An R wrapper around the Inform library
Switch branches/tags
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.

README.md

rinform

Travis-CI Build Status AppVeyor Build status CRAN_Status_Badge codecov

An R wrapper of the Inform v1.0.0 C library for performing information analysis of complex system. As for the Inform library, rinform is structured around the concepts of:

  • discrete empirical probability distributions which form the basis for all of the information-theoretic measures,
  • classic information-theoretic measures built upon empirical distributions (e.g., Shannon entropy, mutual information, cross entropy),
  • advanced measures of information dynamics for time series (e.g., transfer entropy, active information, evidence of integration).

In addition to the core components, rinform also provides a small collection of utilities to deal with time series (e.g., binning of continuous values, black-boxing of time series).

Detailed installation instructions and documentation for the rinform package as well as information about related projects can be found in the online documentation.

If you are using rinform, consider citing the following articles:

  • D.G. Moore, G. Valentini, S.I. Walker, M. Levin. "Inform: Efficient Information-Theoretic Analysis of Collective Behaviors". Frontiers in Robotics & AI. (under review)
  • D.G. Moore, G. Valentini, S.I. Walker, M. Levin. "Inform: A Toolkit for Information-Theoretic Analysis of Complex Systems". In: Proceedings of the 2017 IEEE Symposium Series on Computational Intelligence, Symposium on Artificial Life, 1-8, IEEE Press, 2017. https://doi.org/10.1109/SSCI.2017.8285197

Acknowledgement: This project was supported in part by the grant Emergent computation in collective decision making by the crevice-dwelling rock ant Temnothorax rugatulus provided by the National Science Foundation (NSF grant PHY-1505048).

Installation

The rinform package includes some C code, that is, the sources of the Inform library. You may need some extra tools to install rinform as they are required to compile the Inform source (e.g., Xcode for Mac users, Rtools for Windows users)

Installation from CRAN

To install rinform directly from the CRAN repository you will simply need to type:

install.packages("rinform")

Once installed, you can load rinform prior to use it with:

library(rinform)

Installation from GitHub

To install rinform from its GitHub repository you will need to have installed the devtools package. If you don't have devtools already, you can install it with:

install.packages("devtools")

Load devtools and install the latest stable version of rinform (i.e., master branch):

library(devtools)
install_github("ELIFE-ASU/rinform")

In case you need to use the development version of rinform, you can specify to install from the dev branch:

install_github("ELIFE-ASU/rinform", ref = "dev")

Once installed, you can load rinform prior to use it with:

library(rinform)

Example

Compute the amount of directed information transfered from a process X to a process Y using transfer entropy:

library(rinform)

xs <- c(0, 1, 1, 1, 1, 0, 0, 0, 0)
ys <- c(0, 0, 1, 1, 1, 1, 0, 0, 0)
transfer_entropy(xs, ys, ws = NULL, k = 2)

Examples of each function of the rinform package can be found in the online documentation.

Getting Help

rinform, as its parent library inform, is developed to help anyone interested in applying information-theoretic techniques get things done quickly and painlessly. We cannot do that without your feedback. We host the project's source code and issue tracker on Github. Please create an issue if you find a bug, an error in this documentation, or have a feature you'd like to request. Your contribution will make rinform a better tool for everyone.

If you are interested in contributing to rinform, please contact the developers, and we'll get you up and running!

Resources: Source Repository and Issue Tracker

Copyright and Licensing

Copyright © 2017-2018 Gabriele Valentini and Douglas G. Moore, Arizona State University. Free use of this software is granted under the terms of the MIT License. See the LICENSE file for details.