ImprovedCVactivecontourmodel

这是 Andrew Ng 机器学习讲义《Reinforcement Learning and Control》,包含了离散和连续MDP的内容，比官网的课件全。（官网的缺乏连续MDP部分）(This is Andrew Ng Machine Learning Materials " Reinforcement Learning and Control" , contains a discrete and continuous MDP content than the official website of the courseware whole. (Official website of the lack of continuous MDP section))

说明：  该算法包含极大似然估计的梯度算法和快速ica算法，已经调试成功过，调试时注意所给信号x是行向量还是列向量(The algorithm contains the maximum likelihood estimation of the gradient method and fast ica algorithm, has been debugged, debugging attention given signal x is a row vector or column vector)

k-means 算法接受输入量 k ；然后将n个数据对象划分为 k个聚类以便使得所获得的聚类满足：同一聚类中的对象相似度较高；而不同聚类中的对象相似度较小。聚类相似度是利用各聚类中对象的均值所获得一个“中心对象”（引力中心）来进行计算的。(k-means algorithm accepts input k then n data objects into k clusters in order to make clustering satisfy obtained: the objects in the same cluster high similarity different clustering object similarity small. Cluster similarity is the use of the mean of each cluster of objects that get a " central object" (center of gravity) to perform the calculation.)

some lecture about the extend kalman filter algorithm ,which is very useful for the beginner.if you down load it ,it will be of great convenient for you.

二维交错网格纵横波分离的弹性波正演模拟 matlab仿真(Two-dimensional staggered grid of p-wave elastic wave forward modeling of the separation of the matlab simulation)

说明：  pwm整流电路，即将直流电源逆变为交流电(Rectifying circuit, will be the dc power supply inverter for alternating current)

Mathematic Algorithm for image processing contrast solarisation

to find stastical moments

编程心态管理，自我控制！提高效率。建立目标(Manages itself to move, the formulation goal, the self-surveillance develops the flow )

说明：  L1 Normalization for quadratic regularization