A Multiple-Classifier Framework for Parkinson’s Disease Detection Based on Various Vocal Tests

Table 1

Time-frequency based features presented in Parkinson speech dataset with multiple types of sound recordings [13].

Feature number	Feature name	Group
1	Jitter (local)	Frequency parameter
2	Jitter (local, absolute)
3	Jitter (rap)
4	Jitter (ppq5)
5	Jitter (ddp)
6	Number of pulses	Pulse parameters
7	Number of periods
8	Mean period
9	Standard deviation of period
10	Shimmer (local)	Amplitude parameters
11	Shimmer (local, dB)
12	Shimmer (apq3)
13	Shimmer (apq5)
14	Shimmer (apq11)
15	Shimmer (dda)
16	Fraction of locally unvoiced frames	Voicing parameters
17	Number of voice breaks
18	Degree of voice breaks
19	Median pitch	Pitch parameters
20	Mean pitch
21	Standard deviation
22	Minimum pitch
23	Maximum pitch
24	Autocorrelation	Harmonicity parameters
25	Noise-to-harmonic
26	Harmonic-to-noise