• :

    Whenever decisions under uncertainty are to be made via data analysis, there is a quest for more data to reduce the uncertainty. Today's technology advancement is making such information available in massive volume and constant stream, everywhere and anytime. However, the notion of  "the bigger the better" in today's data environment is not true nor realistic. In the era of data explosion, there is no commensurate growth in human cognitive abilities or physical resources to leverage the data completely. Decision makers are constantly faced with the challenges in extracting right information from ever-growing sources of data. Data acquisition, management, and processing are often costly with limited resources. Therefore, there is a need to evaluate and understand whether the available information is worthwhile beforehand in order to select the most valuable data to maximize efficiency, while ensuring meaningful and robust outcome for the decision makers. Measuring the value of information to distinguish what should and should not be leveraged is not trivial. Although statisticians have been pivotal in the early days of development of value of information theory, a little attention has been provided in the recent decades. Communities from other disciplines have been providing problem-specific, alternative formulations to the traditional statistical methodologies. There have been attempts to provide the elegance of quantifying value of information for dealing with a wide variety of problems in a unified manner. The goal of this paper is to highlight the value of information quantification, inform the practical issues and applications, and discuss a potential unifying framework.

  • : Jade Freeman
  • : Army Research Laboratory
  • : Jade Freeman
  • : statistics
  • : introductory/practitioner
  • : jade.l.freeman2.civ@mail.mil
  • : 301-394-4962
Quantifying Value of Information: Framework and Application