Predication of Parkinson's disease using data mining methods: a comparative analysis of tree, statistical, and...

Area of Science:

• Neurology
• Computer Science
• Data Mining

Background:

• Early prediction of Parkinson's disease (PD) is challenging due to symptom onset in middle to late age.
• Speech articulation difficulties are a key symptom in individuals affected by PD.
• Existing diagnostic methods often identify PD after significant neurological progression.

Purpose of the Study:

• To develop and evaluate predictive models for early Parkinson's disease detection.
• To focus on speech articulation as a primary indicator for PD identification.
• To compare the efficacy of different data mining techniques in classifying PD patients.

Main Methods:

• Utilized three distinct data mining classification methods: tree classifiers, statistical classifiers, and support vector machines.
• Focused analysis on speech articulation difficulty data from individuals with and without PD.
• Assessed classifier performance using accuracy, sensitivity, and specificity metrics.

Main Results:

• The study compared the performance of tree, statistical, and support vector machine classifiers in identifying PD based on speech data.
• Performance metrics including accuracy, sensitivity, and specificity were used to evaluate each model.
• Identified the most accurate data mining model for early detection of Parkinson's disease.

Conclusions:

• Data mining techniques show promise for early Parkinson's disease detection using speech analysis.
• The comparative analysis highlights the strengths of specific classifiers in identifying PD-related speech changes.
• Accurate early identification through speech patterns can potentially improve patient outcomes and treatment strategies.