<address id="fnnj5"><dfn id="fnnj5"></dfn></address>

<address id="fnnj5"><dfn id="fnnj5"></dfn></address>
<sub id="fnnj5"><dfn id="fnnj5"></dfn></sub>
<address id="fnnj5"><dfn id="fnnj5"></dfn></address>

<address id="fnnj5"><listing id="fnnj5"></listing></address>

    云南民族大學學報(自然科學版)

    2020, v.29;No.124(06) 568-576

    [打印本頁] [關閉]
    本期目錄(Current Issue) | 過刊瀏覽(Past Issue) | 高級檢索(Advanced Search)

    一種基于信息熵的加權聚類算法
    A weighted clustering algorithm based on information entropy

    李順勇;崔文秀;荊鵬霏;
    LI Shun-yong;CUI Wen-xiu;JING Peng-fei;School of Mathematical Sciences, Shanxi University;

    摘要(Abstract):

    混合型數據是數值型數據和分類型數據的結合,而真實數據集大部分是混合數據,因此混合型數據聚類問題得到越來越廣泛地關注.主要工作包括:綜合考慮類內熵及類間熵對權重的影響,給屬性賦予新的權重,重新定義了尋找最壞類廣義機制、有效性指標、相異性度量.提出了1種基于信息熵的混合數據加權聚類算法.該算法在5個UCI數據集上比較了5個外部評價指標和1個內部評價指標,其結果均優于與其余兩種算法(~(Liang-)k-prototypes算法,~(Li-)k-prototypes算法).
    Mixed data are the combination of numerical data and categorical data, while the real data set is mostly mixed data, so the problem of mixed data-clustering has received more and more attention. This research focuses on the following: comprehensively analyzing the influence of intra-class entropy and inter-class entropy on weight, giving new weight to attributes, and redefining the generalized mechanism for finding the worst cluster, validity index and dissimilarity measurement. A weighted clustering algorithm for mixed data based on the information entropy is proposed. The algorithm compares the values of 5 external evaluation indexes and 1 internal evaluation index on 6 UCI data sets, and proves to be superior to the other two algorithms(~(Liang)-k-prototypes algorithm,~(Li)-k-prototypes algorithm).

    關鍵詞(KeyWords): k-prototypes算法;混合數據;信息熵;屬性權重;有效性指標
    k-prototypes algorithm;mixed data;information entropy;attribute weights;validity index

    Abstract:

    Keywords:

    基金項目(Foundation): 國家自然科學基金(81803962);; 山西省基礎研究計劃項目(201901D111320);; 山西省研究生教育改革項目(2019JG023);; 山西省回國留學人員科技活動擇優資助項目(2019-13);; 太原市科技計劃研發項目(2018140105000084)

    作者(Author): 李順勇;崔文秀;荊鵬霏;
    LI Shun-yong;CUI Wen-xiu;JING Peng-fei;School of Mathematical Sciences, Shanxi University;

    Email:

    DOI:

    參考文獻(References):

    擴展功能
    本文信息
    服務與反饋
    本文關鍵詞相關文章
    本文作者相關文章
    中國知網
    分享
    新浪爱彩