University of Twente Student Theses

Login

Using statistical learning to predict the probability of a parallel sub-trajectory is getting wrongfully declared at Performation

Grimm, J.H.A. (2023) Using statistical learning to predict the probability of a parallel sub-trajectory is getting wrongfully declared at Performation.

[img] PDF
1MB
Abstract:The aim of this study is to provide a prediction model that can help detect parallel sub-trajectories at risk of getting wrongfully declared, using dashboards and statistical interactions. The required data used to answer the above question is collected per hospital that participated. In total 13 hospitals gave permission to use their data in this research. The models that are suitable for this research are decision trees, logistic regression, and random forest. The most suitable statistical learning method depends on the data that is modeled. For all specialties logistic regression performs best, for ophthalmology this is the decision tree, and for surgery, internal medicine, and the remaining specialties the random forest is best. The advice for Performation is to implement the ophthalmology decision tree, and for the other algorithm import more data to get a higher specificity. The outcome of the model returns a probability of a parallel sub-trajectory is opened correctly. The lower the probability, the higher the chance that this parallel sub-trajectory is wrongfully opened, and will be rejected during checking of sample observations. This probability can be used in the decision making which parallel sub-trajectories should be checked before the declaration.
Item Type:Essay (Master)
Faculty:BMS: Behavioural, Management and Social Sciences
Programme:Industrial Engineering and Management MSc (60029)
Link to this item:https://purl.utwente.nl/essays/94918
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page