Title: Knowledge extraction based on linked open data for clinical documentation

Authors: Mazen Alobaidi; Khalid Mahmood; Susan Sabra

Addresses: Department of Computer Science and Engineering, Oakland University, Rochester, MI, USA ' Department of Computer Science and Engineering, Oakland University, Rochester, MI, USA ' Department of Computer Science and Engineering, Oakland University, Rochester, MI, USA

Abstract: Smart cities are becoming a reality in the near future to transform many sectors and activities in our lives. Smart cities systems such as healthcare systems will have new functionality to improve the quality of life. Electronic health records are an essential component of healthcare systems. They are valuable for medical research, but the information is recorded as unstructured free text. Knowledge extraction (KE) from unstructured text in electronic health records is a problem but still not totally resolved. KE is very challenging because medical language has ungrammatical and fragmented constructions. We have implemented a unique framework KE based on linked open data for clinical documentation (KE-LODC) that generates accurate and high quality triples transforming unstructured text from clinical documentation into well-defined and ready-to-use linked open data for diagnosis and treatment. Our framework proved to produce a large number of highly qualified triple candidates which improves the likelihood of better classification.

Keywords: linked open data; LOD; semantic web; SPARQL; Swoogle; knowledge extraction; clinical documentation.

DOI: 10.1504/IJSPM.2018.091697

International Journal of Simulation and Process Modelling, 2018 Vol.13 No.2, pp.116 - 125

Available online: 02 May 2018 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article