Publishing Persian linked data; challenges and lessons learned

Abstract

This paper discusses the challenges of publishing Persian linked data based on an experience of publishing some academic data from Ferdowsi University of Mashhad dataset. By analyzing the experimental results of the project and classifying the problems, some publisher-oriented solutions are proposed to improve the quality of datasets on the web.

Quality Metrics for Linked Open Data

“…To this end, our approach is based on several significant quality issues identified in already published datasets on LOD. To the extent of our experience in publishing and interlinking academic data [3,4], we found that many of the published datasets suffer from quality issues. We believe that most of these issues have roots in the deficiencies of the sources from where that data is extracted, and they can be avoided if they are identified in the initial stages of publishing.…”

Data accuracy: What does it mean to LOD?

“…It is computed as: (3) This metric is defined to measure the misspelling errors of the values of data type properties. To this end, we have used Lucene spell checker [22] in our implementation.…”

“…Although data quality is an important issue for the successful organic growth of LOD, there are only a very limited number of research initiatives that focus on data quality specifically for LOD. Based on our practical experience in publishing linked data [3] we have observed that many of the published datasets suffer from quality issues such as syntax errors, redundant instances, and incorrect/uncompleted attribute values. We believe that the assessment of data quality before publishing to the LOD cloud can help publishers filter out low-quality data based on the quality assessment results.…”

“…In order to study data quality in depth, it is necessary to study each quality factor separately as well as the properties of the environment that affect it. Since, most of the quality problems of dataset, studied in our experiment [3]were caused by the inaccuracy of their sources; this work focuses on the data accuracy before they are published.…”

Extending FUM-LD framework by including an academic data model

“…In [4] challenges of publishing Persian linked data are discussed then by analyzing the experimental results of the project and classifying the problems, some publisheroriented solutions are proposed.…”