Table 3.

The top 10 feature weights of two random forest regression models separately trained by all features and significant features.

Significant features		All features
Name	Importance	Name	Importance
HMPDD_8_R2	0.107	creak_std	0.023
MCEP_0_skew	0.096	MFCC_delta_deltas_20_std	0.022
MCEP_17_R2	0.068	Peak-to-RMS_median	0.021
MFCC_deltas_10_alph	0.064	MCEP_9_alph	0.021
MCEP_0_kurt	0.062	MCEP_0_skew	0.019
creak_std	0.061	MFCC_deltas_10_intercept	0.016
MFCC_deltas_10_intercept	0.05	HMPDM_8_min	0.015
MFCC_deltas_16_kurt	0.045	MFCC_delta_deltas_4_R2	0.014
MCEP_3_alph	0.036	MFCC_deltas_7_alph	0.014
Age	0.033	HMPDD_8_R2	0.014