A General Framework for Data Uncertainty and Quality Classification

Vanessa Simard1, Mikael Rönnqvist2, Luc Lebel2, Nadia Lehoux2

  • 1FORAC - Université Laval
  • 2Université Laval

Details

11:22 - 11:44 | Wed 28 Aug | 103 | WeAT14.2

Session: Interoperability and Information Management

Abstract

It is often assumed that data used to plan operations and supply chain activities is accurate. But in the presence of uncertainty, this assumption is known not to be entirely true. In this context, it becomes relevant to evaluate if a planning decision is appropriate in light of partially accurate data. This paper proposes a general framework for data analysis in order to provide a quality evaluation of the information used in the decision-making process. To this end we propose a process to quantify data quality by comparing “measured” data to “real” data. We use a hybrid approach combining multiple data quality assessment techniques as well as different alternative sources of historic data. A classification phase then rates and «tags» data for proper consideration for decision-making. Such classification provides insights into the level of uncertainty associated with the data. This paper demonstrates the approach developed using a case study from the forest sector. The approach can be adapted to other industrial sectors.