Confidence level evaluation of LOD resources on CKAN instances

Abstract

Linked Open Data has been an initiative aimed at offering principles for the interconnection of data through machine-readable structures and knowledge representation schemes. At present, there are platforms that allow consuming LOD resources, being CKAN one of the most relevant on a large community made up of governmental organizations, NGOs, among others. However, the resources consumption lacks minimum criteria to determine their validity such as level of trust, quality, linkage and usability of the data; aspects that require a previous systematic analysis on the set of published data. To support this process of analysis and determination of the mentioned criteria, this paper has as purpose to present a method that allows analyzing the dataset current state obtained from the different instances published in CKAN, with the aim of evaluating the levels of trust that can offer from their sources. Finally, it presents results, conclusions and future work from the use of the tool for the dataset consumption belonging to certain instances ascribed to the CKAN platform.

Keywords

• ckan
• linked open data
• machine learning
• open data
• tensorflow
• visual analytics