Functional Graphical Models via Neighborhood Selection Approach
This folder contains codes running multiple methods, namely FPCA-gX, FPCA-gY, FGLasso, PSKL, and FPCA-PSKL (for Model D).
Threshold is set to
This folder contains codes running FPCA-gX method.
Threshold varies under each setting. Penalty parameter
This folder contains codes running FPCA-gX method.
Threshold varies under each setting. Penalty parameter
This folder contains codes running FPCA-gX method. Two datasets: ABIDE for autism and ADHD. Both raw data and time series extracted from raw data can be found in Data folder. Time series are extracted by "ABIDE.read.data.R" and "ADHD.read.data.R" respectively.
"SCV.autism.R" selects and outputs the optimal adjacency matrices using SCV under FPCA-gX method. "spa.ctrl.autism.R" outputs an adjacency matrix with 2% connection density under FPCA-gX method.
"SCV.ADHD.R" selects and outputs the optimal adjacency matrices using SCV under FPCA-gX method. "spa.ctrl.ADHD.R" outputs an adjacency matrix with 2% connection density under FPCA-gX method. "spa.ctrl.ADHD.FGLasso.R" outputs an adjacency matrix estimated by FGLasso method.