Table 4 Summary of the performance of the algorithms for All Participants group, considering Input 3 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	9.56	0.34	0.80	0.64	0.20	0.11	0.29	0.42	11.77	0.58	0.11	0.01	\(-\) 0.19	\(-\) 0.38	0.00
(2) Ensemble trees (LSBoost)	0.26	9.16	0.36	0.76	0.58	\(-\) 0.05	\(-\) 0.17	0.08	0.54	15.55	0.74	0.00	0.00	\(-\) 0.30	\(-\) 0.54	\(-\) 0.07
(3) Linear SVR	0.10	3.93	0.14	0.93	0.87	0.06	0.02	0.10	0.16	4.71	0.21	0.94	0.88	\(-\) 0.07	\(-\) 0.14	\(-\) 0.01
(4) Quadratic SVR	0.31	10.41	0.44	0.78	0.60	0.12	\(-\) 0.03	0.27	0.20	5.61	0.26	0.92	0.84	\(-\) 0.07	\(-\) 0.16	0.01
(5) Cubic SVR	0.13	4.64	0.18	0.92	0.85	0.12	0.07	0.16	0.18	5.04	0.25	0.93	0.86	\(-\) 0.07	\(-\) 0.15	0.01
(6) Gaussian SVR	0.09	3.37	0.12	0.94	0.88	0.01	\(-\) 0.03	0.05	0.19	5.36	0.26	0.92	0.85	\(-\) 0.07	\(-\) 0.16	0.01
(7) Linear regression	0.67	21.98	0.99	0.79	0.63	0.61	0.35	0.88	0.37	10.23	0.51	0.46	0.21	\(-\) 0.27	\(-\) 0.42	\(-\) 0.11
(8) Lasso regression	0.11	4.13	0.15	0.93	0.86	0.09	0.05	0.13	0.47	13.25	0.65	0.01	0.00	\(-\) 0.31	\(-\) 0.50	\(-\) 0.11
(9) Ridge regression	0.15	5.38	0.20	0.92	0.85	0.14	0.10	0.19	0.32	9.32	0.44	0.61	0.37	\(-\) 0.20	\(-\) 0.33	\(-\) 0.06
(10) Binary decision tree	0.21	7.40	0.25	0.83	0.69	0.05	\(-\) 0.03	0.14	0.54	15.31	0.73	0.14	0.02	\(-\) 0.29	\(-\) 0.52	\(-\) 0.06
(11) GR (K.-exponential)	0.11	4.22	0.15	0.92	0.84	0.06	0.01	0.10	0.38	10.47	0.55	0.23	0.05	\(-\) 0.25	\(-\) 0.42	\(-\) 0.08
(12) GR (K.-squared exponential)	0.10	3.60	0.12	0.94	0.87	0.03	\(-\) 0.01	0.07	0.23	6.57	0.32	0.87	0.76	\(-\) 0.12	\(-\) 0.22	\(-\) 0.01
(13) GR (K.-matern 32)	0.09	3.46	0.11	0.94	0.88	0.02	\(-\) 0.02	0.05	0.23	6.44	0.34	0.87	0.76	\(-\) 0.15	\(-\) 0.25	\(-\) 0.04
(14) GR (K.-matern 52)	0.10	3.70	0.13	0.92	0.84	\(-\) 0.01	\(-\) 0.06	0.03	0.23	6.50	0.33	0.87	0.76	\(-\) 0.13	\(-\) 0.24	\(-\) 0.03
(15) GR (K.-rational quadratic)	0.09	3.55	0.12	0.93	0.87	0.01	\(-\) 0.04	0.05	0.23	6.57	0.32	0.87	0.76	\(-\) 0.12	\(-\) 0.22	\(-\) 0.01
(16) ETSVR-Kernel Linear	0.13	4.75	0.17	0.86	0.74	0.03	\(-\) 0.03	0.09	0.25	7.38	0.34	0.80	0.63	\(-\) 0.14	\(-\) 0.25	\(-\) 0.04
(17) Kernel ridge regression	0.12	4.41	0.16	0.89	0.79	0.03	\(-\) 0.02	0.09	0.29	8.29	0.38	0.72	0.52	\(-\) 0.16	\(-\) 0.28	\(-\) 0.05
(18) Nyström ridge regression	0.13	4.70	0.16	0.88	0.78	0.01	\(-\) 0.05	0.06	0.21	6.19	0.27	0.86	0.74	\(-\) 0.05	\(-\) 0.14	0.04
(19) DNNE	0.46	15.70	0.64	0.73	0.54	0.37	0.19	0.55	0.35	9.88	0.52	0.31	0.10	\(-\) 0.18	\(-\) 0.34	\(-\) 0.01
(20) kNN weighted mean	0.39	13.83	0.47	0.71	0.50	0.26	0.13	0.40	0.55	16.21	0.76	0.15	0.02	\(-\) 0.25	\(-\) 0.50	0.00
(21) RKNNWTSVR	0.13	4.61	0.16	0.89	0.79	0.05	0.00	0.11	0.26	7.55	0.36	0.79	0.62	\(-\) 0.18	\(-\) 0.29	\(-\) 0.07
(22) LTSVR	0.23	8.16	0.29	0.54	0.29	\(-\) 0.09	\(-\) 0.19	0.00	0.41	11.66	0.57	0.37	0.14	\(-\) 0.28	\(-\) 0.45	\(-\) 0.11
(23) Stepwise glm	0.13	5.10	0.18	0.87	0.75	0.02	\(-\) 0.04	0.09	0.45	13.31	0.58	0.13	0.02	\(-\) 0.22	\(-\) 0.40	\(-\) 0.03
(24) Neural networks	0.13	4.56	0.16	0.90	0.80	\(-\) 0.04	\(-\) 0.10	0.01	0.20	6.35	0.23	0.92	0.85	\(-\) 0.09	\(-\) 0.16	\(-\) 0.02

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 4 Summary of the performance of the algorithms for All Participants group, considering Input 3 as training data.

Search

Quick links