nmf

将采用训练函数 traingdx 和“提前停止”相结合的方法来训练 BP 网络，以提高 BP 网络的推广能力(Will use the training function traingdx and " stop ahead" a combination of methods to train BP network to improve the generalization ability of BP network)

基于非周期随机共振的实现，求最大相关系数有明显的效果(Implementation based on aperiodic stochastic resonance )

化工上的程序控制 系统：二阶串联回馈热搅拌器，使用2个热交换器分别对2个串联的搅拌槽加热。 控制策略：我们将利用2个控制阀来控制热交换器，以维持T2和T4的温度在理想的控制范围之内 可计算：GP，GV，GC，误差值时，PID 可做动态测试 附上文件操作及说明档(Chemical process control on System: Second-order hot mixer series feedback, the use of two heat exchangers in series, respectively, for two stirred tank heating. Control strategy: We will use the two control valves to control the heat exchanger to maintain the temperature in the T2 and T4 within the scope of the ideal control Calculated value: Gp, Gv, Gc, the error value, PID Dynamic testing can be done Attach file operations and help files)

Solves matrix equation of type AX+XB+XCX+D = 0 for X being non-square

说明：  一阶RC低通滤波器与二阶RC低通滤波器特性分析及其阶跃响应对比(First-order RC low-pass filter with the second-order RC low-pass filter characteristics of its step response comparison)

MATLAB优化模型求解，对于等效法的构建(MATLAB optimization model, for according to build)

二维数据平滑，跟定二维数据，每9个点拟合，中间点带入拟合的曲线，得到平滑后的数据(2 d data smoothing, with 2 d data, fitting every nine points, intermediate points to the fitting curve, get data after smoothing)

NSCT多尺度分解，能够得到多方向的分解向量(The NSCT multi-scale decomposition can be multi-directional decomposition vector)

说明：  涉及AR参数计算，并用实际数据进行预测结果还可以。涉及AR参数计算，并用实际数据进行预测结果还可以。(Involving the AR parameter calculation with actual data, and the forecast results can also be.)

emd经验模态分解，适用于对含噪信号处理，优于小波去噪的信号去噪方法。(Empirical mode decomposition)