Automated Meta-Actions Discovery for Personalized Medical Treatments

1 online resource (120 pages) : PDF

2014

University of North Carolina at Charlotte

Healthcare, among other domains, provides an attractive ground of work for knowledge discovery researchers. There exist several branches of health informatics and health data-mining from which we find actionable knowledge discovery is underserved. Actionable knowledge is best represented by patterns of structured actions that inform decision makers about actions to take rather than providing static information that may or may not hint to actions.The Action rules model is a good example of active structured action patterns that informs us about the actions to perform to reach a desired outcome. It is augmented by the meta-actions model that represents passive structured effects triggered by the application of an action. In this dissertation, we focus primarily on the meta-actions model that can be mapped to medical treatments and their effects in the healthcare arena. Our core contribution lies in structuring meta-actions and their effects (positive, neutral, negative, and side effects) along with mining techniques and evaluation metrics for meta-action effects. In addition to the mining techniques for treatment effects, this dissertation provides analysis and prediction of side effects, personalized action rules, alternatives for treatments with negative outcomes, evaluation for treatments success, and personalized recommendations for treatments. We used the tinnitus handicap dataset and the Healthcare Cost and Utilization Project (HCUP) Florida State Inpatient Databases (SID 2010) to validate our work. The results show the efficiency of our methods.

doctoral dissertations

Computer science
Health services administration

Ph.D.

Actionable Rules
Action Set
Data-Mining
Medical Treatment Patterns
Meta-Actions
Side-Effects

Information Technology

Ras, Zbigniew

Studnicki, James
Zadrozny, Wlodek
Yang, Jing
Kazemi, Mohammad

Thesis (Ph.D.)--University of North Carolina at Charlotte, 2014.

This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s). For additional information, see http://rightsstatements.org/page/InC/1.0/.

Copyright is held by the author unless otherwise indicated.

Touati_uncc_0694D_10581

http://hdl.handle.net/20.500.13093/etd:1361