Data Mining Techniques Implementation To Improve Healthcare Among Diabetic Patients

dc.Location2016 T 58.6 A43
dc.SupervisorDr Sherief Abdullah
dc.contributor.authorAlBanna, Ghania Aref
dc.date.accessioned2017-03-02T13:17:48Z
dc.date.available2017-03-02T13:17:48Z
dc.date.issued2016-12
dc.description.abstractMedical data mining is an emergent field and, on overcoming its facing challenges such as privacy of documentation and ethical use of information about patients, voluminous and heterogeneous data, and imprecise and erroneous data, medical data mining can be as powerful as that in any other common field such as ecommerce and marketing. Traditional research could not overcome completely these challenges and only hypotheses based on anthropological approaches are tested. Unlike traditional research, this dissertation discusses predictive analysis and knowledge discovery of trends and patterns from databases in the medical field. Retrieval of clinical medical data is helpful in conducting different learning techniques. Performance of different classification techniques is compared and ensemble learning of best classifiers is tested. The analysis showed that ensemble learning via bagging predicts best the percentage of diabetic adolescents who are most prone to hospital readmission and more susceptible to join the “Diabetic Self-Management Educational Support Program”. This predictive classification helps in leveraging the healthy psychological status of the patients (social and medical), reducing readmission costs (economic), and pre-hypothesizing (scientific) relationships between different parameters based on different patterns and trends predicted by machine learning techniques.en_US
dc.identifier.other2014128042
dc.identifier.urihttp://bspace.buid.ac.ae/handle/1234/974
dc.language.isoenen_US
dc.publisherThe British University in Dubai (BUiD)en_US
dc.subjectdata miningen_US
dc.subjecthealthcareen_US
dc.subjectdiabetic patientsen_US
dc.titleData Mining Techniques Implementation To Improve Healthcare Among Diabetic Patientsen_US
dc.typeDissertationen_US
Files
Original bundle
Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
2014128042.pdf
Size:
4.36 MB
Format:
Adobe Portable Document Format
License bundle
Now showing 1 - 1 of 1
Name:
license.txt
Size:
1.71 KB
Format:
Item-specific license agreed upon to submission
Description: