Prediction of the success of health therapy using machine learning

Thesis Type Bachelor
Thesis Status
Currently running
Student Samuel Plangger
Thesis Supervisor
External Supervisor
Michael Tschuggnall

In Austria electronic health records have been stored since 2013. This data has proven to be a very useful tool for future patient care.
Especially in the domain of cancer diseases, this data has been extensively used in combination with machine learning algorithms to provide better predictions 
for survivability, recurrence and susceptibility. This procedure hasn't found wide adoption in other medical fields until now.

In this thesis this methodology is applied on a dataset of 2500 patients with bone and joint diseases to make predictions regarding therapy success.
To achieve accurate predictions, the dataset undergoes a cleaning and analysis phase. 
Afterwards, predictions are generated with various machine learning algorithms.