A machine learning approach to the automatic classification of female uroflowmetry measurements

Baas, S.P.R. (2016) A machine learning approach to the automatic classification of female uroflowmetry measurements.

This is the latest version of this item.

[img]
Preview
PDF
858kB
Abstract:Uroflowmetry is a cheap, simple and noninvasive test that is often first used for patients with possible lower urinary tract dysfunctions. This test is usually automated to a high extend and the result is a graph of the urine voiding speed (ml/s) vs. the time (s). For uroflowmetry measurements obtained from women, these measurements are currently classified subjectively by one or more physicians. The measurements are classified into one of four groups, each indicating a set of underlying dysfunctions. In this research, it is investigated if this classification process can be successfully automated by constructing multiple automatic classification methods. For constructing these classifiers, a dataset of measurements and classifications by hospital staff from the University Medical Center in Utrecht is used. One of the constructed classifiers is an improved questionnaire resulting from former research, the other classification methods are constructed using machine learning methods. All classifiers are evaluated on a set of chosen performance measures. The ultimately chosen classifier is the regression forest classifier, which was shown to have a good overall performance and gives an estimated accuracy of 96.7% for the diagnosis of new patients.
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:31 mathematics, 44 medicine, 54 computer science
Programme:Applied Mathematics BSc (56965)
Link to this item:http://purl.utwente.nl/essays/70570
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page