The previous version implementes the algorithms proposed in Zhou et al. (2024), which aims to estimate the high-dimensional networks from longitudinal data using Gassian grapical models. Though the overarching goal of the package is the same, i.e., explore the possible associations between high-dimensional data to improve the estimations of networks. this updated version add three important functionalities, which are

1. Estimation of heterogeneous networks for longitudinal data. The models in previous version (i.e., in the 2024 paper) assumed a stationary process for the longitudinal data. This might not be the case in many situations, e.g, the antibody network before and after vaccination, metabolite network before and after the initialization of the treatment for cancer patients. In this version of lglasso, the function lglasso are extended to accommodate such important scenarios.

2. Extension to general clustered data. This extension is motivated by our study of metabolome data in different tissues of mice. We have the metabolome data in colon, ileum, portal blood peripheral blood from same mouse. It has been observed that these data are closely correlated which is not a surprise since they are from same mouse. When it comes to interaction networks, this package extends the algorithms for longitudinal data to such general clustered data and can estimate both general and tissue-wise networks.

3. This version provide function CVlglasso to facilitate the selection of tuning parameter.