Identifying changes in dynamic plantar pressure associated with radiological knee osteoarthritis based on machine learning and wearable devices

Table 2 Significantly different features between groups

Number	Features	Mann–Whitney U	p value
1	f_Runloadr_std	721	< 0.001
2	Age^a	1567.5	0.009
3	f_L8PPP_std^a	784	0.011
4	f_Rpeak2_std^a	759	0.013
5	f_RYcopstd_std^a	1318	0.017
6	BMI	1341	0.02
7	f_L1PPP	781	0.021
8	f_L1MAXPG	755	0.025
9	f_L1MINPG	1329	0.029
10	f_L7MINPG	1306	0.03
11	f_L8MINPG	1371	0.031
12	f_R1MAXPG	748	0.032
13	f_RYcopmean	1380	0.033
14	SI_Xcopmean	1314	0.035
15	f_L1MINPG_std	775	0.039
16	f_L8MINPG_std	774	0.04
17	f_Lpeak2_std	791	0.042
18	f_R1MAXPG_std	779	0.049

BMI body mass index, R right, L left, PPP peak plantar pressure, MINPG minimum pressure gradient, MAXPG maximum pressure gradient, cop the center of pressure, SI symmetry index, std standard deviation, peak2 the second pressure peak, unloadr unloading rate
^aOptimal features included in model training

ISSN: 1743-0003