Algorithmic Definition of Patient Trajectory Similarity
Name
Sander Tamm
Abstract
The digitalisation of medical data has provided the possibility to compare patients with each other. The objective of this Master's thesis is to define an algorithm, which could measure the similarity between a pair of patients based on their previous disease records. Two diagnose-based algorithms are defined, based on the International Classification of Diseases (ICD). One of them is based on the hierarchy of ICD and the other focuses on the contextual similarities of diagnoses. In addition, diagnose rarity, severeness and whether the diagnose is chronic or acute is considered. These algorithms are used to develop two different patient-based algorithms, which have different benefits and caveats. Features and comparisons of the two algorithms are provided by using synthetic data.
Graduation Thesis language
Estonian
Graduation Thesis type
Master - Data Science
Supervisor(s)
Jaak Vilo
Defence year
2023