本文關鍵詞： 多標簽學習 特征降維 主成分分析 非負矩陣分解 相似矩陣　出處：《閩南師范大學》2017年碩士論文　論文類型：學位論文

【摘要】：在多標簽學習中,多標簽數(shù)據(jù)的每個樣本含有多個標簽,標簽與標簽之間也不是獨立存在的。多標簽數(shù)據(jù)的維數(shù)較高,增加了數(shù)據(jù)挖掘的復雜度和難度。近些年來如何高效地處理多標簽數(shù)據(jù),成為研究者們研究的一個熱點問題。特征降維能降低多標簽數(shù)據(jù)的維度、縮小數(shù)據(jù)規(guī)模,提高多標簽學習的性能。本論文提出了兩種多標簽學習特征降維算法:(1)基于主成分分析的多標簽學習特征降維算法(MLFR-PCA)。首先該算法利用PCA原理將原始數(shù)據(jù)投影到低維空間,對數(shù)據(jù)進行密集和去噪處理。其次算法將數(shù)據(jù)的所有標簽作為一個整體,在標簽與特征之間引入稀疏回歸,建立起標簽空間與特征空間的聯(lián)系,以此構(gòu)造數(shù)據(jù)降維的目標函數(shù)。然后結(jié)合2,1l范數(shù)對算法進行優(yōu)化處理,最終實現(xiàn)降低多標簽數(shù)據(jù)維數(shù)的目的。(2)基于非負矩陣分解的多標簽學習特征降維算法(MLFR-NMF)。首先該算法用特征矩陣與非負矩陣的乘積構(gòu)建特征空間的相似矩陣。其次將數(shù)據(jù)的所有標簽作為一個整體,利用已有方法構(gòu)造標簽空間的相似矩陣。然后在特征空間的相似矩陣與標簽空間的相似矩陣之間引入最小二乘法,建立起標簽空間與特征空間的聯(lián)系,以此構(gòu)造數(shù)據(jù)降維的目標函數(shù)。最后結(jié)合2l范數(shù)對算法進行優(yōu)化處理,以實現(xiàn)降低多標簽數(shù)據(jù)維數(shù)的目的。以上兩種特征降維算法可以直接對多標簽數(shù)據(jù)進行降維,不需要轉(zhuǎn)化多標簽數(shù)據(jù)為單標簽數(shù)據(jù),這樣不僅減少了轉(zhuǎn)化過程引起的工作量增大問題,也避免了因轉(zhuǎn)化不準確帶來的后續(xù)問題。此外,算法將數(shù)據(jù)的所有標簽作為一個整體參與目標函數(shù)構(gòu)造,這樣可以在不破壞標簽結(jié)構(gòu)的情況下,有效利用標簽信息實現(xiàn)降維。通過在真實數(shù)據(jù)集上的實驗,表明了兩種算法效果良好。
[Abstract]:In multi-label learning, each sample of multi-label data contains multiple tags, and the labels and tags do not exist independently. The dimension of multi-label data is higher. In recent years, how to deal with multi-label data efficiently has become a hot issue for researchers. Feature dimensionality reduction can reduce the dimension of multi-label data and reduce the scale of data. In this paper, we propose two multi-label learning feature reduction algorithms: (1) Multi-label learning feature reduction algorithm based on principal component analysis (PCA) and MLFR-PCAA algorithm. Firstly, this algorithm uses PCA principle to project raw data into low-dimensional space. Secondly, the algorithm takes all labels of data as a whole, introduces sparse regression between labels and features, and establishes the relationship between label space and feature space. The objective function of data dimension reduction is constructed, and the algorithm is optimized with 2L norm. Finally, the purpose of reducing the dimension of multi-label data is realized.) the multi-label learning feature reduction algorithm based on non-negative matrix factorization is proposed. Firstly, the product of feature matrix and non-negative matrix is used to construct the similarity matrix of feature space. Take all the labels of the data as a whole, The similarity matrix of the tag space is constructed by using the existing methods, and then the least square method is introduced between the similarity matrix of the feature space and the similarity matrix of the label space, and the relation between the tag space and the feature space is established. Finally, the algorithm is optimized with 2l norm to reduce the dimension of multi-label data. The above two feature dimensionality reduction algorithms can directly reduce the dimension of multi-label data. There is no need to convert multi-label data to single-label data, which not only reduces the increased workload caused by the conversion process, but also avoids the subsequent problems caused by inaccurate transformation. The algorithm constructs all the tags of the data as a whole to participate in the objective function, which can effectively use tag information to reduce the dimension without breaking the tag structure. The results show that the two algorithms are effective.
【學位授予單位】：閩南師范大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：TP181

【參考文獻】

相關期刊論文 前3條

1 劉松;張德賢;;基于權(quán)重差異和類別關聯(lián)的互信息改進研究[J];計算機應用研究;2014年07期

2 張翔;鄧趙紅;王士同;蔡及時;;極大熵Relief特征加權(quán)[J];計算機研究與發(fā)展;2011年06期

3 龔建興;韓超;邱曉剛;黃柯棣;;構(gòu)建可擴展的HLA聯(lián)邦成員架構(gòu)[J];系統(tǒng)仿真學報;2006年11期

相關博士學位論文 前2條

1 陳自潔;多標簽分類問題的圖結(jié)構(gòu)描述及若干學習算法的研究[D];華南理工大學;2015年

2 朱林;基于特征加權(quán)與特征選擇的數(shù)據(jù)挖掘算法研究[D];上海交通大學;2013年

相關碩士學位論文 前4條

1 李玲;多標簽分類中特征選擇算法研究[D];浙江師范大學;2015年

2 王文峰;K-L變換的研究及其在圖像壓縮編碼中的應用[D];沈陽理工大學;2008年

3 翟亞利;非負矩陣分解及其在中文文本挖掘中的應用[D];國防科學技術大學;2007年

4 蘇映雪;特征選擇算法研究[D];國防科學技術大學;2006年

，

本文編號：1504808