基于ReliefF-DDC特征選擇算法的非侵入式負(fù)荷識(shí)別研究
2021年電子技術(shù)應(yīng)用第7期
邵 琪1,包永強(qiáng)2,姜家輝1,張旭旭1
1.南京工程學(xué)院 電力工程學(xué)院,江蘇 南京211167;2.南京工程學(xué)院 信息與通信工程學(xué)院,江蘇 南京211167
摘要: 提取有效的負(fù)荷運(yùn)行數(shù)據(jù)特征對(duì)于提高非侵入式負(fù)荷識(shí)別的精度具有重要作用。針對(duì)數(shù)據(jù)特征選擇欠佳導(dǎo)致負(fù)荷識(shí)別準(zhǔn)確率不高的問題,提出了一種基于ReliefF-DDC特征選擇算法,用于降低特征維數(shù)減少復(fù)雜度,改善負(fù)荷識(shí)別效果。首先,利用ReliefF算法分析各特征與類別的關(guān)系計(jì)算特征權(quán)重,篩選無關(guān)特征;其次,利用DDC算法計(jì)算特征之間與類別的互信息分析相關(guān)性,根據(jù)特征子集評(píng)價(jià)度量刪除冗余特征;最后,采用孿生支持向量機(jī)(TWSVM)作分類器進(jìn)行負(fù)荷識(shí)別。實(shí)驗(yàn)表明,所提出的算法在提升分類效果的同時(shí)減少了運(yùn)行時(shí)間。
中圖分類號(hào): TN911;TM714
文獻(xiàn)標(biāo)識(shí)碼: A
DOI:10.16157/j.issn.0258-7998.200524
中文引用格式: 邵琪,包永強(qiáng),姜家輝,等. 基于ReliefF-DDC特征選擇算法的非侵入式負(fù)荷識(shí)別研究[J].電子技術(shù)應(yīng)用,2021,47(7):74-77,82.
英文引用格式: Shao Qi,Bao Yongqiang,Jiang Jiahui,et al. Research on non-intrusive load identification based on ReliefF-DDC feature selection algorithm[J]. Application of Electronic Technique,2021,47(7):74-77,82.
文獻(xiàn)標(biāo)識(shí)碼: A
DOI:10.16157/j.issn.0258-7998.200524
中文引用格式: 邵琪,包永強(qiáng),姜家輝,等. 基于ReliefF-DDC特征選擇算法的非侵入式負(fù)荷識(shí)別研究[J].電子技術(shù)應(yīng)用,2021,47(7):74-77,82.
英文引用格式: Shao Qi,Bao Yongqiang,Jiang Jiahui,et al. Research on non-intrusive load identification based on ReliefF-DDC feature selection algorithm[J]. Application of Electronic Technique,2021,47(7):74-77,82.
Research on non-intrusive load identification based on ReliefF-DDC feature selection algorithm
Shao Qi1,Bao Yongqiang2,Jiang Jiahui1,Zhang Xuxu1
1.School of Electrical Engineering,Nanjing Institute of Technology,Nanjing 211167,China; 2.School of Information and Communication Engineering,Nanjing Institute of Technology,Nanjing 211167,China
Abstract: Extracting effective characteristics of load operation data plays an important role in improving the accuracy of non-intrusive load identification.In this paper, a ReliefF-DDC feature selection algorithm was proposed to reduce feature dimension, reduce complexity and improve load recognition.Firstly, ReliefF algorithm was used to analyze the relationship between each feature and category, calculate feature weight, and screen irrelevant features.Secondly, DDC algorithm is used to calculate the mutual information analysis correlation between features and categories, and redundant features are removed according to feature subset evaluation measurement. Finally, twin support vector machine(TWSVM) is used as classifier for load recognition. Experiments show that the algorithm proposed in this paper improves the classification effect and reduces the running time.
Key words : ReliefF;DDC;TWSVM; feature selection; load identification
0 引言
非侵入式負(fù)荷監(jiān)測(cè)法(Non-Intrusive Load Monitoring,NILM)為實(shí)現(xiàn)智能電網(wǎng)和用戶之間的互動(dòng)提供了數(shù)據(jù)支持,該方法在接戶線入口處安裝傳感器,采集總負(fù)荷的電壓、電流等電氣量數(shù)據(jù)進(jìn)行分析,細(xì)化系統(tǒng)數(shù)據(jù),從而辨識(shí)家用電器的類別及運(yùn)行狀態(tài)[1]。相比于侵入式負(fù)荷監(jiān)測(cè)法(Intrusive Load Monitoring,ILM),NILM具有成本低、用戶接受度高、后期維護(hù)方便等優(yōu)勢(shì),但是該方法對(duì)于負(fù)荷分解算法的要求較高。特征提取和負(fù)荷識(shí)別作為NILM中兩大關(guān)鍵技術(shù)[2],為NILM的發(fā)展提供了強(qiáng)有力的技術(shù)支持。特征選擇作為處理已提取特征的重要手段,是目前研究的熱點(diǎn)之一。
本文詳細(xì)內(nèi)容請(qǐng)下載:http://ihrv.cn/resource/share/2000003659。
作者信息:
邵 琪1,包永強(qiáng)2,姜家輝1,張旭旭1
(1.南京工程學(xué)院 電力工程學(xué)院,江蘇 南京211167;2.南京工程學(xué)院 信息與通信工程學(xué)院,江蘇 南京211167)
此內(nèi)容為AET網(wǎng)站原創(chuàng),未經(jīng)授權(quán)禁止轉(zhuǎn)載。