How aastha2104/parkinson-disease-prediction works
This is a research prototype for classifying Parkinson’s disease from speech features, not a deployed diagnostic application. Its main workflow loads a local dataset, optionally rescales features, splits training and validation data, fits classifiers, and reports accuracy and Matthews correlation. Separate scripts compare multiple algorithms, evaluate a Praat-feature variant, and run KNN independently. The README documents broader comparisons across splits and repeated trials; the available excerpts do not show how benchmark.py implements those steps, so its connections are left unverified.
Main entry points: data.csv Speech Features · algorithm_comparison.py Validation Metrics · knn.py KNN Evaluation · algorithm_comparison_praat.py Praat Comparison
Components 12
Speech Data
- Speech Features
data.csv - Parkinsons Data
parkinsons.csv
Model Experiments
- Algorithm Comparison
algorithm_comparison.py - Praat Comparison
algorithm_comparison_praat.py - Scaled Comparison
rescaled_data_algorithm_comparison.py - KNN Evaluation
knn.py - Benchmark Runs
benchmark.py - Feature Scaling
knn.py - Classifier Fitting
algorithm_comparison.py
Evaluation
- Validation Metrics
algorithm_comparison.py - Scores
Other components
- Researcher
Connections 15
- Researcher → Algorithm Comparison: runs experiment
- Speech Features → Algorithm Comparison: loads
algorithm_comparison.py - Speech Features → Praat Comparison: loads
algorithm_comparison_praat.py - Speech Features → Scaled Comparison: loads
rescaled_data_algorithm_comparison.py - Speech Features → KNN Evaluation: loads
knn.py - Speech Features → Feature Scaling: scales
knn.py - Feature Scaling → KNN Evaluation: provides features
knn.py - Algorithm Comparison → Classifier Fitting: fits models
algorithm_comparison.py - Praat Comparison → Validation Metrics: evaluates models
algorithm_comparison_praat.py - Scaled Comparison → Validation Metrics: evaluates models
rescaled_data_algorithm_comparison.py - KNN Evaluation → Validation Metrics: scores predictions
knn.py - Classifier Fitting → Validation Metrics: predicts validation
algorithm_comparison.py - Validation Metrics → Scores: prints scores
algorithm_comparison.py - Parkinsons Data → Benchmark Runs: benchmark input
- Benchmark Runs → Scores: produces results