Skip to main content

Table 3 Grid search values for HMM hyperparameter optimization. Grad = gradient of the linear regression fit, var = signal variance and poly fit = the first three coefficients of the second-order polynomial fit

From: Hidden Markov Model based stride segmentation on unsupervised free-living gait data in Parkinson’s disease patients

Parameters

Values

Window size [ms]

100, 220, 500

Feature combinations

[raw] / [raw, grad] / [raw, var, poly fit]

Number of GMM components

1, 3, 5, 8

Number of states for stride model

5, 10, 15, 20, 25

Number of states for transition model

3, 5, 8, 12