Finger Tapping Clinimetric Score Prediction in Parkinson's Disease Using Low-Cost Accelerometers

<table>A global model was trained based on the 86 observations, each one of them being associated with 18 features as well as one SMD consensus score. Since all the observations were used for the construction of the global model, the predictive performance of the model could not be evaluated on an independent dataset.</table>

Computational Intelligence and Neuroscience

fig4

Figure 4

Figure 4: Finger Tapping Clinimetric Score Prediction in Parkinson's Disease Using Low-Cost Accelerometers 