这些方法对选择的贝叶斯网络结 构首先基于梯度的优化(gradient-based optimization)或用 EM(expectation-maximization)算法进行最大后验参数估计, 然后使用拉普拉斯近似(Laplace approximation)或贝叶斯信 息标准(...
基于12个网页-相关网页
To obtain a constant gradient magnetic field outside unilateral nuclear magnetic resonance (NMR), optimal design method based on finite element method (FEM) and topology optimization is proposed.
为了在单边核磁共振仪器外产生梯度恒定的磁场，提出利用基于拓扑优化和有限元法的优化设计方法。
In order to attain a high-resolution image, the algorithm is based on the MAP algorithm, solving the optimization by proposed iteration steps with using the gradient projection method.
该方法基于MAP算法，通过利用梯度投影的方法对重建结果不断进行迭代优化得到最终的理想高分辨率影像。
The gradient optimization algorithm based on the stochastic finite element was adopted to analyze the reliability of the frame with stochastic parameters.
基于随机有限元的梯度优化算法引入拥有随机参数的框架可靠度分析。
应用推荐