Data Quality Indicators Composition and Calculus: Engineering and Information Systems Approaches

Big Data phenomenon is a result of novel technological developments in sensor, computer and communication technologies. Nowadays more and more data are produced by nanoscale photonic, optoelectronic and electronic devices. However, their quality characteristics could be very low. The paper proposes...

Full description

Saved in:
Bibliographic Details
Published in:Sensors & transducers 2015-02, Vol.185 (2), p.140-140
Main Authors: Reznik, Leon, Lyshevski, Sergey Edward
Format: Article
Language:English
Subjects:
Accuracy
Automation
Calculus
Consistency
Data processing
Indicators
Information systems
Mathematical analysis
Network security
Optoelectronic devices
Sensors
Transducers
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Big Data phenomenon is a result of novel technological developments in sensor, computer and communication technologies. Nowadays more and more data are produced by nanoscale photonic, optoelectronic and electronic devices. However, their quality characteristics could be very low. The paper proposes new methods of the data management with huge data amounts that is based on associating of data quality indicators with each data entity. To achieve this goal, one needs to define the composition of the data quality indicators and to develop their integration calculus. As data quality evaluation involves multi-disciplinary research, various metrics have been investigated. The paper describes two major approaches in assigning the data quality indicators and developing their integration calculus. The information systems approach employs traditional high-level metrics like data accuracy, consistency and completeness. The engineering approach utilizes signal characteristics processed with the probability based calculus. The data quality metrics composition and calculus are discussed. The tools developed to automate the metrics selection and calculus procedures are presented. The user-friendly interface examples are provided.
ISSN:2306-8515
1726-5479