Methodology for Evaluating Similarity in Health Trajectories
Name
Karel Paan
Abstract
The vast digitalization of health data has opened a door towards personalized medicine.By methodically searching for patients with similar health trajectories it is possible topredict diagnoses and prevent diseases or alleviate their effect. Prior knowledge frombioinformatics can be leveraged because the data shares a similar structure and problemswith medical data. This thesis aims to find the best method for identifying similar healthtrajectories inside a certain set of patients. To test the method a script is made that usesinput generated from patients’ health data and returns clusters with similar trajectories.The generator’s signal is used as an indicator of effectiveness of the method. At the endof the thesis, the implemented method is described with ideas on improving it.
Graduation Thesis language
Estonian
Graduation Thesis type
Bachelor - Computer Science
Supervisor(s)
Jaak Vilo
Defence year
2021