Distance Metric Learning for Large Margin Nearest Neighbor Classification
Distance Metric Learning for Large Margin
Nearest Neighbor Classification假设Large margin nearest neighbor classification (LMNN)凸优化问题能量分类器(energy-based classifier)
Nearest Neighbor Classification)
假设
首先作者基于了以下的假设:对于输入的任意样本xix_ixi 对于其相邻的K个样本,应该让与其相同类别的样本尽量靠近
原创
2020-11-09 11:46:00 ·
625 阅读 ·
1 评论