Skip to main content

Table 3 Classification results of the set of Q = 36 individuals (18 individuals with the Parkinson’s disease and 18 controls) using two features (stride lengths and gait velocities) and different models with selected transfer functions (TF), spread (S) and misclassified number of values in brackets

From: Motion tracking and gait feature estimation for recognising Parkinson’s disease using MS Kinect

Classification model System parameters Accuracy K-fold cross validation
K = Q K = Q/2
Perceptron Hardlim 88.9 0.11 (4) 0.14 (5)
RBFN 36-7-1 Radial (S:1) 88.9 0.11 (4) 0.08 (3)
RBFN 36-18-1 Radial (S:0.1) 91.7 0.17 (6) 0.14 (5)
NN 36-4-2 TF: sigmoid 97.2 0.08 (3) 0.11 (4)