loggle v1.0

0

0th

Percentile

Local Group Graphical Lasso Estimation

Provides a set of methods that learn time-varying graphical models based on data measured over a temporal grid. The underlying statistical model is motivated by the needs to describe and understand evolving interacting relationships among a set of random variables in many real applications, for instance the study of how stocks interact with each other and how such interactions change over time. The time-varying graphical models are estimated under the assumption that the graph topology changes gradually over time. For more details on estimating time-varying graphical models, please refer to: Yang, J. & Peng, J. (2018) <arXiv:1804.03811>.

LOGGLE (LOcal Group Graphical Lasso Estimation)

Description

The R package loggle provides a set of methods that learn time-varying graphical models based on data measured over a temporal grid. loggle is motivated by the needs to describe and understand evolving interacting relationships among a set of random variables in many real applications, for instance, the gene regulatory networks over the course of organismal development, and the dynamic relationships between individuals in a community over a few years. loggle estimates time-varying graphical models under the assumption that the graph topology changes gradually over time.

loggle has been applied to S&P 500 stock price dataset, where the interacting relationships among stocks and among industrial sectors in a time period that covers the recent global financial crisis can be revealed. Detailed description of S&P 500 stock price dataset is in ?stockdata.

For more details on estimating time-varying graphical models and the package, please refer to: Estimating Time-Varying Graphical Models https://arxiv.org/abs/1804.03811.

Dependencies

Please make sure to install the following package dependencies before using R package loggle. R with version later than 3.0.2 is needed.

install.packages(c("Matrix", "doParallel", "igraph", "glasso", "sm"))


Installation

The R package loggle can be installed from source files in the GitHub repository (R package devtools is needed):

library(devtools)
install_github(repo="jlyang1990/loggle")


Main functions

• loggle: learn time-varying graphical models for a given set of tuning parameters.
• loggle.cv: conduct model selection via cross validation for learning time-varying graphical models.
• loggle.cv.select: conduct model selection for time-varying graphical models based on cross validation results from loggle.cv.
• loggle.cv.vote: learn time-varying graphical models for a given set of tuning parameters via cv.vote.
• loggle.refit: conduct model refitting given learned time-varying graph structures.

Contact

Please report any bugs to jlyang@ucdavis.edu.

Functions in loggle

 Name Description loggle.refit A function to conduct model refitting given learned graph structures example_source Source code for generating time-varying graphs loggle A function to learn time-varying graphical models loggle.cv A function to learn time-varying graphical models via cross validation loggle.cv.h A function to learn time-varying graphical models via cross validation (with h fixed) loggle.cv.select A function to conduct model selection based on cross validation results loggle.cv.vote A function to learn time-varying graphical models using cv.vote loggle.cv.select.h A function to conduct model selection based on cross validation results (with h fixed) stockdata_source Source code for stock prices of S&P 500 companies from 2007 to 2016 stockdata Stock prices of S&P 500 companies from 2007 to 2016 example An example dataset of time-varying graphs No Results!