计算机工程与设计
計算機工程與設計
계산궤공정여설계
COMPUTER ENGINEERING AND DESIGN
2014年
8期
2864-2871
,共8页
杨娜%邱建林%潘阳%卞彩峰%陆鹏程
楊娜%邱建林%潘暘%卞綵峰%陸鵬程
양나%구건림%반양%변채봉%륙붕정
k-means算法%主成分分析%熵值法%聚类有效性函数%玉米良种选育
k-means算法%主成分分析%熵值法%聚類有效性函數%玉米良種選育
k-means산법%주성분분석%적치법%취류유효성함수%옥미량충선육
k-means algorithm%principal component analysis%entropy value method%clustering validity function%maize seed breeding
为降低 k值选取对聚类效果的影响，用改进的聚类有效性函数来选取k值。综合考虑了类内相似性、类间差异性，通过UCI数据库中的几个数据集和k-means算法来验证其正确性，并应用于玉米良种选育。为确保得到的玉米良种集中不混有劣种，对聚类得到的玉米良种集进行主成分分析和熵值法的综合评价。实验结果表明，采用该函数在玉米良种集中得到了优良的玉米种子，验证了其正确性。
為降低 k值選取對聚類效果的影響，用改進的聚類有效性函數來選取k值。綜閤攷慮瞭類內相似性、類間差異性，通過UCI數據庫中的幾箇數據集和k-means算法來驗證其正確性，併應用于玉米良種選育。為確保得到的玉米良種集中不混有劣種，對聚類得到的玉米良種集進行主成分分析和熵值法的綜閤評價。實驗結果錶明，採用該函數在玉米良種集中得到瞭優良的玉米種子，驗證瞭其正確性。
위강저 k치선취대취류효과적영향，용개진적취류유효성함수래선취k치。종합고필료류내상사성、류간차이성，통과UCI수거고중적궤개수거집화k-means산법래험증기정학성，병응용우옥미량충선육。위학보득도적옥미량충집중불혼유렬충，대취류득도적옥미량충집진행주성분분석화적치법적종합평개。실험결과표명，채용해함수재옥미량충집중득도료우량적옥미충자，험증료기정학성。
To reduce the effect of the k value selection on clustering results ,the improved clustering validity function was used to select the k value .The similarity within the class and differences between classes were taken into consideration .The correctness of the function was verified by several datasets of the UCI database and the k-means algorithm .Finally the function was applied in the maize seed breeding .To ensure the fine corn set without mixing with inferior seeds ,the comprehensive evaluation by the principal component analysis and the entropy value method was made .Experiments prove the validity of the improved clustering validity function ,and good corn seeds were gotten in maize seed breeding .
