“当前”统计模型(Kalman)

有限元平面问题的求解相关程序。 包括质量矩阵，刚度矩阵，主程序，绘图程序等。(Finite element solution of plane problems related procedures. Including the mass matrix, stiffness matrix, the main program, mapping procedures.)

fft的文件，fortran环境下运行！欢迎指正！(fft documents, fortran environment running! Please correct me!)

说明：  对时间序列进行滤波，去掉噪声，得到平滑后的数据(Filters a time series via Lanczos method (cosine filter).)

给定求解区间，用最小二乘法给出函数的零点，并给出迭代次数(Given the solution interval, the zeros of the function are given by the least square method, and the number of iterations is given.)

3D流场计算，程序简单非常容易看懂。适合初学者学习(3D flow field)

说明：  线性判别分析（LDA，全称Fisher Linear Discriminant Analysis）算法的C#实现源码，根据stprtool box for matlab中的LDA.m编写。用到MathNet库做相关的矩阵运算，使用zedgraph控制绘图。有简单的测试数据。( C# source code of linear discriminant analysis (LDA, full name of the Fisher Linear Discriminant Analysis) algorithm , according to LDA.m of stprtool box for matlab . Use MathNet library to do matrix operations,and use zedgraph control for drawing. Thers has a simple test data.)

用MATLAB实现模拟立方体的重构,是计算机视觉中三维重构的重要的应用实验(Simulating cube reconstruction with MATLAB is an important application experiment of 3D reconstruction in computer vision.)

利用坎门站的潮汐观测数据计算调和常数，并利用主要分潮和浅水分潮进行潮汐预报(Kanmen station observation data using tidal harmonic constants calculated, using mainly tidal and shallow water tidal surges were forecast...)

加权总体最小二乘matlab工具箱，包含各种总体最小二乘算法的源代码(weighted total least squares toolbox)

为了提高三级倒立摆系统控制的响应速度和稳定性,在设计Mamdani 型模糊推理规则控制器控制倒立摆系统稳定的基础上, 设计了一种更有效率的基于Sugeno 型模糊推理规则的模糊神经网络控制器。该控制器使用BP 神经网络和最小二乘法的混 合算法进行参数训练,能够准确归纳输入输出量的模糊隶属度函数和模糊逻辑规则。通过与Mamdani 型控制器的仿真对比, 表明该Sugeno 型模糊神经网络控制器对三级倒立摆系统的控制具有良好的稳定性和快速性,以及较高的控制精度。(In order to improve the three-level control of inverted pendulum system response speed and stability, in the design of Mamdani-type fuzzy inference rules of the system controller to control the stability of inverted pendulum on the basis of a more efficient design based on Sugeno-type fuzzy inference rules of fuzzy neural network controller. The controller is the use of BP neural network and hybrid least squares training algorithm parameters can be accurately summed up the amount of input and output fuzzy membership function and fuzzy logic rules. Mamdani-type controller with a simulation comparison shows that the Sugeno-type fuzzy neural network controller for the three-tier control of inverted pendulum system with good stability and fast, as well as a higher control precision.)