Date of Award

11-2-2020

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Information Science

First Advisor

Serhan Dagtas

Abstract

Today, more and more systems make real-time decisions based on the data received from sensors. But it is inevitable that sensors will fail, or send bad data – leading to faulty real-time decision-making. Using this poor-quality data in real-time decision-making can lead to deadly consequences. Traditional Data Quality (DQ) foundations call for “fixing” the poor DQ before the data can be used. However, time is often of the essence in a real-time decision-making environment, leaving little or no time to “fix” poor-quality data before it can be used in the decision-making process. The primary objective in real-time decision-making is not to correct the poor-quality data, but rather to make the best decision (within milliseconds, if not microseconds) with the information at hand. This dissertation looks at how poor-quality data can be used to make real-time decisions. This dissertation research presented a framework to identify the reliability of poor-quality data in real-time decision-making and certified the steps outlined in the framework. The framework more than anything opens data quality to this new arena of real-time decision-making where we need to start using data now and then figure out how to fix the poor-quality later. We have to start thinking in terms of, how much poor data quality can a system deal with before the real-time decision made becomes unreliable – rather than throwing up our hands and saying we can’t make real-time decisions because we have poor-quality data.

Download

Share

COinS