The code takes a multiple returns vector and generate covariance martix. It has 4 different GRACH models and it uses the AIC/BIC test to select best fitted model for each serie.
Input:
X: NxM M series of length N of returns
NumFac: the munber of factor to be used in the CPA process. If NumFac<1 the percentage of varince will be used to determin the number of factors
Lag: Lag for the covariance forecast
Samp: The umber of periods from the last period to use to calculate the matrix
OUTPUT:
Cov:Covariance matrix
Corr: correlation matrix
Cov: forecast Covariance
PCorr:forecast correlation
PRt: forecast return
PRterr: forecast return error
NumFac: number of factors used in the PCA
COEFF: is a P-by-P matrix, each column containing coefficients
for one principal component.
SCORE: correspond to observations, columns to components.
ev: eigenvalues of the covariance matrix of X, in
LATENT.
Example: P is a NxM matrix of prices with N periods and M assets.
Rt=tick2ret(P);
[Cov Corr PCov PCorr PRt Model PRterr NumFac COEFF,SCORE,ev]=uni_multi_garch(Rt,0.8,1,N-1)
comments and corrections are welcome. I'm using this to calculate the variance of my portfolios and I have more than 3 years experiance of managing the the volatility of my portfolios using this code.
The code is based in the following paper:
Multivariate GARCH with Only Univariate
Estimation
引用
Tal Shir (2024). MULTIVARIATE GARCH BASED ON PCA (https://www.mathworks.com/matlabcentral/fileexchange/47071-multivariate-garch-based-on-pca), MATLAB Central File Exchange. 取得済み .
MATLAB リリースの互換性
プラットフォームの互換性
Windows macOS Linuxカテゴリ
- AI, Data Science, and Statistics > Statistics and Machine Learning Toolbox > Dimensionality Reduction and Feature Extraction >
- Computational Finance > Econometrics Toolbox > Conditional Variance Models >
- Computational Finance > Econometrics Toolbox > Multivariate Models >
タグ
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!