Bayesian Changepoint Test Method for High-Dimensional Data Based on Inspect Projection
Multiple changepoint detection in high-dimensional data has become a hot topic.This paper proposes a Bayesian changepoint detection method based on Inspect projection for high-dimensional data with Gaussian noise.This method utilizes Inspect projection to calculate the optimal projection direction through singular value decomposition(SVD),projects the high-dimensional data onto a one-dimensional space along this direction,and incorporates Bayesian prior information to perform changepoint detection on the dimension-reduced data.Numerical simulations show that this method outperforms the Inspect method under different settings of sample size n,dimension p,and changepoint sparsity k.Finally,this method is applied to the microarray dataset(ACGH)of bladder tumor patients.
万方数据
维普
