|
Abstract:
|
The rapid development of medical imaging technologies, including anatomical magnetic resonance imaging (MRI), functional magnetic resonance imaging (fMRI) and electroencephalography (EEG), allows researchers to predict disease status/clinical outcomes by analyzing these images. Classical regression methods turn the images stored as multi-dimensional arrays into a vector. This process not only damages the structure of the image data but also has been challenged by ultra-high dimensionality. Traditional tensor regression substantially reduces the number of parameters to be estimated and facilitates the capturing of the structure of data via the product-linked parameters. We propose a tensor regression model with functionally linked parameters, which can characterize more complicated structural correlations in images.
|
