自动提取方法。依据非结构化信息描述概念,构建一个四面体结构,实现对非结构化数据的整体描述。采用小波降噪方法去除非结构化信息中的非正常流信息,保证特征提取的效率。在此基础上,运用多特征判别方法实现对有价值特征与无价值特征的判别,从而实现对非结构化信息特征的自动提取。实验结果表明:所提方法能够实现对非结构化信息特征的高效提取,且提取结果准确性较高,所得结果具有可靠性。Due to the large proportion of unstructured information in the Internet of things and the rapid growth of information,in order to solve the problem that the existing methods can not effectively deal with the characteristics of massive unstructured information,the accuracy of feature extraction results is not high and the time-consuming problem is long.This paper proposes an automatic feature extraction method of unstructured information in the Internet of things.According to the concept of unstructured information description,a tetrahedral structure is constructed to realize the overall description of unstructured data.The wavelet denoising method is used to remove the abnormal flow information from the unstructured information to ensure the efficiency of feature extraction.On this basis,the multi feature discrimination method is used to distinguish the valuable features from the non valuable features,so as to realize the automatic extraction of the unstructured information features.Experimental results show that the proposed method can effectively extract the features of unstructured information,and the extraction results are accurate and reliable.

关 键 词：物联网 非结构化信息 特征提取 小波降噪 多特征判别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]