UMAT_Hashin3D

根据电磁参数,介电常数实部和虚部，磁导率的实部和虚部，计算反射损耗。(calculating the reflection loss from the electromagnetic parameters)

奇异值快速分解程序，相比其他奇异值分解程序，速度快，解的精度高。(Fast singular value decomposition process, compared to other singular value decomposition process, fast, high precision solution.)

提供两种非线性动态方程线性化的方法实例，精确线性化方法和反馈线性化方法(Provides two linear equations of nonlinear dynamic methods examples of exact linearization and feedback linearization method)

说明：  有限元法分析土体渗流的源程序，采用Pascal语言，比较好(Source program of finite element analysis of soil seepage)

based classification (SRC) has been widely used for face recognition (FR). SRC first codes a testing sample as a sparse linear combination of all the training samples, and then classifies the testing sample by evaluating which class leads to the minimum representation error. While the importance of sparsity is much emphasized in SRC and many related works, the use of collaborative representation (CR) in SRC is ignored by most literature.

我们需要成百上千的分类器来解决现实世界的分类吗 我们评估179分类17种分类器（判别分析，贝叶斯，神经网络，支持向量机，决策树，基于规则的分类器，升压、装袋、堆放、随机森林和其他合奏，广义线性模型，线性，偏最小二乘法和主成分回归，logistic回归、多项式回归、多元自适应回归样条等方法），实现在WEKA，R（有或没有插入包），C和Matlab，包括所有目前可用的相关分类。(Do-we-Need-Hundreds-of-Classifiers-to-Solve-Real-World-ClassificationProblems)

说明：  matlab程序，有约束最优化问题直接法--复合形法(matlab program, constrained optimization problems are the direct method- the complex method)

幂律分布的数据案例，这个案例数据有助于理解幂律分布的计算公式。(The case of power-law distribution of data, case data will help to understand the power-law distribution formula.)

基于交替方向乘子法的非负矩阵分解算法，主要用于盲分离(Alternating direction multiplier method based on non-negative matrix factorization algorithm, mainly used for blind source separation)

The only information passed from difeq to solvde is the matrix of derivatives