An enhanced concept based approach medical information retrieval to address readability, vocabulary and presentation issues

Kontagora, Ibrahim Umar (2019) An enhanced concept based approach medical information retrieval to address readability, vocabulary and presentation issues. Doctoral thesis, Universiti Tun Hussein Onn Malaysia.

Preview	Text 24p IBRAHIM UMAR KONTAGORA.pdf Download (2MB) | Preview
	Text (Copyright Declaration) IBRAHIM UMAR KONTAGORA COPYRIGHT DECLARATION.pdf Restricted to Repository staff only Download (2MB) | Request a copy
	Text (Full Text) IBRAHIM UMAR KONTAGORA WATERMARK.pdf Restricted to Registered users only Download (3MB) | Request a copy

Abstract

Querying of health information retrieval for health advice has now become a general and notable task performed by individuals on the Internet. However, the failure of the existing approaches to integrate program modules that would address the information needs of all categories of end-users remains. This study focused on proposing an improved framework and designing an enhanced concept based approach (ECBA) for medical information retrieval that would better address readability, vocabulary mismatched and presentation issues by generating medical discharge documents and medical search queries results in both medical expert and layman’s forms. Three special program modules were designed and integrated in the enhanced concept based approach namely: medical terms control module, vocabulary controlled module and readability module to specifically address the information needs of both medical experts and laymen end-users. Eight benched marked datasets namely: Medline, UMLS, MeSH, Metamap, Metathesaurus, Diagnosia 7, Khresmoi Project 6 and Genetic Home Reference were used in validating the systems performance. Additionally, the ECBA was compared using three existing approaches such as concept based approach (CBA), query likelihood model (QLM) and latent semantic indexing (LSI). The evaluation was conducted using the performance and statistical metrics: P@40, NDCG@40, MAP, Analysis of Variance (ANOVA) and Turkey HSD Tests. The outcome of the final experimental results obtained shows that, the ECBA consistently obtained above 93% accuracy rate results on Medline, UMLS and MeSH Datasets, 92% on Metamap, Metathesaurus and Diagnosia 7 datasets and 91% on Khresmoi Project 6 and Genetic Home Reference datasets. Also, the statistical analysis performance results obtained by each of the four approaches: ECBA, CBA, QLM and LSI shows that, there is a significant difference among their Mean Scores, hence, the null hypothesis of no significant difference was rejected.

Item Type:	Thesis (Doctoral)
Subjects:	Q Science > QA Mathematics > QA76 Computer software
Divisions:	Faculty of Computer Science and Information Technology > Department of Software Engineering
Depositing User:	Mrs. Sabarina Che Mat
Date Deposited:	22 Jun 2021 03:45
Last Modified:	22 Jun 2021 03:45
URI:	http://eprints.uthm.edu.my/id/eprint/65

Actions (login required)

View Item