ePrints.FRI - University of Ljubljana, Faculty of Computer and Information Science

Predicting Parkinson's Disease with Voice Analysis on a Smartphone

Andrej Zupanc (2017) Predicting Parkinson's Disease with Voice Analysis on a Smartphone. EngD thesis.

Download (718Kb)


    Early diagnosis can have significant effect on disease progression, its treatment and patient's quality of life. However, some diseases are incurable, and doctors can only help relieve symptoms. One of such is Parkinson's disease, a neurodegenerative disease marked by tremor, slowness of movement, muscular rigidity and difficulty with speaking. The aim of this paper was to develop a system for early diagnosis of Parkinson's disease which could recognize signs of Parkinson's disease in a person's voice. For this purpose, a mobile application, an API interface and a classifier were developed. The API interface saves voice recordings made by the mobile application, then analyses and classifies them with the classifier. After the classification is done, the API interface sends the result back to the mobile application which informs its user about the outcome of their voice analysis. The application was developed for Android operating system. The API interface is based on the Flask library. Different classifiers using libraries Scikit-learn and Keras were developed. Then, the most appropriate classifier was chosen and implemented into the API interface. An example of how the application can be used is also described.

    Item Type: Thesis (EngD thesis)
    Keywords: machine learning, Parkinson's disease, mobile application, imbalanced data
    Number of Pages: 43
    Language of Content: Slovenian
    Mentor / Comentors:
    Name and SurnameIDFunction
    akad. prof. dr. Ivan Bratko77Mentor
    Link to COBISS: http://www.cobiss.si/scripts/cobiss?command=search&base=51012&select=(ID=1537381571)
    Institution: University of Ljubljana
    Department: Faculty of Computer and Information Science
    Item ID: 3819
    Date Deposited: 09 Mar 2017 08:41
    Last Modified: 16 Mar 2017 10:16
    URI: http://eprints.fri.uni-lj.si/id/eprint/3819

    Actions (login required)

    View Item