现代计算机：下半月版
現代計算機：下半月版
현대계산궤：하반월판
Modem Computer
2012年
21期
8-11
,共4页
缺失值%缺失值填充%kNNI算法%MkNNI算法
缺失值%缺失值填充%kNNI算法%MkNNI算法
결실치%결실치전충%kNNI산법%MkNNI산법
Missing Data%Missing Data Imputation%kNNI Method%MkNNI Method
针对k最近邻填充算法（kNNI）在缺失数据的k个最近邻可能存在噪声，提出一种新的缺失值填充算法——相互k最近邻填充算法MkNNI（Mutualk-NearestNeighborImputa—tion）。用于填充缺失值的数据，不仅是缺失数据的k最近邻，而且它的k最近邻也包含该缺失数据．从而有效地防止kNNI算法选取的k个最近邻点可能存在噪声这一情况。实验结果表明．MkNNI算法的填充准确性总体上要优于kNNI算法。
針對k最近鄰填充算法（kNNI）在缺失數據的k箇最近鄰可能存在譟聲，提齣一種新的缺失值填充算法——相互k最近鄰填充算法MkNNI（Mutualk-NearestNeighborImputa—tion）。用于填充缺失值的數據，不僅是缺失數據的k最近鄰，而且它的k最近鄰也包含該缺失數據．從而有效地防止kNNI算法選取的k箇最近鄰點可能存在譟聲這一情況。實驗結果錶明．MkNNI算法的填充準確性總體上要優于kNNI算法。
침대k최근린전충산법（kNNI）재결실수거적k개최근린가능존재조성，제출일충신적결실치전충산법——상호k최근린전충산법MkNNI（Mutualk-NearestNeighborImputa—tion）。용우전충결실치적수거，불부시결실수거적k최근린，이차타적k최근린야포함해결실수거．종이유효지방지kNNI산법선취적k개최근린점가능존재조성저일정황。실험결과표명．MkNNI산법적전충준학성총체상요우우kNNI산법。
As there may be noise in choosing the k nearest neighbors of missing data using the k-nearest neighbor imputation (kNNI) algorithm, puts forward a new imputation method, Mutual k-Nearest Neighbor Imputation (MkNNI), for missing values. Data for imputing missing value are the k nearest neighbors of the missing data and at the same time, their k nearest neighbors include the missing data. Thus, noise is effectively prevented in selecting nearest neighbors. We experimen- tally evaluate our approach and demonstrate that MkNNI is much better than the kNNI method in imputed accuracy ~enerallv.
