摘要
An algorithm for unsupervised linear discriminant analysis was presented. Optimal unsupervised discriminant vectors are obtained through maximizing covariance of all samples and minimizing covariance of local k-nearest neighbor samples. The experimental results show our algorithm is effective.
An algorithm for unsupervised linear discriminant analysis was presented. Optimal unsupervised discriminant vectors arc obtained through maximizing covariancc of all samples and minimizing covariancc of local k-nearest neighbor samples. The experimental results show our algorithm is effective.
基金
TheHighTechniqueProgramofChina(No.2001AA135091)andtheNationalNaturalScienceFoundationofChina(No.60275021)
