A Machine Learning Model for Prediction of Parkinson’s Disease using Vocal Features

Parkinson’s disease is a progressive neurological disorder which affects motor skills and is characterized by symptoms such as muscle rigidity, tremors and difficulties in maintaining equilibrium and coordination. Detecting Parkinson’s disease involves analyzing the patient’s medical background and conducting multiple imaging tests like MRI, Computed Tomography, and DaTscan, along with physical and neurological assessments. All these steps are essential to validate a definitive diagnosis. Recent researches have proved that speech variations in speech can be considered as a quantifiable indicator to detect Parkinson’s disease in its early stages.This method is both non-invasive and economically efficient in contrast to alternative diagnostic procedures. To achieve a speech-based diagnosis system, we have tried a feature selection algorithm in combination with classification algorithms and a stacking classifier to combine these base models. In the process of feature selection, we have applied the mutual information gain method in combination with three classifiers: K-nearest neighbors(KNN), Support Vector Machine (SVM) and XGBoost. To evaluate various permutations, we’ve made use of the speech dataset provided UCI. To address the issue of imbalanced data, we’ve applied the Synthetic Minority Oversampling Technique (SMOTE). The stacking classifier has shown the best performance, attaining an accuracy rate of 97.05%.