坐标下降法
- 网络coordinate descent;Coordinate Descent Method
坐标下降法-
提出了一种求解无约束优化问题的并行坐标下降法，并给出了该并行算法的数值实验结果，进而讨论了该算法的加速倍数及并行效率。
A parallel coordinate descent method for solving unconstrained optimization problems is proposed . The result of numerical experiment is given , an acceleration multiple and parallel efficiency of the algorithm are discussed .
-
将低层次的权重直方图特征和高层次的动作语义属性融合，采用隐支持向量机结合坐标下降法替代单一支持向量机分类方法求解最终动作识别模型的局部最优解。
The low-level weight histogram and high-level semantic attributes are fused together and the latent Support Vector Machine ( SVM ) combined with coordinate descent is adopted to find the local optimum of the prediction model instead of the single SVM classifier .