Table 3 Summary of the performance of the algorithms for all participants group, considering input 2 as training data.

From: On the prediction of tibiofemoral contact forces for healthy individuals and osteoarthritis patients during gait: a comparative study of regression methods

Function	1st Knee contact peak (N/body weight)								2nd Knee contact peak (N/body weight)
Function	MAE	RPE	RMSE	R	R2	MDF	LCI	UCI	MAE	RPE	RMSE	R	R2	MDF	LCI	UCI
(1) Ensemble trees (bagging)	0.27	9.59	0.35	0.58	0.33	0.03	\(-\) 0.09	0.15	0.45	12.65	0.61	0.10	0.01	\(-\) 0.24	\(-\) 0.44	\(-\) 0.05
(2) Ensemble trees (LSBoost)	0.33	11.29	0.41	0.45	0.20	\(-\) 0.05	\(-\) 0.19	0.09	0.61	19.17	0.76	0.16	0.03	\(-\) 0.17	\(-\) 0.42	0.09
(3) Linear SVR	0.29	10.29	0.35	0.78	0.60	0.23	0.13	0.32	0.37	11.11	0.47	0.39	0.15	\(-\) 0.05	\(-\) 0.21	0.11
(4) Quadratic SVR	0.29	10.17	0.37	0.64	0.41	\(-\) 0.05	\(-\) 0.18	0.07	0.29	10.04	0.40	0.40	0.16	\(-\) 0.20	\(-\) 0.32	\(-\) 0.08
(5) Cubic SVR	0.22	7.80	0.29	0.53	0.28	\(-\) 0.06	\(-\) 0.16	0.04	0.52	15.25	0.67	0.08	0.01	\(-\) 0.28	\(-\) 0.49	\(-\) 0.07
(6) Gaussian SVR	0.19	6.85	0.22	0.87	0.76	\(-\) 0.15	\(-\) 0.20	\(-\) 0.09	0.49	15.28	0.57	0.06	0.00	0.08	\(-\) 0.12	0.27
(7) Linear regression	0.20	7.37	0.25	0.77	0.59	0.12	0.05	0.20	0.53	15.39	0.68	0.28	0.08	\(-\) 0.16	\(-\) 0.38	0.07
(8) Lasso regression	0.20	7.09	0.24	0.72	0.51	0.06	\(-\) 0.02	0.14	0.62	17.98	0.79	0.32	0.10	\(-\) 0.23	\(-\) 0.49	0.03
(9) Ridge regression	0.27	9.76	0.37	0.58	0.34	0.15	0.03	0.26	0.59	17.45	0.72	0.28	0.08	\(-\) 0.02	\(-\) 0.26	0.23
(10) Binary decision tree	0.46	15.84	0.60	0.41	0.16	0.11	\(-\) 0.10	0.31	0.55	16.37	0.68	0.28	0.08	\(-\) 0.09	\(-\) 0.32	0.15
(11) GR (K.-exponential)	0.22	7.68	0.25	0.86	0.74	\(-\) 0.19	\(-\) 0.25	\(-\) 0.13	0.46	14.08	0.53	0.09	0.01	\(-\) 0.05	\(-\) 0.23	0.13
(12) GR (K.-squared exponential)	0.19	6.76	0.24	0.85	0.72	\(-\) 0.12	\(-\) 0.19	\(-\) 0.05	0.42	12.75	0.52	0.09	0.01	0.00	\(-\) 0.17	0.18
(13) GR (K.-matern 32)	0.19	6.64	0.22	0.89	0.79	\(-\) 0.16	\(-\) 0.21	\(-\) 0.11	0.48	14.30	0.59	0.22	0.05	\(-\) 0.08	\(-\) 0.28	0.12
(14) GR (K.-matern 52)	0.20	7.06	0.23	0.89	0.79	\(-\) 0.17	\(-\) 0.23	\(-\) 0.12	0.48	14.36	0.56	0.07	0.01	\(-\) 0.08	\(-\) 0.27	0.11
(15) GR (K.-rational quadratic)	0.20	6.94	0.23	0.89	0.79	\(-\) 0.17	\(-\) 0.22	\(-\) 0.12	0.47	14.12	0.54	0.06	0.00	\(-\) 0.06	\(-\) 0.24	0.13
(16) ETSVR-Kernel linear	0.32	11.29	0.44	0.57	0.32	0.21	0.07	0.34	0.47	14.13	0.58	0.06	0.00	0.02	\(-\) 0.18	0.21
(17) Kernel ridge regression	0.37	12.95	0.54	0.55	0.31	0.24	0.07	0.40	0.51	15.42	0.62	0.07	0.01	0.02	\(-\) 0.19	0.24
(18) Nyström ridge regression	0.68	23.44	0.90	0.23	0.05	\(-\) 0.12	\(-\) 0.43	0.18	0.75	22.06	0.89	0.44	0.19	\(-\) 0.04	\(-\) 0.35	0.26
(19) DNNE	0.44	15.34	0.53	0.71	0.50	\(-\) 0.02	\(-\) 0.20	0.16	0.70	22.09	0.85	0.33	0.11	\(-\) 0.32	\(-\) 0.59	\(-\) 0.05
(20) kNN weighted Mean	0.49	16.93	0.58	0.53	0.28	\(-\) 0.49	\(-\) 0.60	\(-\) 0.38	0.46	13.23	0.61	0.02	0.00	\(-\) 0.25	\(-\) 0.44	\(-\) 0.06
(21) RKNNWTSVR	0.28	10.28	0.39	0.67	0.45	0.20	0.09	0.32	0.42	12.32	0.55	0.15	0.02	\(-\) 0.11	\(-\) 0.29	0.08
(22) LTSVR	0.34	11.96	0.48	0.53	0.28	0.19	0.04	0.34	0.62	18.55	0.75	0.28	0.08	0.02	\(-\) 0.24	0.28
(23) Stepwise glm	0.18	6.37	0.26	0.69	0.47	0.00	\(-\) 0.09	0.09	0.48	14.46	0.60	0.24	0.06	\(-\) 0.21	\(-\) 0.40	\(-\) 0.02
(24) Neural networks	0.16	6.12	0.20	0.79	0.63	0.01	\(-\) 0.06	0.08	0.65	20.61	0.79	0.35	0.13	0.19	\(-\) 0.07	0.45

Best results (i.e. highest accuracy) are in bold.
RMSE root mean squared error, R Pearson correlation coefficient, (R\(^2\)) the coefficient of determination, MDF mean delta force, LCI lower confidence interval, UPF upper confidence interval, GR Gaussian regression, K Kernel.

Back to article page

Table 3 Summary of the performance of the algorithms for all participants group, considering input 2 as training data.

Search

Quick links