【摘要】：在物聯(lián)網(wǎng)多設(shè)備通信過程中,其差異化的數(shù)據(jù)在分類種類上存在較大的問題,導(dǎo)致識別過程存在異常數(shù)據(jù)不收斂或者無窮計算的問題。為了解決這一問題,提出基于多約束標(biāo)簽的異常數(shù)據(jù)檢測方法,在對海量的異常數(shù)據(jù)進行分類的過程中,引入可以約束標(biāo)簽異常特征的多個約束條件,對物聯(lián)網(wǎng)通信數(shù)據(jù)進行降維限制處理,避免擴大化的搜索過程,再運用支持向量機在限制區(qū)域內(nèi)完成檢測分類。實驗結(jié)果表明,利用該算法能對海量的物聯(lián)網(wǎng)通信異常數(shù)據(jù)進行自動學(xué)習(xí)過程的搜索,提高異常數(shù)據(jù)檢測的準(zhǔn)確性。
[Abstract]:In the process of multidevice communication in the Internet of things, there is a big problem in the classification of the differentiated data, which leads to the problem of abnormal data convergence or infinite computation in the identification process. In order to solve this problem, a method of anomaly data detection based on multi-constraint tags is proposed. In the process of classifying large amounts of abnormal data, several constraints that can constrain the abnormal features of labels are introduced. In order to avoid the expansion of the search process, support vector machine (SVM) is used to complete the detection and classification in the restricted area by reducing the dimension of the communication data of the Internet of things. The experimental results show that the algorithm can automatically search the massive abnormal data of Internet of things communication and improve the accuracy of anomaly data detection.
【作者單位】： 宿遷學(xué)院信息工程學(xué)院;江蘇大學(xué)計算機科學(xué)與通信工程學(xué)院;
【基金】：宿遷市科技計劃項目(Z201445,S201410,Z201448) 宿遷學(xué)院科研基金項目(2013KY13)
【分類號】：TP391.44;TN929.5

【相似文獻】

相關(guān)期刊論文 前1條

1 張棟;于寧莉;許愛華;;箭載GPS異常數(shù)據(jù)檢測與剔除[J];載人航天;2013年05期

，

本文編號：2421106