A quantitative approach for modelling the influence of currency of information on decision-making under uncertainty

Abstract

Stored information, used to support decision-making, can be outdated. Existing metrics for currency provide an indication about the correspondence between this stored information and its real-world counterpart. In the case of low currency, this information cannot be effectively used to support decision-making, although the decision-maker can probably learn from it. Our first objective is to develop an extended metric referring to currency, which provides an indication about the real-world information at the time of measurement, based on the stored information. Thus, the decision can be adjusted and the value of the stored information increased. Therefore, as a second objective, we propose a quantitative approach for modelling the influence of currency on decisionmaking by extending the normative concept of the value of information. Finally, we demonstrate the relevance of our approach by applying it to two real-world scenarios from the field of sales management in customer relationship management.

Potential Problem Data Tagging: Augmenting information systems with the capability to deal with inaccuracies

“…Markov models have been used to model the probabilities that attribute values change, and can estimate when data will become inaccurate (Wechsler and Even, 2012;Zak and Even, 2017). Various useful currency metrics have been developed using probability theory (Heinrich et al, 2009;Heinrich and Hristova, 2016;Klier, 2011, 2015).…”

“…The existing currency models rely on being able to estimate the transition probabilities for data value changes; hence, they are ideally suited to examples where these can be estimated, such as marital status (Heinrich and Hristova, 2016), or the changing status of a process (Zak and Even, 2017). However, contrast this to the quantity of items at a warehouse location transitioning from 4 to 5 items, and what the associated probability is.…”

Requirements for Data Quality Metrics

“…2 both under consideration of the additional information, which means, ( 2 , , 2 , ) − ( 1 , , 1 , ) [for details cf. Heinrich and Hristova 2016]. Thereby, the application of the data quality metric is economically efficient and therefore justifiable with respect to the decision-oriented framework (cf.…”

“…Each data quality dimension provides a par-ticular perspective on the quality of data views. As a result, researchers have developed corresponding metrics for the quantitative assessment of these dimensions for data views [e.g., Ballou et al 1998;Hinrichs 2002;Even and Shankaranarayanan 2007;Heinrich et al 2007;Fisher et al 2009;Heinrich et al 2009;Blake and Mangiameli 2011;Heinrich et al 2012;Wechsler and Even 2012;Heinrich and Klier 2015;Heinrich and Hristova 2016]. Metrics assessing such data quality dimensions for data views and data values stored in IS are in the focus of this paper.…”

Assessing data quality – A probability-based metric for semantic consistency

“…From a methodological and decisionoriented perspective, both benefits can be estimated by comparing the effects resulting from decisions with respectively without considering the metric results and the improved data quality (for a detailed discussion cf. Heinrich & Hristova, 2016;Heinrich, Hristova, Klier, Schiller, & Szubartowicz, 2018).…”