GitHub - xkKevin/Compute-H2: H2 is a nonparametric analysis aiming at quantifying the amplitude correlation of a signal Y on a signal X, independently of the type of the relation between the two signals. This technique has been shown to be particularly suitable for the analysis of EEG/stereotactic-EEG signals in the context of epilepsy.

The nonlinear regression coefficient h2 is calculated as follows:

The average of the signals is set to zero. Xi and Yi are the amplitude values of the ith samples of the X and Y signal.
A scattergram is made: the amplitude values of Yare plotted as a function of the corresponding amplitude values of X.
The ordinate is split into equal-sized bins. Within each bin, the average of the y-values is calculated and called Qj for bin j. The midpoint of bin j is called Pj.
The points (Pj, Qj) with (Pj+1, Qj+1) is called gj(x) and the whole piecewise curve is called f(x). The deviation of each sample from the curve is calculated: Yi - f(Xi) for (Xi, Yi). The sum of the square of the deviations is called the unexplained variance. h2 is the variance in all the y-values (called the total variance) minus the unexplained variance, all divided by the total variance.

Name		Name	Last commit message	Last commit date
Latest commit History 8 Commits
FCDataAnalysis		FCDataAnalysis
3.23（初步二分法（测试失败））.ipynb		3.23（初步二分法（测试失败））.ipynb
Butterworth滤波.ipynb		Butterworth滤波.ipynb
Computer H2.ipynb		Computer H2.ipynb
README.md		README.md
data0324.mat		data0324.mat
sz1H2.ipynb		sz1H2.ipynb
优化h2+滤波.ipynb		优化h2+滤波.ipynb
打开mat文件.ipynb		打开mat文件.ipynb
计算H2.ipynb		计算H2.ipynb
采样二分（略有成效）.ipynb		采样二分（略有成效）.ipynb