hope

说明：  包含了悉尼科技大学徐亦达老师的课程课件，有贝叶斯理论、EM算法、卡尔曼滤波、HMM、粒子滤波、概率论相关知识等(The file includes knowledge about EM, HMM, particle filter, kalman filter and so on.)

计算回测程序，多因子选股模型，单因子测试(Calculation of back - test program, multi - factor stock - picking model, the single factor test)

说明：  一般的最小二乘通常是一次拿到全部的数据， 对所有数据进行统一优化计算得到模型系数。递推最小二乘是以一种递推的方式计算最小二乘，每次使用最新的测量值，来不断更新模型系数。(General least squares usually get all the data at one time, and perform unified optimization calculations on all the data to get the model coefficients. Recursive least squares is to calculate the least squares in a recursive way, each time the latest measured value is used to continuously update the model coefficients.)

Covariance Matrix adaptation ES for optimizing

说明：  最小二乘法寻找最佳函数匹配，用以数值计算，优化曲线(The least square method is used to find the best function matching for numerical calculation and curve optimization.)

谐波分析，加窗fft插值算法，幅值校正；多次谐波；峰值搜寻(Harmonic analysis；fft)

这个帖子中我想讨论的是移动窗口多项式最小二乘拟和平滑方法，粗糙惩罚方法，以及kernel平滑方法。(Posts in this discussion I think are moving window least squares polynomial fitting smoothing method, crude methods of punishment, as well as the kernel smoothing method.)

说明：  利用python进行因子分析,因子分析可以进行降维!(Using Python factor analysis, factor analysis can reduce dimensionality!)

外点牛顿法，十个参数，可以解决带有约束条件的非线性规划问题。(External point Newton, with dozens of parameters that can solve linear programming problems with constraints)

可以用最小二乘法来提取趋势项。在MATLAB中可利用它的函数来计算(Least squares method can be used to extract the trend items. It can be used in MATLAB to calculate the function)